Old thread is at bump limit and rapidly (lol) approaching page 10. Someone made a new thread and linked it in the old one but it appears to have been deleted(?).
Pastebin info: http://pastebin.com/9UgLjyND
>>977908
fail crop on the OP image...
anyway, I found a can of spray-on PCB photoresist at the thrift store for a dollar. It looks pretty old, possibly 80s or older. Should it still be good? I have a darkroom already set up for developing film and prints and stuff, so my plan here is to take my camera and point it at my computer screen, get a nice sharp negative on a single sheet of film, then load that into my enlarger, allowing me to make as many copies of boards as I want without fucking around with ink transfer or printers or anything.
>>977915
other side of the board
>>977908
Any idea why the other thread was deleted?
>>977915
Try dry film photo resist Kolon Accuimage.
What is the best kind of decoupling capacitor for an audio spectrum analyzer circuit.
Aside from some digital circuitry to handle the multiplexing it's all analog. No microcontrollers or anything fancy like that.
I have a 470uF electrolytic capacitor for bulk decoupling on the Vcc input on both my PCBs. I need to know what to use for local decoupling. I have:
3x TL074 quad op-amps for 10 2nd order MFB bandpass filters, an input L-R active mixer, and an inverting amplifier
5x TLC2272 for active rectification for my peak detector
1x CD4017 Johnson counter
3x CD4016 quad switch (basically acting like a 1-10 analog mux)
1x CMOS 555 timer (12Vp 2kHz 50% duty cycle)
1x LM3915 LED driver
10x 4G3Y3R LED bars (each has a 2N3904 NPN transistor to switch power to the anodes when a signal is received on the base from the CD4016s)
Everything runs off 12VDC and the max current draw is somewhere around 100-130mA.
What kind of decoupling caps should I use for all the above?
I originally had 2.2uF polyester film capacitors but the PCB footprint is rather large (22.5mm pin pitch) and the max board size I can order is 60sq in. I also had similar 1uF film capacitors for AC coupling (15mm pin pitch) which I will also have to change for something smaller preferably.
Ceramic seems to be recommended but I didn't think it was well suited for low frequency applications. I also consider Al-polymer electrolytic caps (which is what I was planning to use for the LED bars 10uF).
>>978134
You'll probably be fine with ceramic caps. They can cause minor distortion in audio circuits under some very specific circumstances, but since you're making a spectrum analyzer, the circuit is just feeding into a visual output; it's not like it will affect the sound.
Further reading on capacitor types: http://sound.westhost.com/articles/capacitors.htm
Also, the CD4066 analog cmos switch chips are usually better suited for audio than the 4016. Then again, for the same reasons as above, super high fidelity is not really an issue. I think they are pin-for-pin compatible if you wanted to switch them out, though.
>>978062
As /jp/ would lovingly say,
New
Shitty
Janitors
>>978169
>Also, the CD4066 analog cmos switch chips are usually better suited for audio than the 4016
I'll bear that in mind for any related work I do in the future but I can't make that change now unfortunately. I'm doing this as a design project for a class and I really wanted to have this board ordered a week or two ago but I was having technical problems I had to resolve first.
I'll look into some ceramic caps though, the smaller footprint is more important than a small amount of distortion at this point.
>>978134
I would sprinkle around 0.1uF 0805 ceramics. Unless it's for AC coupling the signal path I wouldn't worry too much about second-order characteristics.
But your layout is probably going to be a bigger impact on how well your circuit than anything else.
>>978134
Local decoupling caps need low series resistance and inductance, which normally favours high-k ceramic capacitors.
Where relatively large values are required, it's common to use a tantalum capacitor of the required value in parallel with a smaller ceramic capacitor.
The non-linearities of ceramics aren't an issue when used as decoupling caps. It's only an issue if they're used in the signal path.
>>977915
>It looks pretty old, possibly 80s or older. Should it still be good?
A bottle of 30 years old reactive chemical with a nominal shelf life of 1-2 years, stored who knows how? Well, it might still work, particularly if you play with the exposure and developer a bit.
>plan here is to take my camera and point it at my computer screen
Are you trolling? Well, whatever.
>get a nice sharp negative on a single sheet of film
On normal 35mm camera film? If your photoresist is ancient shit, at least you should maximize the contrast, as playing with low contrast films is suffering. You're also likely to need a positive.
>load that into my enlarger
Your photoresist is likely to need UV and - at least when compared to normal photographic paper - massive amounts of it.
If you insist on going this route, at least consider making a 1:1 sheet film positive (using the highest contrast sheet film you can find) first and using some decent UV lamp/source for the exposure. Expect to waste lots of photoresist and PCB for experimentation.
>>978280
>Are you trolling? Well, whatever.
Yes, it's jank as fuck, but it will work. Scan lines aren't a problem with the shutter speed set longer than 1s.
>On normal 35mm camera film? If your photoresist is ancient shit, at least you should maximize the contrast, as playing with low contrast films is suffering. You're also likely to need a positive.
It's medium format. The magnification between the negative and the board is less than 2x. I also have some x-ray film and some other stuff that's more intended for reprographic rather than photographic use.
>If you insist on going this route, at least consider making a 1:1 sheet film positive (using the highest contrast sheet film you can find) first and using some decent UV lamp/source for the exposure.
Yeah, I forgot about the UV part. I doubt my enlarger will work. I do have a contact printer box though.
>>978282
Well, good luck.
I've used film in the past and the principal problem was always contrast, particularly how opaque/black the film was. The shittier the resist, exposure and development were, the more critical good films were.
How can I make liquid flux from paste flux or rosin flux? Paste form is just rosin and petroleum jelly melted together, isopropyl alcohol didn't work.
>>978317
I tried "thermal-paste-cleaning-90%-something" IPA. The liquid got a little amber, but I still didn't feel confident using it.
How are you supposed to use solid rosin, anyway?
Yes, I tried solid rosin and paste rosin flux.
>>977908
I just ruined part of my ebay 15 dollar 3 axis DRO by putting a tin can voltage regulator in its socket backwards.
An inductor blew up and I'm thinking the regulator probably is fucked as well as a bunch of other shit.
I'm thinking of going to goodwill and buying a wall ac/dc adapter to power it off of but I don't know the exact power requirements of the unit...
>>978515
Show us an ebay link to your unit, then.
Power supply voltage has to match, current has to be equal or above power draw.
>>978542
It's pretty old and I'm pretty sure you won't be able to find out what power it needs beyond "plugs into the wall"
the blown regulator is PIC625 which outputs 60v, which obviously wouldn't jive with a bunch of 74 series logic and some 7 segment displays
Could I just find a common ground and the power rail and give it 5v?
>>978549
74 series seems to work on 5V. Why were you using a 60V regulator in the first place?
I'm not knowledgeable in PICs or anything of the sort, but for 7 segment displays and 74 logic it should work. Check your display resistor values for working with 5V
>>978578
That's just what came with it
stupid question - say i have a ferrite coil with a tap. i only want to input up to the tap. what do i do with the other end of the coil? can i just leave it float, or do i have to do something with it?
i do not at this moment wish to get into winding my own coils. basically i am looking at buying premade loopstick 680uH with a tap at 470uH.
Data sheet says 180-450uH loopstick supported by this IC.
wanted to connect from rfgrnd to loopstick grnd and ami to tap.
picture unrelated.
>>978510
Solid rosin is used for reconditioning soldering irons, but you can use it as regular flux used on a PCB.
https://www.youtube.com/watch?v=yTV8gtSX6tg
You're apparently supposed to grind it up and leave it to soak (for a few DAYS) in IPA. Note that IPA dries like a bitch, so you'll have to let it dissolve in some kind of airtight container.
>>978649
You can let it float. There will be some voltage induced on that part of the coil but it's harmless.
>>978649
You can let it float, even though the self-resonance of the unconnected part will have some minor effect on the other part.
