Why are serial ports dying /g/?
Nowadays it's hard to find mobos that even have a serial header let alone a db-9 port on the back
usb adapters are also not up to par.
What has the world come to?
>>54812507
Shuttle makes a few fanless celeron models with 2 x DB9. Good enough for DOS-based radio programming at a reasonable price with new hardware.
>>54812507
Capitalism dictates that things should be done to maximise profit and minimise loss.
Serial ports cost money and don't bring much profit due to consumer demand.
>>54812507
USB adapters are definitely up to par. I don't know what you're talking about. Serial ports are still used extremely often in the network administration.
gee let's wonder why
>not hotswappable
>transfer speed in bytes / s
>pins can easily bend
>1002 old proprietary serialshit devices that never had an adapter made for anything else
>IRQs, 2016
>screwing in the shitty port
>What has the world come to?
Something better, while you yearn for the past, luddite.
>>54812708
inb4 REEEEEEE at pic
>>54812786
Are they seriously still using mini USB ports?
>>54812708
USB adapters are not up to par because they can't be. ever.
USB uses a polling based method to detect whether a device has something to say or not, vs serial's superior interrupt based method.
Also, serial is used anywhere where actual shit is being done, besides muh games.
Examples besides networking would be hardware development/debugging, which are impossible to do with USB adapters because of the above statement.
>>54812907
>USB uses a polling based method to detect whether a device has something to say or not, vs serial's superior interrupt based method.
Who cares? It's 2016. Hell, let's pretend it's 2001. Your hardware is more than powerful enough to poll at a rate higher than any interrupt frequency you would want for any real-world application.
>Also, serial is used anywhere where actual shit is being done, besides muh games.
Uh. USB is used virtually everywhere where actual shit is being done.
I don't know what kind of weird, faux-nostalgic bullshit you've filled your head with, but forget it.
>Examples besides networking would be hardware development/debugging, which are impossible to do with USB adapters because of the above statement.
Uh. Explain how the following is possible:
1) I've prototyped systems using Arduino and TI's MSP430 boards, which connect over USB.
2) I worked at a hardware company when we transitioned from RS232 to USB (and later Bluetooth). How did we do that?
>>54812786
>not hotswappable
They can be, through software.
>transfer speed in bytes / s
You can easily slap a 100 megahurtz clock to your chip and have as much as speed you want. usual speeds are low because the standard never got updated
>pins can easily bend
literally what? do you put your dick in them?
>IRQs, 2016
>Current year
>spinlocking
>1002 old proprietary serialshit devices that never had an adapter made for anything else
>serial
>proprietary
>screwing in the shitty port
Don't screw the port if you hate stability then
>>54812952
Arduino and similar microcontroller dev boards use an FTDI USB to serial converter on the board.
Most, if not all, microcontroller architectures such as MCS-51, AVR and 68k are natively Serial.
>>54812507
You can still find these ports on custom motherboards such as those used in workstations, which also tend to have a PCI slot for legacy hardware.
>>54812954
>the standard never got updated
Yes, it did. We call it USB now.
>>54812991
>Arduino and similar microcontroller dev boards use an FTDI USB to serial converter on the board.
...yeah. Which kind of demonstrates the point, eh?
>>54813040
It also demonstrates the point that product designers have to build in USB support, which incurs additional cost and potentially causes issues, for no reason other than "because who has a serial port".
See also: Recent counterfeit FTDI fiasco.
>>54813062
>which incurs additional cost and potentially causes issues
Uh. I *really* think you're too young to remember RS-232.
"Causes issues" my ass. USB Plug and Play is a fucking godsend. I'm tired of these faux-nostalgia posts by /g/tarded Millennials.
>>54812507
They're not dying, I can bet they'll be around for another good 10 or 20 years. But yes it's true that they're dying from consumer computers because of lack of demand. On workstations and business grade machines they're still well and good.
Tbh their data transfer speeds are way too slow for any data intensive applications, the only plus point about them are their ease of application in hardware (one IC does all comm), hence making them good for embedded applications, and the fact that they just send raw data and you get to interpret it however you want, so no need for drivers either
>>54813097
nobody is arguing against plugging in your shitty mouse into the computer and expect it to work retard. use USB for that.
This is about rs232 ports being eliminated even though they are used everywhere.
>>54813097
>he doesn't know what FTDIgate was
Keep being a consumer whore normie, then.