An alternative would be to unwind the offending part. And since the inductance is somewhat too high for your IC, you might want to unwind a small part of the 470µH part as well. If you do that, remember that the inductance is proportional to the square of the turns and you should remove only 20% or so. If it uses Litz wire (multistranded, quite likely), it's probably better to keep your hands off from the 470µH part, though.
remeber my VCO form the other thread?
i put it on a panel
>>978785
>>978786
I also added a transistor to the saw wave to bring it up to the level of the square.
yesterday I got my linear power for the synth running. so much smoother than the damn smpsu i used.
>>978788
>transformer/inductor looks rated to like 6000 amps
>5000 caps
I don't understand
>>978816
You should spend like.. 10$ and get a printed board.
>>978816
cont;
please understand that this is 3 rails with separate transformer taps, capacitors and regulators on a single board.
>>978817
and spend the as much time designing it, then wait for weeks before i can build it?
prefabing something so simple isn't justifiable to me.
i cranked this out in one day.
>>978811
>>transformer/inductor looks rated to like 6000 amps
are you high?
look at the cup next to it for scale.
this is like a 100Ws or something.
>>977908
I shorted something on a board and my green led turned red
the thing is it only has 2 leads? what the fuck?
>>978855
not sure what you are trying to say
>>978855
Massive overheating can change the LED die material enough for the color to change.
Not sure if this is related to what you were trying to say.
can i use a simple XOR gate as phase detector for sine waves or is it only used for square waves?
>>978884
could work
OR - add a schimtt trigger to make it a squarewave, then use the xor maybe?
>>978884
Use the phase detector 2 circuit of a CD4046.
>>978884
If the sine wave has the correct bias and amplitude, feeding it to a digital logic gate will turn it into a square wave.
>>978855
> the thing is it only has 2 leads?
You can get red-green LEDs with two leads. The two LEDs are anti-parallel (i.e. in parallel, but the directions are reversed).
I have decided I want some non-metallic screwdrivers and needle-nose pliers. Seen anywhere I can buy some?
The first post in the "DIY injuries" thread (showing an exploded capacitor) reminded me that a few times I've killed electronics projects because I dropped a screwdriver or pliers on them while they were powered up.
I know about insulated screwdrivers, but they wouldn't have been any help, since just the metal tips being conductive would have done the damage.
[--And anyway, you can make insulated screwdrivers yourself just by putting heat-shrink tube around the metal shafts--]
I see there is plastic pot trimmer tools around that are flat-blade screwdrivers, but that plastic is pretty cheap and there's no phillips type available.
,,,,I am thinking more of something like glass-fiber-reinforced plastic, not regular polyethylene,,,
And I don't recall ever seeing plastic needlenose pliers that weren't kids toys.... :\
I could go buy some fiberglass rod at the hardware store and make my own screwdrivers myself, but making any decent pliers (even just for grabbing wires) would be a bitch. And I can't help but think that somebody else has had the same problem....
>>978185
>>978192
I agree with this too. Small value ceramic is the way to go.
If you want to be extra sure that you are getting good filtering, place (1) 0.1uf cap in an 0805 package and (1) 0.01uf cap in an 0603 package as close to each power pin as possible. If you want to be triple sure, add an additional (1) 0.001uf cap in an 0402 package. May be overkill though.
what's the cheapest i can get a "fully functional" pickit for? the chinese knockoff arduinos work just fine when you install their shady usb driver, is that the case here too?
>>979031
don't fuck with chinese arduinos, get the real atmel explained mini from Atmel and you get full AVR studio support with breakpoints and debugging (for 10 bucks)
as for pics, I imagine the chinese ones will be adequate
>>979031
The knockoffs will work with the official drivers, but don't get a pickit, get a K-150 for 10$.
>>978981
you know I'm drawing a huge blank right now, there was an online company I found a few years ago just browsing the internet, that sold spark proof tools.. they used some sort of non-conductive metal. I want to say its aluminum bronze but its not.. IT was some sort of bronze colored alloy tho.
>>977908
how can I cop a powersupply with a 5v and 30-40v rail?
I don't want to waste too much time fiddling with a smps, that's boring
>>979110
how much current on each rail?
>>979110
Most bench power supplies have a fixed 5V rail and one or more variable rail(s) up to 30V.
ebay has old models for like 50$. Search for HP ones, they're really old but still quality psu's.
>>979113
it needs ~ 1 amp on the high voltage rail, and the low voltage rail is for the driver IC of a motor so less than 100 mA
>>979110
> how can I cop a powersupply with a 5v and 30-40v rail?
There's no inherent reason these need to come from a single PSU. Separate PSUs may be easier to find than one which has both. You can get 5V@100mA from any USB port (including phone-charger wall-warts).
I need a more confortable alternative to pic related.
even with improvised keycaps the short stroke and loud click make them uncomfortable. What should be my keyword for searching?
>>978951
>>978889
>>978887
thanks i will look into it. i want to mess around with PLL and would like to make a simple one to create sine waves of different frequency, thus the phase detector. i don't really want to use full blown PLL IC
now what i don't get is how i would realise the frequency divider. use simple T-flip-flops? wouldn't i need quite alot for certain factors then? it's supposed to be for kHz range first then i want to move up towards MHz
>>979232
just search for "push button switch" i guess
>>979232
Push button switch. Check the datasheet for travel and actuating force curve. Curves with very sharp edges suggest clickiness. Check the mechanism (if specified) as well, dome swiches are inherently noisy. Usually the manufacturer also says if some particular model is clicky.
>improvised keycaps
There are plenty of very similar switches with keycaps. You just need to buy them separately. Switches like that usually have square or cross-shaped shaft.
>>978981
You can get plastic trimming screw drivers with phillips, square and hex heads. But yeah, they aren't meant for anything requiring force.
You can also buy ceramic tools. Some are full ceramic, some have ceramic only in the business end or are just coated with ceramic. Pliers with plastic caps on the jaws also exist.
>>979236
A full-blown PLL IC would be easier to use, though.
TFFs aren't really available as separate ICs. You need to use DFFs or just counter/divider ICs. 4040 is well-known general purpose divider IC and 74hc161 a well-known counter. If the input frequency is in kHz or low MHz range, you'll need to square the input signal first.
There are also dedicated dual modulus dividers for high-frequency PLLs.
>>979232
>I need a more confortable alternative to pic related.
>even with improvised keycaps the short stroke and loud click make them uncomfortable. What should be my keyword for searching?
You can get those cheapy buttons with keycaps already included, but I agree--they still just don't feel that nice. But they are cheap and it costs a lot more money to get away from them tho.
compounding the problem is that most nicer + cheaper push buttons are panel-mount, with no easy way to put them right on a PCB. The common video-game push buttons are like this: they use a microswitch and they work well but they're huge and panel-mount only.
search for "computer keyboard switches" or for "computer keyboard buttons". this gets you the buttons that are used in computer keyboards.
Note that for these kinds of buttons, the buttons and keycaps are usually sold separately. The buttons can cost $2 each and the caps can cost $1 each,,,, even on the cheap China sites.
Considering that, you can just go to walmart and buy a cheapo keyboard for under $10, and scrap that.. Those keys already have keycaps and would cost you only ten cents each. :>|
-----
If you want PERFECT buttons, then go to Digi-key and search for NKK push-button switches. They will cost you ~$10+ each, but they look and feel fantastic.
>>979245
>A full-blown PLL IC would be easier to use, though
sure it's easier, but i won't learn anything in the progress and won't have any fun. i get the theory and all but want to try and implement it in a simple way myself
>You need to use DFFs or just counter/divider ICs
counter actually sound like a really good idea
>If the input frequency is in kHz or low MHz range, you'll need to square the input signal first
because the counter wouldn't recognize it as a clock otherwise?
>>979125
so ugh, an 30-40volt powersupply and a shunt regulator?
put an effect resistor in series with the regulators of it get too hot
>>979251
>because the counter wouldn't recognize it as a clock otherwise?
logic chips often have quite specific requirements on their inputs, such as min. rise time, and if you dont provide what they need, the chip's behavior is not guaranteed.
>>979256
>shunt regulator?
shunt regulators are crap, dude, very inefficient. a series regulator, like a 7805, or a switching equivalent is simpler, cheaper andless wasted heat.
>>979251
>but i won't learn anything in the progress
I disagree. There are plenty of PLL ICs which require you to know pretty well what you're doing and to provide your own loop filter and VCO and possibly other stuff as well. Many of them also allow easy experimentation on topics you don't fully understand yet.
Also, considering the questions you're asking, I see a non-negligible chance of you failing to get your own design to work, which tends to be demotivating and counterproductive.
>won't have any fun
Well, the choice is yours.
>because the counter wouldn't recognize it as a clock otherwise?
Or counts multiple edges or counts totally wrong.
>>979265
>7805
that's what i meant, i got the terminologies mixed up :D
>>979267
>considering the questions you're asking
i'm just not familiar working with logic ICs so i like to double check first, ya know
>>979269
Like I said, the choice is yours.
As an afterthought: if you think you really need TFFs, then those are easiest to emulate using JK FFs.
>>979236
> now what i don't get is how i would realise the frequency divider. use simple T-flip-flops? wouldn't i need quite alot for certain factors then? it's supposed to be for kHz range first then i want to move up towards MHz
You can just use a programmable down-counter: use the VCO as the clock, so it counts down one step per cycle, when it reaches zero, reload it with the division modulus.
Traditionally, divide=by-10/11 counters were used, which divide by either 10 or 11 depending upon the state of an input pin. The output is used to clock a pair of programmable counters so that the overall division ratio is N=10*A+11*B for chosen A and B.
This allows you to get any N>=100 in unit steps, while the A/B counters are clocked at 1/10th or 1/11th of the VCO frequency. Only the dual-modulus counter (which is a simpler circuit) needs to run at the full VCO frequency.
Nowadays, you can get programmable counters which run well into the hundreds of MHz range.
Hello
So I got this not working tv for free.
This board had this very intense noise.
After visual inspection, I changed 9 capacitors, because top of them indicated that they are bad.
So I plugged it in again, noise is gone, but now it does these 2-3 clicks per second. If I unplug all connection from board and just leave input connection it does 1 click per second. Can't determine where it comes from.
Can it be one of transformers?
I marked in picture that noise comes from somewhere in the middle.
Can it be that big ass capacitor?
Any ideas what can cause that click noise?
>>979312
>caps are literally inside the heatsink
planned obsolescence people
>>979322
maybe just shielding that the chinks thought could also work as a heatsink?
>>979324
well the rectifiers are mounted right next to the caps, they're going to get bretty hot
>>979353
'Never attribute to malice that which is adequately explained by stupidity.'
Besides, right next to hot components is usually the most natural place for electrolytics in power supplies.
>>979384
never fucking reply to me again unless you're contributing to the thread
>>979391
>waaah don't reply to me!!!
Fuck off. I'm not even that guy.
>>979391
Don't reply to me or my son ever again!
Will circuit simulators ever not be complete trash?
>>979598
Why you are call them trash?
>>979709
I'm not the other dude, but
> tried for an hour to get a simple 555 astable to oscillate in a certain simulator
> zero success, even if i've built dozens exactly like it in real life
> one of the sample designs that came with the program was a 555 astable
> loaded it up and it worked fine
> recreated it exactly from scratch: wont oscillate
> realize the program was specifically designed to frustrate people
> for example, if you have the oscilloscope display on, and you make a small change to the circuit, the scope wont work unless you close it and open it again
> seems all engineering software was designed in the user-hostile era of the 1990's
>>979391
>never fucking reply to me again unless you're contributing to the thread
pretty sure this is just bait.
seen this a few times on diy now.
no need to reply to it.
>>979742
yeah i hate them too alright. there is always some little detail you miss and nothing works. you just have to get used to it, then everythings fine
What's the best way to jumper all 6 screws on the right side of this terminal block. Was going to just use wire but maybe some Anons have a cleaner way.
>>979312
So the update.
I made tv work, finally!!
So where was the problem.
I took out big ass capacitor out just to see it's parameters, and look what I found. A hole in it. I put isolating tape over it, soldered it back to it's place and tv works.
Gonna order new capacitor and change it later.
Best place for used test equipment? Looking for just a basic tektronics or similar analog scope but eBay prices seem rediculous
>>979851
I love my ebay tektronix tds 360 I bought for 150$
Can someone explain to me like I'm retarded how to enable a 3.4v 1.9A signal with a 4.5v 50mA control signal with a transistor/mosfet/IGBT/whatever?
>>979851
local uni
>>979873
igbts are ``great'' for high voltages, and bjts are great for when you're a penny-pinching jew or when you're doing things other than pure switching. none of these things apply here so let's use an n-channel mosfet.
to simplify, an nmos is a 3 pin electronic on-off switch. it has +, -, and control pins that are called drain, source, and gate respectively. to turn it on, you have to bring the gate pin ~3 volts higher than the source pin. the gate doesn't need any constant supply of current like with a bjt, it only cares about voltage.
to use it, you connect the source to ground, the drain to your load, and the gate to one of your microcontroller's output pins as shown in the pretty picture. you can connect it in other ways but those are less convenient.
since mosfets aren't ideal, you also need the two resistors in the bottom of the picture. the right resistor makes sure the mosfet stays off by pulling it to ground. you don't need this if you make sure your uc's control pin is always programmed as a high or low output (ie not an input), but you're a terrible programmer and you know it so this helps prevent confusing mistakes. the left resistor limits the current the mosfet draws when you turn it on. even though mosfets are "voltage controlled", they have a small input capacitance that will need to be charged and discharged when you turn it on or off. that can draw a big peak current for a very short time, and the resistor limits that.
what's the thing you're switching? depending on what it is, you might need more components like a diode.
>>979846
i dont think you need a new cap. the hole is just in the plastic, which is only used for isolation. you can probably just jam a bit of thick paper between the cap and the wire on the PCB to keep them from shorting. your tape will probably melt like the plastic did.
>>979873
mosfets were created by satan and work by black magic. if you want something you can understand, you get a darlington NPN power transistor, like the TIP142 (or any power darlington) and you wire it like in the pic. connect your 3.4V supply where it says 5V, and GPIO is your 4.5V control signal goes. connect the grounds of both supplies together, and to the emitter of the TIP142.
>>979895
there's a million different n-channel mosfet models so i won't just give you a part to use. you're probably picking them out of a bin at a university or something anyway. the things to look for in a model that will work for you are:
>threshold voltage (vgs th), how much higher the gate voltage needs to be than the source. should be <3.5v.
>resistance (rds on), the resistance of the mosfet when it's turned fully on. needs to be <250 milliohms unless you're heatsinking it. 250 milliohms is really pushing it without a heatsink.
the rest doesn't really matter for what you're probably doing.
>>979898
npns are dumb, ignore this vicious shitposter and his endless lying.
>>979895
also i wasn't paying attention with that bottom resistance. pretend it says >=27k. the reason is that 27k resistors have an orange stripe on them instead of a yellow stripe, which looks nicer imo.
>>979902
>npns are dumb, ignore this vicious shitposter
posted by one of satan's minions for sure.
did you know you can destroy a mosfet just by touching it? how absurd is that. avoid at all cost because, even if it survives your handling, if the circuit doesnt work, you're left to wonder, ''is it dead or did i just wire it wrong." power NPNs, on the other hand, can survive anything.
http://www.irf.com/technical-info/appnotes/an-955.pdf
>>979922
ignore the document it's a false flag by the bjt lobby
>>979895
it's a motor control IC
the IC will be the one switching the inductive load so it should be OK
>>977908
I've got a machine with an EPROM (27C64) hooked up to a Z180.
I can't find any docs about this particular machine's setup, so I need to read the firmware off the EPROM and, eventually, write new firmware.
Is there anywhere I can get a sub-$300 programmer, and can programmers typically dump the existing ROM to a file?
>>979964
what are you switching then, the supply to the ic? every full or half bridge driver i've seen just has logic inputs to turn it on or off. i'm sort of confused.