That, and >>>/v/
>>54812507
>foxconn
man, time to fap
>I have more serial cables lying around than VGA cables
Actually pisses me off when I need to find a VGA cable
>>54812952
>Uh. USB is used virtually everywhere where actual shit is being done.
I work at a datacenter, serial connections are still used for connecting to networking and other equipment when a direct connection is needed.
>>54813097
I used serial devices growing up and they were basically plug and play, you just needed to tell the software which serial port to use and you only needed to to that once unless you changed what port you used. To tell the truth I think we should bring that back for devices to keep complete retards from using computers and shitting up the internet further.
>>54813028
/thread
>>54812507
My Asrock X99 WS-E has 2 headers but then it is for workstation use.
>>54813062
>that product designers have to build in USB support
>incurs additional cost and potentially causes issues
You do know that goes with any interface, do you? I'd rather use an FTDI than a MAX232.
Also USB support is trivial these days, outside of the maker-tier MCU market at least.
It virtually costs nothing in hardware when you chose to use a controller with an USB PHY peripheral. Firmware development is easy with a fitting USB stack library. There are no excuses in professional product development nowadays to use either RS232 or the slightly less awful FT232.
>for no reason other than "because who has a serial port".
Here's a few hints on why USB is better:
* more powerful than RS232
* smaller size and footprint
* plug and play
* is a comprehensive standard for many different solutions which just works
>>54812786
My friend, you misunderstand.
>>54812507
Low board space cost for embedded USB devices.
As a hardware designer myself, I thank a lot to have that little space because devices are more portable and thinner.
For the same reason, they can even be used to charge a portable device, and nowadays you may expect that everyone has a microUSB spare cable to charge their devices with. That even spares me the work of designing a special charger myself.
Just buy a decent Sabrent adapter and get over yourself.
-t. Network engineer who's field machine doesn't have a serial port.
>b-b-b-buy a USB adapter u fags
t. all you retards that don't work with CNC machinery
>>54812507
Because they're really only used for board configuration and industrial stuff now.
By the way, there's a ton of industrial boards with lots of COM ports on them. Look at the low-end Atoms for examples, you'll often find them and RS-232 there as well.
>>54814229
You seem to be missing the point.
It is not about what is "better", it is about having it available, so designers and customers alike don't have to think about compatibility.
>>54814571
>My edge case doesn't apply to a general recommendation meant to cover the other 95-99% of use cases
Boohoo. :(
because usb adapters work fine
alternative: get a serial pci card you fucking idiots. the chances of windows 95 running on modern hardware (for your precious cnc machine) is virtually zero in the first place.
and before you have a brain aneurysm, my work has multiple windows 2000 and even windows nt machines that run ICS and equipment like cranes. however they are all virtualized
>>54814776
>abloobloobloo the only people in the universe that use RS-232 say that USB adapters are fucking useless so they can go fuck themselves
Go fuck yourself
>>54814778
>somehow windows 95/200/nt are relevant to serial port
>>54814831
USB adapters will work for 99% of the applications, with most of the problems stemming from using modern hardware and newer OS's which is the fault of the ancient hardware needing it.
>>54814844
they are. anything modern will use a usb serial port just fine, or you just buy a serial card for the motherboard and call it a day.
the only reason this is an issue is if you are using an old as fuck piece of equipment where the software was made for fucking windows 3.1 or something old and wont run on anything new. which means the only time this is an issue is if there is a major hardware failure such as motherboard, in which case you cannot replace it and any modern shit you buy isnt going to be able to run windows 3.1
if you are going to try and look smart, at least have some sort of real world experience to back it up with
I currently need to figure out some solution for the laser engraver I have it uses an heavily modified version of embedded windows xp and the hardware is slowly dying. I need to come up with something relatively new that I can use in place of it but I can't find anything close to modern with serial ports.
>>54812507
I understand that you really want to talk about that old thing everyone liked but please choose another topic. We have a new serial bus now, updated and better suited for modern needs, and adapters for comm with old hardware.
>>54812954
>serial apologists
I should have expected this, but I didn't know it was this bad. Stick to IRC, luddite.
I'll just say it once more
>SERIAL APOLOGISTS
>>54812507
>Nowadays it's hard to find mobos that even have a serial header
The vast majority of them do have it, actually.
>>54812786
The great advantage of serial ports is that they are available to software before USB drivers are loaded and initialized, though. It can be pretty hard to debug firmware, bootloaders and/or early kernel booting without a serial port.
>>54812786
>>IRQs, 2016
What are you on about? There's literally no piece of hardware that doesn't use IRQs in 2016.