>>979898
>>979895
This was very helpful
What I had been doing was buying anything with a power transistor package (the one with a heatsink) and I actually have a few tip120s and tip125s lying around, so I'll just try again (although I haven't been able to get them to work)
Also I heard the EEV blog lad say IGBTs are more durable, while MOSFETs have better specs but will blow up easier.
Maybe I can just remain blissfully ignorant to different types of transistor for a while longer... I don't plan on designing any amplifiers.
>>979965
I presume you could build one out of a $5 Pi Zero.
>>979968
My IC is a Toshiba TB6560AHQ which isn't an Hbridge, it's a chopper type bipolar stepper motor driver
It has a low power control circuitry input and a high power input, which it switches itself.
>>979965
New from China, used from eBay.
Or build your own. Reading is very easy and writing isn't that difficult either. Or substitute an EEPROM when you need to write your own stuff on it.
>>979969
IGBTs are preferred for high voltage, MOSFETs for lower voltage. Typical Ids*Rds[on] for a MOSFET with Vds[max] below a couple of hundred volts is less than typical Vce[sat] for an IGBT, meaning lower power dissipation.
One other difference between them is that MOSFETs conduct in either direction, while BJTs (incl. IGBTs) only conduct one way (collector-to-emitter for NPN, the reverse for PNP).
This gives MOSFETs an advantage for driving inductive loads, as you can use them as synchronous rectifiers (Ids*Rds[on] is typically lower than the voltage drop of a diode, even Schottky diodes).
>>979979
decent chart to expand on this
>>979974
> It has a low power control circuitry input and a high power input, which it switches itself.
In which case, you'll probably need the switch (FET/BJT) on the high side.
Which is a bit of a nuisance. In this case, you should just be able to use a p-channel MOSFET or PNP transistor, as your control voltage is above the power rail. In the more common case where the control voltage is below the power rail, it typically requires two or three transistors depending upon whether you can use an inverted control signal and/or whether you can use a pFET/PNP as the switching element.
Why do the arduino relay shields are set up in a way that you shift the relais by pulling a pin to GND. I suppose they shift transistors who shift the actual relays but still: Why dont they shift it by shifting an arduino pin to HIGH?
>>980002
It's easier to drive a load on the high side (between the positive rail and the output) than on the low side (between the output and ground), particularly if the load voltage is close to or above the logic voltage.
See >>979991 for examples of driving a low-side load (high-side switch). Driving a high-side load (low-side) switch only needs one transistor.
>>980018
So all the transistors in the >>979991
examples are used to gap the bridge between the logic voltage and the required voltage for the final transistor that can shift the 25V?
In a case where the logic voltage level was 25V (hypothetically speaking) i still wanted to put all the transistors in between (or even better using a low side switch) to be sure i dont stress the limits of the microcontroller?
In the case of the arduino (5V logic, 5V coil relay), do they just want to make sure the atmega doesnt fry itself by generating 5V at to much of a current?
>>977908
/ohm/ you guys might be able to help me.. I dropped my LG G3 in water while fishing, it fucked the screen.. but only the LCD. I'm going to buy the whole front bezel just so I don't have to dick with trying to keep lint and dust out between the touch glass and LCD.. my question is. is there a way to waterproof my phone while I have it apart? or am I better off just buying a waterproof case and dealing with the bulkyness.
>>980196
unfortunately no, waterproofing has to start at the design stage. the only holes in your phone are ones that really need to be there for it to function fully.
>>980197
I figured the microphone and charger holes could be plugged with some sort of rubber.. and microphones/speakers could be covered in a super thin plastic. guess I'm buying a case!
I'm admittedly new to electronics.
I have a SPDT slider switch with 3 pins
I don't know which pins to wire up, I'm only gonna be using it as a spst switch
how do
>>980196
If you intend to use your phone while you're fishing I'd suggest a lifeproof case, which doesn't add much bulk - make sure you attach a float to the lanyard loop if you're near deep water. Waterproof doesn't help you if your phone is at the bottom of the ocean.
If you only intend to have your phone one you and just need to have look at it occasionally I'd recommend one of those transparent dry bags - they're cheap and easy to take on and off, but it's kinda annoying to use the screen through the plastic and people will have difficulty hearing you on calls. Wear it on a lanyard around your neck, or put it in your pocket and tie the lanyard to your belt.
There is no practical way to /diy/ waterproof your phone internally. Anything you put over the circuit boards has a good chance of either; a) not actually being waterproof, b) causing the casing not to fit properly back onto the phone, or c) fucking up the heat management and making your phone cook/throttle itself to death. There was at one point a company which promised to spray something onto people's phone circuit boards for $50 that would make them resistant- it was a scam and they just gave back people's phones unaltered.
Sealing the edges might seem like a good idea, but again, you'd have no way of knowing whether it had worked, and it's likely to cause problems if you have to open your phone for repairs again.
Any more major modifications, like 3D printing / machining a casing that seals with an o-ring, will be incredibly time consuming / expensive (though fun), and likely add more bulk than a thin waterproof case.
>>980210
hi christian
if your switch looks like pic related, that's the diagram. if you're confused and you have a multimeter just measure resistances to figure it out.
>>980215
And heck, if you're a cheapskate, just chuck your phone in a ziplock bag when you're near water.
Daily reminder it's misleading to think of transistors as switches. Rather, they should always be treated as amplifiers. Is: a MOSFET connected to a microcontroller isn't switching a load. The MOSFET is amplifying the gate sign and using the source as a power supply just like (in principal) an open loop op amp.
>>980276
when you're explaining transistors to a newbie that only wants to make his motor spin, this qualifies as intentional obfuscation
why do white LEDs need more volts than other LEDs
>>980296
chemistry
>>980276
Uhh dude think of a MOSFET as a voltage-controlled resistance. With a capacitor at the gate.
>>979991
the second circuit has a major-league bug in it.
>>980034
> So all the transistors in the >>979991 examples are used to gap the bridge between the logic voltage and the required voltage for the final transistor that can shift the 25V?
They're needed if the logic voltage is below or barely above the load voltage (a BJT needs a base-emitter voltage of ~0.7V to turn on; a MOSFET needs a gate-source voltage of at least 3V, sometimes more).
I> In a case where the logic voltage level was 25V (hypothetically speaking) i still wanted to put all the transistors in between (or even better using a low side switch) to be sure i dont stress the limits of the microcontroller?
If the load requires more current than the logic can source, you need at least one transistor. If you can put the load on the high side, that's all you need, even if the load voltage differs from the logic voltage
> In the case of the arduino (5V logic, 5V coil relay), do they just want to make sure the atmega doesnt fry itself by generating 5V at to much of a current?
The uC typically can't provide enough current to drive a relay coil directly, so a transistor is used as a current amplifier. Putting the load on the high side means that you get nearly 5V (rather than 4.3V if it was on the low side), and would also allow you to control it from a uC using 3.3V logic (you'd need a 5V supply for the relay shield).
>>980359
You don't actually need a resistor if Q2's beta is known and reasonably stable. R2 limit's Q2's base current which in turn dictates its collector current. But a resistor is probably wise in most cases (and you don't have to be so picky about Q2).
Also, Q1's base should ideally be pulled up to the supply rail for faster turn-off (BJTs turn off faster if you actively pull the base to the emitter voltage rather than just letting it float there).
>>980296
> why do white LEDs need more volts than other LEDs
Quantum physics.
Shorter wavelength means more energy per photon which means a higher voltage. The energy of a photon is E=h*c/λ where h is the Planck constant, c is the speed of light, and λ is the wavelength. h*c ~= 1.24e-6 eV.m, so for λ=700nm (red), E=1.77eV, while for λ=400 (blue) E=3.1eV. To get a 3.1eV photon, you need at least 3.1V.
Most white LEDs are actually a blue or UV LED which drives a white phosphor. This allows any colour temperature or profile to be obtained by choosing the appropriate phosphor.
The other possibility for a white LED is to use matched red, green and blue LEDs, but that still requires at least 3.1V for the blue, and they're more expensive.
>>980276
It's misleading to think of transistors as amplifiers if you're intentionally driving them between off and saturation (and often going to great lengths to avoid the region in between).
>>980360
The relay coil i was using has a coil that draws a current of 30ma. Apparently every arduino pin can give out up to 40 ma.
>>980380
iirc an entire 328p port can do 40ma, and each pin can do 20ma. i'm too lazy to check the datasheet but you might want to.
If an application circuit of an IC's data sheet doesn't include any decoupling caps and pull-ups, do I still need them?
>>980481
disregard this i suck cocks
40ma is pin rating, 200ma for vcc
>>980485
I would, just to be safe.
>>980485
It depends upon the IC.
Decoupling caps are typically only needed for synchronous logic, where many gates change state in sync with the clock.
Pull-ups are only needed on open-collector outputs or on inputs which would otherwise be left floating.
Decoupling caps typically aren't shown in schematics, pull-ups normally are.
not entirely sure if this is the perfect place for this question, but i was thinking of making a new pair of speakers for myself, and although i can do the electronics of it all, i was wondering if it's actually worth it, and how much of it depends on things such as acoustics of the enclosure
i.e. how hard is it to make speakers that actually sound good?
>>981017
Building speakers is primarily about the design of the enclosure. The electronics are negligible, just a matter of choosing the right speakers and a crossover
>>981039
yeah, i've been reading since posting
do you know any good resources off hand?
>>978788
That looks pretty cool. I've been wanting to build a modular synth for a little while now, could you post your schematic?
>>981113
pretty much this: http://www.modularsynth.ru/wp-content/uploads/2013/10/psu_bipolar_schematic_v1.2.png
except my 5v has its own transformer tap with rectifier and caps, not just regulated from the 12v.
i did that for the lulz, if 5v from the 12v rail is good enough for doepfer, its good enough for us.
FUCKING COCKSMOKING MOTHERFUCKING ASSHOLES
WHY IS IT SO HARD TO REPLACE CAPACITORS ON A FUCKING MOTHERBOARD
I've been working on this shit all day and this solder is just not coming off. Any ideas?
>>981564
Are you using a hot-air wand or a soldering iron? Is it hot enough for lead-free solder?
How are you removing the solder? Suction or braid? Melting solder won't make it magically vanish.
>>981574
Hakko 926 station, the temperature isn't calibrated right but it's definitely hot enough. Tried both suction and braid, neither one is working.
To any electronic engineers here, do you have any recommendations for introductory textbooks on image processing? If they include lots of matlab stuff that would be a bonus.
Working in my uni on image processing related stuff in the summer and being a second year all I've done so far is a class on general signal processing, would like to have a better grasp of the subject before I start
Would running a 12v fan on a 5v power source kill the motor with extended use? If yes, how to make simple 5v to 12v converter? Electronic newb here, working on my first project.
>>981603
> Would running a 12v fan on a 5v power source kill the motor with extended use?
No.
If it's a conventional mechanically-commutated DC motor, it will just run slowly. If it's a "brushless DC" motor, it probably won't work.
> how to make simple 5v to 12v converter?
Boost converter.
The MC33063/MC34063 is fairly straightforward to use, although you need to understand the basic operation of a boost converter in order to choose the correct component values based upon input/output voltage, output current and switching frequency.
You can get self-contained modules which incorporate the IC, inductor and capacitors, but they tend to be more expensive than the individual parts.
hehe
assmann
>>981719
fucking saged
people like you are why we cant have nice things
![latest[1].jpg](https://i.imgur.com/t7uago0.jpg "latest[1].jpg")
>>981721
If you expect a no-nonsense attitude, I do not know why you choose to use 4chan.
Also, sage is not supposed to be an indication of displeasure. This post is a sage because it's not worth bumping the thread just to call you a faggot.
how bad is the forward voltage tolerance on typical jellybean leds? the datasheet for my part doesn't specify and i haven't ordered it yet so i can't just measure a few.
Anybody play with ESP8266 stuff? All the stuff I've seen is some Lua crap or whatever. Is there an open toolchain for this shit?
>>981846
>Is there an open toolchain for this shit?
https://github.com/pfalcon/esp-open-sdk, it's based on the original Espressif SDK but has some more FOSS components.
I found it awful to work with though, callbacks and void pointers everywhere - so I'm using Arduino, which was ported to the ESP8266 (https://github.com/esp8266/Arduino)
There is also another C SDK from Espressif that uses RTOS, but I haven't tried it.
>>981850
Thanks. Sounds like a pain in the ass. I must be getting old.
>>981887
you can program it with the arduino ide.
not sure how far the libraries have gotten, been a while since i messed with it.
i got it to connect to ssids and set up a specific IP. and respond with http pages when adressed in a browser.
Ihad no real use for it so i never got any further.
>>981765
> how bad is the forward voltage tolerance on typical jellybean leds?
Pulling a few datasheets at random, typical specs for Vf@20mA for red indicator LEDs are min=1.85V, typ=2V, max=2.5V.
Distribution tends to be roughly Gaussian, so most of them will be much closer to the typical value. The specs basically just say that if anything outside that range is considered defective.
If you look at the row of LEDs on the front of a typical consumer-grade ethernet router, the LEDs all have roughly the same brightness. And I'm fairly sure they aren't using a constant-current source or sourcing tolerance-matched parts.
>>981577
>removing anything on a board
>not just breaking it apart desoldering the legs out with a flathead or pliers
Shiggy
>>979742
Even pspice 9.3 or whatever is decent. I used it to test a VTO oscillator, and it even replicated the behavior of the 555's internal resistors on the CV line.
Multisim 13 or 14 is literally the normal go-to for this shit.
>>978981
Hardware stores sell this grip-coat stuff
http://www.homedepot.com/p/Plasti-Dip-14-5-oz-Black-Rubber-Coating-11603-6/202196703
You could dip the metal end of a screw driver in it, and then cut off the rubber at the driving end.
>>978788
I agree about the prefabbing and the other circuit nonsense. Stripboard is a marvelous thing.
Are there any ICs that emulate potentiometers that take a variable DC input to control the wiper position? The only ones I can find require a digital interface with a microcontroller or some bullshit.
>>982461
You could probably make one with a MOSFET if you keep it in the triode region assuming you don't need an exact resistance.
>>981765
In my younger days I made the mistake of paralleling LEDs. There was a noticeable distribution in brightness because the low Vf ones ate most of the current. But that's the pathological case.
>>982466
> In my younger days I made the mistake of paralleling LEDs. There was a noticeable distribution in brightness because the low Vf ones ate most of the current.
In that situation, even minor variations in Vf will result in substantial changes in current, as the slope of the V-I curve is much steeper than the load line.
I'm not suggesting that variations don't exist, just that the stated range of 1.85V-2.5V significantly overstates what you will find in practice.
>>982461
Potentiometer or variable resistor?
A voltage-controlled potentiometer with one end grounded is an analogue multiplier: the output voltage is the voltage at the ungrounded end multiplied by the control voltage.
If you need a variable resistor, it's rather hard to construct a generic voltage-controlled resistor, which is why you don't get them as ICs.
The usual approach is to use a FET, but the details depend upon the specifics, so you need to design the circuit for the specific application. Mostly, you need to decide upon the operating point so that you can linearise the response around that point.
The digital case is much easier: you have a chain of fixed resistors and each input bit simply switches a particular resistor in or out of the chain.
The easiest way to construct an equivalent analog device is probably to drive a digital resistor from an ADC.
>>982456
thanks, that still looks a lot better.
the while reason the regulators and the heatsink is wired and no on board is that i plan on using good old to-3 packages and a flat heatsink on the side of the cabinet I've yet to build.
I'm working on a 1v/octave design, if i'm sufficiently happy with its tracking range, temp drift and features I might fab a set of 10 or so just to have some nice looking VCO's with my name on it.
I'll probably send the leftovers to other diy synth dudes for free just for bragging rights.
There was a guy on here who built some big LED array panels for parties. What was the LED driver chip used?
I am stocking up on some cheaper/more-interesting parts and thought of getting a LED driver of some kind.
There is lots of articles on the TLC5940 (16 channel PWM LED driver), and I'm pretty sure the guy here said the chip had 16 or 18 channels.
The DIPs are around a buck each from China-land, so buying ten "for no reason" isn't wasting much money.
>>978816
https://www.youtube.com/watch?v=cjFLN6B0wVU
Sometimes it's just better to go for two common capacitor values to build up your total capacitance on the rail (like x2 1000uF instead of a say 10uF + 220uF + 33uF..), else you increase your risk of running into the antiresonance peak.
>>982484
>The easiest way to construct an equivalent analog device is probably to drive a digital resistor from an ADC.
i found info on how to do this, but i want to avoid microcontrollers due to cost and board space constraints. Analog multipliers are even more expensive in my experience and I also need to have all of the terminals floating. I'm basically trying to simulate a multi-gang pot, but at this point it seems like the space- and cost-effective way to do it is to just use a "real" multi-gang pot.
>>982559
you can get a shitty little 8 pin pic with spi and a 10 bit adc for 60 cents, if that's really too much you should reconsider whether you need this functionality.
>>982501
If you haven't heard of MFOS, I'd reccommend looking into his designs for any pointers, that and electronotes if you can get 'em. I'm building MFOS's sound lab MK2 right, and it's been a carnival of disaster, unfortunately. It's not like the circuits don't work in theory, it's just putting them into practice that's a bitch. Do you use matched transistors anywhere, or do you have some kind of one chip design?
>>981586
Fundamentals of Multimedia, ZeNian Li
Introduction to Data Compression, K. Sayood
Multimedia Signals and Systems, Mrinal Kr. Mandal
Color Image Processing and Applications, K.N. Plataniotis
Digital Video Processing, A. Murat Tekalp
The 1st one was the textbook we used in my multimedia course, the rest were additional references. Most of them are online.
>>983038
not him but i had a class on laplaces but it never once covered any practical applications, nor were they in the syllabus, so i just stopped going to it
what do you actually do with them in particular?
>>983157
Uhh it's everywhere dude. Ever do AC circuit analysis?
>>983167
can you give me an actual answer please? if you don't want to then that's alright too, you can just not reply
>>983157
The Laplace, Fourier and Z transforms are all convolution transforms, i.e. they have the property that the product of the transforms of two functions is the transform of the convolution of the functions:
T(f*g) = T(f).T(g)
where * is convolution and . is multiplication.
A consequence of this is that the transform of the derivative or integral of a function is an algebraic expression involving the transform of the function itself. So applying the transform to a differential equation yields an algebraic equation. This makes them the standard procedure for solving non-trivial ordinary differential equations (differential equations with one independent variable).
The Fourier transform is used for periodic functions, the Z transform for discrete (sampled-time) functions, the Laplace transform for continuous-time functions. They're effectively all slightly different formulations of the same basic transform.
>>983157
you got a lot of mathy answers and i hate math so ill give you one too
look up "control systems", that shit is ALL laplace transforms.
imagine a firing computer - lots of numeric inputs, lots of numeric outputs, and hundreds of relationships that are used to map the inputs to the outputs. laplace transforms let you take those relationships and turn them into clean diagrams that are easy to analyze
so my senior design project is to design a discrete opamp with 10MHz bandwidth and ~60dBV gain
i was thinking of taking an integrated LM747 schematic and converting it to discrete, but i found that 747s only have about 1MHz bandwidth.
how can i increase the bandwidth of this design? there's already a differential amplifier and a feedback network in the design and those are the only things i was taught that could increase the bandwidth of an opamp
>>983220
Why would you start from an IC design? Current mirrors are cheaper than resistors in an IC, but with discretes you're not going to get the matching that you would on a die.
>>983220
> there's already a differential amplifier and a feedback network in the design and those are the only things i was taught that could increase the bandwidth of an opamp
These increase the closed-loop bandwidth by making the closed-loop gain less dependent upon the open-loop gain.
Bandwidth is normally defined as the frequency at which the gain drops to 3dB below the peak gain. By lowering the peak gain, they lower the "3dB below peak" level and thus increase the frequency at which this occurs.
>>983237
> Current mirrors are cheaper than resistors in an IC
Using a current mirror in the collectors of a long-tailed pair increases the single-ended gain compared to using resistors.
>>983237
because i frankly failed the class that teaches you how to do multistage/feedback amplifiers so i wanted to outsource as much of the thinking as possible
right now im googling audiophile discrete opamp schematics but i doubt im gonna get 10mhz bandwidth out of any of them
>>983245
It's just gm R bro. Current source does give you a more stable bias though.
>>983248
The only things in that schematic which will affect the frequency response are the 30pF capacitor and the transistors. Choose transistors with a high transition frequency.
Also: an IC has lots of parasitic capacitance due to the small distances involved.
It shouldn't take long to feed that schematic into SPICE and see what kind of bandwidth you're getting.
OTOH:
> because i frankly failed the class that teaches you how to do multistage/feedback amplifiers
Well, designing an op-amp wasn't exactly the best project for you ...
>>983264
the cuck who assigns these stupid projects assigns them based off the ~opposite~ of what your discipline is
>go to uni's library
>find ancient copy of art of electronics
>last check out date was 1990
nice
>>983349
Do these books actually help anyone or are they just a placebo? I've never got any practical knowledge out of a textbook, and very little theoretical knowledge relative to the invested time.
I'd there a set of basic electronic components that I can buy? I am getting more into electronics, and I want done big boy starter components to give me something to work with.
>>983381
search "_____ kit" on ebay for resistors, inductors, capacitors (ceramic and electrolytic), and transistors. you can get a big bundle for $10-15, if you don't mind the long delivery time.
other things you're usually better off buying on a case by case basis.
>>983385
the jellybeans aren't for building circuits on their own with, they're for just being able to throw in different resistors and capacitors when you're experimenting with an actual circuit.
also you can get 100 decent through-hole pushbuttons for $1, might as well add those to the list. who doesn't want a big box full of hundreds of bags of dumb bullshit you got for (almost) free?
>>983386
>hundred decent through hole push buttons for $1.00
Source?
>>983428
China, obviously
I have a female XLR to male 3.5mm TRS, for a microphone.
Having a look at the xlr side of it, pin 2 goes to a single wire. And pins 1,3, and the casing all go to a 2nd wire. Does that sound right? Will that just send a mono signal down?
>>983459
Here's how it connects through to the TRS side
So I have an RC circuit hooked up to a comparator, with a resistor divider on the other input, to have the comparator raise its output to vcc for a set period of time when power is applied, then stop after the time is reached. This is just to virtually press a button for a couple seconds when power is applied. It works great but not if the power is cycled quickly since the capacitor needs to discharge. Any idea to remedy this?
>>983476
Diode from capacitor to VCC is the usual solution. This assumes that VCC itself drops quickly and that the power drain is high enough to discharge the capacitor.
>>983459
>for a microphone
>Will that just send a mono signal down?
>microphone
Yes?
>>982916
Much appreciated
Hey fellas, I am building the circuit at the bottom of pic related for my amp. Will the stf3nk80z need a heat sink? The B+ is 405 volts. My calculations have it running at 122C @30C ambient, (spec says 150C is limit) but I could be wrong it has been a while since I took analog ICs and I am not accustomed to using many sand components. Thanks!
>>983576
Also I am using the to-220fp variety which has a Θjc of 5 C/W
Where can I get a Gunn diode for use as an oscillator in the 30-40GHz frequency range. I can't find any on typical websites like digikey or mouser.
Also links on information on how to build and work with EHF RF oscillators would be appreciated. I can find very little info on how to actually use a Gunn diode as an oscillator, I though I read something about needing a cavity resonator but I don't know the specifics.
>>983437
So no source then. Your basically just talking out of your ass.
>>983585
You generally order exotic stuff like that directly from manufacturer or from their sales representative.
Do you have any instruments or experience on working at that frequency range or anywhere near to it?
>>983605
>Do you have any instruments or experience on working at that frequency range or anywhere near to it?
Unfortunately no. I'm trying to get experience with it but I seem to be running into walls left and right. As far as instruments I probably don't have anything suited to the task either. I have two oscilloscopes, one analog 50MHz and one 100MHz 200MSa/s MSO but I figure I'd probably need a pretty high end spectrum analyzer to work with signals in that frequency range and that kind of equipment is way out of my budget. Even old spectrum analyzers on ebay can go for up to $2000 and they usually only go to 10GHz, maybe 20.
>>983609
So you bolt the diode on some resonator based on a guess, fed via a bias filter based on a guess and matched to an antenna or whatever based on a guess. Then you give it some power and try to guess if it works - at all.
Dude, start from something bit easier. At the very minimum, start from something you can actually measure.
>>983590
suck my cock http://www.ebay.com/itm/100Pcs-Tactile-Push-Button-Switch-Tact-Switch-6X6X5mm-4-pin-DIP-/141926240182?hash=item210b768fb6:g:~9wAAOSwxH1T2z3e
>>983621
::sucks anons cock::
>>981564
>WHY IS IT SO HARD TO REPLACE CAPACITORS ON A FUCKING MOTHERBOARD
this is an old question but i just noticed it today.
removing the old cap isnt too hard, just heat and pull out a bit at a time, alternating the pins. the real big problem is cleaning the holes for the new cap. the only trick i found effective was to use a sewing needle (stainless of course). heat the solder in the hole, stick the needle into it, let it solidify, and twist off the needle with small pliers, leaving a clean hole.
there are dedicated desoldering needles you can buy, using a different principle, but i've never tried 'em. this jewtube video looks legit: https://www.youtube.com/watch?v=Lw7XSJ-JSIM
>>983381
>I'd there a set of basic electronic components that I can buy? I am getting more into electronics, and I want done big boy starter components to give me something to work with.
it depends on what sources are good for you / what country you are in.
In the USA, it is very cheap to get this stuff from China. What I did was go on aliexpress and search for different kinds of parts, or assortments of parts. it shows how many sales there were of an item and that is a pretty good indicator of how useful it is. The shipping can be slow, but thats why you stock up in advance.
There is also Amazon, but a lot of China people lists stuff on there too now. And they charge more due to the Amazon fees, but the shipping is still slow. So if I have to deal with slow China shipping, then I want the lower China-direct price.
Everything isn't cheaper: for a component that has only one or two non-China manufacturers, the China price may be higher than elsewhere.
It does feel like I'm buying a lot of stuff I'll likely never use, but oh well.
For a USAfag, the stuff from China is so much cheaper its still a better deal even if you only use 10% of it.
You can get bags of 100-ct common parts for $1.25 w/free shipping, where US places would charge $.75 ~ $1 each for them--PLUS a shipping charge on top of that.
ePacket is kinda jacking over US companies, but oh well. I can't fix that (pic related)
https://www.washingtonpost.com/news/storyline/wp/2014/09/12/the-postal-service-is-losing-millions-a-year-to-help-you-buy-cheap-stuff-from-china/
$2 arduinos! party on dude
I just started an EE graduation course, I never before tried to play with electronics and my last semester teacher that was suposed to introduce me to it was a shitter. Now in the current class/lab my teacher makes me think that all of electronics rely on opamps. Do you really use them that much?
>>983673
nope. electronics was once all fundamental components (transistors, resistors, caps), then collections of components integrated into chips used as building blocks, then more complex chips that perform a full function, like a radio receiver chip. op-amps are in that second category, so are in decline, but, just like the fundamental components, will always be needed.
basically OP-AMP and transistors are popular in academia coz there's lots of math involved, which is the primary way to separate those worthy of a diploma, and future taxi drivers or presidential candidates.
>>977908
I'm studying electronics, recently built several successful effect pedals for bass (including designing and making the PCBs) but I'm still not feeling very confident when it comes to guitar electronics. I failed to repair an old bass for someone recently and felt really shitty about that, I somehow got a really nasty hum on it and couldn't get rid of it no matter what i tried....
Anyway, my question is, why do people do blend potentiometers like pic related and not in a more simple way? With a standard linear poti, both pickups connected to one end of the poti, with the middle going to volume/tone. Shouldn't the added resistance drop the voltage/reduce current and therefore make one of the pickups more quiet than the other?
>>983707
Basically, would this work?
>>983650
Or you could be an actual dude and get old electronics scrap effectively for free and rip parts from there.
What is the purpose of having a 1n4004 with the anode to the ground like in this schematic?
>>983673
> Do you really use them that much?
I suspect that a lot of practical circuits will replace the op-amp with a transistor or two.
An ideal op-amp is a perfect amplifier: infinite gain, infinite input impedance, zero output impedance. Real op-amps come quite close to that ideal, which makes it very easy to design analogue signal-processing circuits which actually work without getting bogged down in mundane details like transistor gain curves, biasing, etc.
Once you understand the basic concepts behind the circuit, you can think about the required tolerances, which are often loose enough that you replace the op-amp with something simpler.
Real circuits do actually use op-amps, but they don't use them for absolutely everything (I'm assuming that the examples you're seeing don't contain any discrete transistors, just op-amps and passives).
>>983726
to protect 100uF C from an inverse polarization
>>983694
>basically OP-AMP and transistors are popular in academia coz there's lots of math involved,
Uhm.. The math isn't that complicated, I haven't been introduced to it's inner working, just the way it works and what we can do it to it.
>>983736
I just wished I could see how to do the same thing with "discrete" components. Altough the teacher seems to think that using a op amp makes thing easier, I hate the black box aproach and it fucks with my learning methods. And the very simple circuits we do in op-amps look like dogshit in the protoboard and the tight wiring makes it prone to error. I just feel I could learn much more if I used discrete components.
>>983717
yes, it would work but not as well. with the standard circuit, you can silence an input by turning the pot to the other extremity. with your design, you couldnt go below, lets say, 20% or whatever. depends on the pot values.
>>983726
it doesnt JUST protect the cap against reverse polarity. it protects the entire circuit, most notably the OP-AMP itself which would likely be killed if you left the battery in backwards for more than an instant. as is, the diode will fizzle and burn protecting its fellow components for a while before it dies and is no longer able to protect them.
>>983773
Okay, that makes sense. And I suppose you'd lose quite a bit of volume if you have both pickups active (blend in the middle position) since you have to use a linear poti for this. With the standard design it is logarithmic and reverse logarithmic, meaning the resistance is actually very small in the center position. I might just test it out on an old bass and see if it makes sense.
Also, could someone point me in the right direction when it comes to tuner circuits? I assume I'll use a schmitt trigger to make the signal more easy to read and then simply have a small µC count it, but maybe there's a smarter solution. I'll make a bass self-tuning for a project later this year, so we'll need a tuner circuit and probably a µC to get the information and address small motors for the tuning machines.
>>983793
i thought musician types had special ears that allowed them to tune their instruments by ear. you're blowing the mystique with your talk of instruments and auto-tuners.
>>983819
Tuning by ear is a simple skill that you learn early on (literally just listening for beating), everyone can do that. But you can't tune by ear if you're on stage and an auto-tuner is simply a neat thing to have.
>>982707
my first VCO was MFOS.
but I'm getting into eurorack, so his pcb formats wont fit my rack.
i REALLY want to build his subcommander, shame about his cancer, cant buy from there any more.
this is my diy vco
>>978785
>>978786
its kinda of crude, but tracks fine over 5 octaves and I think its a good basis to build more advanced units.(more waveforms etc.)
http://schmitzbits.de/vco4069.html
http://www.sdiy.org/knas/ars2/4069vco.xhtml
>>982707
i forgot to mention i also bought his make: book.
the MFOS guy , that is.
I'm a bit underwhelmed honestly.
it explains the basics of common modules and has sliced up schems from his noisetoaster.
not bad, but its not really what i needed.
>>983725
That's a big amount of work just to save like 5 bucks.
Maybe if you want the soldering exercise.
>>983938
thats right.
dive and harvest for specialty parts, cool switches and transformers(cheap, but expencive to ship)
but diving for regular parts? nah.
im getting into synth and sound stuff, and the filmcaps are flying out of the cabinet at this point.
i MIGHT harvest some if i find it convenient. its nice to have a desoldering gun,
>>983935
Does he talk about the VC filters in that book? Filters is where things start to get a little hairy. I think the next step is combing through Jurgen Haible's designs and trying to trace through those schematics. At least, that's what I'm on right now.
>>983947
he does, but again its a simple explanation in one part of the book, then a specific article about the VCF in his noisetoaster.
better just read on the web tbqh.
the chem for his noisetoaster is on his page anyway.
the op-amp chapter seems very good thought, i need to brush up on that..
>>978171
janitors can't delete threads only posts.
>>983576
Buddies pls
>>983745
An opamp looks like this on the inside, which is why students tend to start with the black box approach. Trying to solve this sort of circuit for every op amp would just be an enormous headache for everyone. Although discrete op-amps (1 op amp per chip) are basically only used in analog and high frequency applications, any basic integrated circuit will have thousands of amplifiers packed into an area the size of your pinky nail.
If you start designing circuits you'll quickly learn that a bottom up approach is basically impossible, you'll always want to start by specifying more general functions you want your circuit to do, then going down into the specifics that will satisfy the conditions you need
>>983694
op amps and transistors are popular in academia because they are how all digital electronics function, don't be retarded
>>983725
>Or you could be an actual dude and get old electronics scrap effectively for free and rip parts from there.
The practicality of that is very debatable.
Most consumer electronics now are a few VLSI chips and a smattering of tiny SMD diodes, resistors and caps; there's not a huge variety of useful discrete parts to find in most items.
>>983576
>122C @30C ambient, (spec says 150C is limit
That 150C is the absolute maximum value. If your calculations are correct, you're running it really hot. And if it's inside a tube amp, that 30C assumption for ambient temp sounds rather optimistic. Based on that info alone, heatsink sounds like a good idea.
Is your calculation for power dissipation right is another matter. A quick guess about the drain current of LND150 suggests bit less than 1mA in rest, leaving 100V across the big fet, which then results in around 300mW quiescent power dissipation. If you don't use any heat sink, the relevant thermal resistance is Rthj-a, which is 62C/W for the isolated TO220. That would suggest a temperature rise of around 20C. The power dissipation will be different when the oscillator is running, but not so much that a heatsink is needed.
Or so I think. You'll see what really happens when you build it. If the transistor runs really hot (burns your fingers), use a heatsink.
Btw, your tube amp schematic looks suspicious. Are you sure it's drawn correctly? For example, the speaker negative side isn't grounded, which pretty much defeats the feedback loop.
>>984020
>>984020
Thank you for the response, I have many of the necessary transistors so I guess build and see is the way I will play it. The schematic I posted is one I cribbed from an ax84 forum post. My amp is already built and working. I would surmise that the person who created this schematic didn't explicitly show the speaker ground because they were using an uninsulated speaker jack, as is somewhat common, which grounds to chassis. On a guitar amp building forum I imagine this is simply taken for granted, and they were mostly trying to exhibit the nifty solid state oscillator that works at tube voltages that they had created.
Thanks again.
>>984011
You'd be surprised of the wide range of age in things you find lying around in dumpsters
>>984139
>as is somewhat common, which grounds to chassis
when i built my tubeamp i read merlins(valvewizard) grounding chapter in his book.
maybe he overstated it but he claimed stargrounding was really important. so that's what i did.
is your speakers return path through the chassis?
the speakers carries by far the highest current, you get no noise or feedback?
i also got my design form ax84.
randy fay's daisycutter.
>>984453
> maybe he overstated it but he claimed stargrounding was really important. so that's what i did.
From an audio perspective, ideally there shouldn't be any loops in the ground network. If there are loops, their area should be minimised.
From a safety perspective, any mains-powered device which isn't double-insulated should have its chassis connected to earth (safety ground).
If you need to connect two mains-powered devices together via a signal cable, ground loops should be broken by isolating the signal connection (typically using an audio transformer), never by breaking the earth connection on the mains supply. You cannot rely upon signal leads (particularly ones which can be unplugged) to provide earthing.
/ohm/ I'm thig fagit:>>983745
I had a filter laboratory today in which there was no previous explanation except for two paragraphs then just "make filters". I didn't studied calc lll yet so I never saw fourrier. My question is, is there a book on general electronics? Not something on academic scope but somewhere that explains at least something and how to do shit that werks.
>>984509
the Art of Electronics, 3rd edition, or 2nd edition is the 3rd is difficult to find on torrent sites. it's very comprehensive, and not too theoretical, so it's just right.
>>984514
Many thanks anon, I'm a autist can learns by taking a big book and reading that shit then experimenting. Learned C in the same way in a couple weeks.
>>984515
That learns** Jesus fucking phones
>>984519
second edition is just fine for most things. when they have lists of components and their characteristics, like page 501 of the 3rd edition, then the older edition may have some obsolete parts, and be missing newer better parts.
>>984453
I did the hoffman/marshall bus style grounding. Mine is a rebuilt BF Twin and it is quieter now than it was before. All pots and the speaker jacks ground on the chassis.
There seems to be a lot of disagreement in guitar amp building about the absolute best grounding scheme. Probably because most of these guys are guitarists with tech skills and not EEs a lot of the time. Hoffman says star grounding is gay and sucks. I went with his scheme because it is easy and I was eager to get playing again (my amp had been destroyed in a flood).
>>984523
Thank you anon for your answers.
>>984503
i never said my chassis wasn't earthed.
it is earthed where the starground bolt is connected.
this amp can dial crazy gain so i wanted to be extra dilligent.
i prototyped it on a wooded panel before mounting it in a chassis and it was so noisy you wouldn't believe.
Are source impedance and input impedance the same thing?
>>984829
No. Source impedance is output impedance.
Can we also talk microcontrollers here?
I want to make an X (say, 64) channel digital recorder.
What I've sketched out thus far looks like pic related.
Clock ticks away at a specified sample rate setting each register at the same moment.
This then triggers the controller to walk through every register,
enabling it onto the bus,
then pushing it to a computer via USB.
It hopefully can do this before the next clock tick.
Is this even a good way to go about it?
Lot's of shit today instead does SPI,
which I just don't know how to think about - I like my buses.
What sort of controllers should I be looking at?
Do they even make these components/ICs like I want anymore?
>>984961
You will have a hard time carrying 64 channels of ANYTHING over usb or spi, unless your sample rates are low enough to make this worthless for audio and RF applications. For DC measurements (e.g. a logic analyzer or short probe) it might be ok.
Really you should be looking into DSPs rather than generalized microcontrollers/microprocessors and an interface like Eth-RJ45 (or big parallel port, or firewire 800). I can guarantee the built-in SPI or USB capabilities on an arduino-tier uC will be an order of magnitude too slow for you.
>>984961
The first thing to consider is that microcontrollers typically only have a single ADC. If they have e.g. 16 analogue inputs, they're just multiplexing the single ADC between them. So if you sample all 16 inputs, the samples will be skewed (each input will be sampled at a slightly different time). The same is usually true for discrete multi-channel ADCs.
So if you want a 64-channel recorder and need synchronous sampling, you're going to need 64 ADC chips. If you don't care about synchronous sampling, you can get 8-channel and 16-channel ADCs, which cuts the number of ADC chips to 8 or 4.
The main issue with that many channels is that the total data rate is likely to exceed the capacity of an SPI bus. If you're talking about audio, 64 16-bit channels at 48kHz equates to ~50Mbit/sec of raw data, and framing is going to add a lot of overhead to that.
So you either need a controller with multiple SPI interfaces, or multiple controllers which simply compress the data (collating multiple samples into a single burst transmission to reduce the framing overhead), or you need to find an ADC with a parallel interface (which will significantly reduce the number of options; no-one likes parallel buses).