>>54812507
Because they were superceded by USB two decades ago and adapters exist for the legacy hardware that still uses them
>>54818148
this is true for consumer hardware, not so much for industrial/embedded systems
serial is simple and cheap, and for many tasks, will be fast enough forever
it makes far more sense to add a serial UART to a cheap microcontroller than it does USB for the same purpose
basically, USB is overkill for many tasks. serial isn't going anywhere
>>54818267
so there must still be industrial motherboards being made for use in these systems that have serial ports. Either that, or PCI/PCI-E cards.
>>54818439
i'm not even talking about desktop pc's, anon
>>54818267
>serial was cheap in the 90s and is therefore still acceptable now simply because it was there the whole time
everyone has upgraded. God /g/ is so fucking stuck in the past, we can't even use USB.
>>54818786
like i said, many things will never need more speed than what serial can provide
you're suggesting such things use more complex/expensive hardware/software/firmware for no reason other than "it's newer"
>>54813062
>building in USB support
Why is this a bad thing
>>54818267
>basically, USB is overkill for many tasks
This. Especially DIY projects. It's easy to work with serial, implementing USB is a pain in the ass.
>USB adapters
Universally shit
![DELOCK_82768_01[1].png](https://i.imgur.com/CDrdcMM.png "DELOCK_82768_01[1].png")
>>54812507
literally every new motherboard has still serial port headers you fucking moron.
Just buy pic related for $5.
>>54819438
in the case of small microcontrollers, a usb controller could be a significant portion of the die, it's wasteful if the additional speed isn't needed
USB rightly took over in the consumer space where the additional speed and the convenience of using one port for so many things was beneficial
but in other cases, USB is unnecessarily complex, and only adds additional cost
>>54819528
To support my point here is the cheapest motherboard I could find.
https://www.caseking.de/asrock-c70m1-amd-a50m-mainboard-amd-apu-c-70-mbar-134.html
>>54819699
>>54819528
/thread
>>54812786
>>54812952
>>54813097
>>54814229
>>54814280
>>54814962
>>54817307
>>54818148
>>54818786
>>54819438
Problem: need to transfer 1 byte from client to host.
Serial:
>client pulls tx line down
>host: looks like we have some data coming our way. lets standby for the data to come
>client proceeds to send the byte over. pulls tx line down again at the end
>host: packet received. Hey! OS sempai.. here is a byte sent by client on port n.
>OS sees the byte.
USB:
>client pulls D+ line up to determine it is using full speed.
>host: holy shit. new device.. let's sends every client 4 bytes of sync data so they can arrange their clocks.
>host: SHUT THE FUCK UP EVERY BODY. I'm talking to device address 0 now. I'm about to give you a request. also, here is a crc5 for this 3 byte package.
>client: ...
>other clients: ...
>host: okay here is another 4 bytes of 0 & 1 for every client. so they can arrange their clocks again.
>host: Whoever the fuck was told they are about to receive a request. here goes. it's a 10 byte package asking you to identify yourself. also here is a crc16 for you to check.
>client: here is a 4 byte 0 & 1 so we can sync our data master.
>client: I got your request successfully, host.
>host: Okay you know the drill. 4 bytes of garbage coming down everybody's throat.
>host: every client shut the fuck up. I'm asking client address 0 to send me whatever they were asked to send before. here is a crc5 for you btw.
>client: Hello host. here is 4 bytes of garbage. Also I'm a usb2 device with vid = x, pid = y. I have an interface that will choose their own class. Make sure you don't send me over 8 bytes of data in one go because I can't handle it. the rest of the details will come in the next packet. have fun. here's your complementary crc16.
>host: Let's just everyone sync our clocks again. Also, whoever the fuck sent that last packet, message received.
to be continued.
>>54813097
Yeah, for consumer purposes, there is no place for RS-232. For engineering, there's still a place for RS-232, and there are pretty easy ways to get RS-232 where ever you need it. I really don't see why the tech nerds here are complaining.
>>54820003
>host: too soon? lolno. 4 bytes of sync happening again. I'm talking to device address 0 now. I have something important to announce. here is your crc5 btw.
>host: Let's all just sync again for good measure. I announce that the message you were trying to send me has been received. here is your complementary crc16 to ensure you got my message correctly.
>host: Hey you fucking OS. I have a fucking USB device trying to connect. Should I address it?
>OS: give it a fucking address. 0x0c has a nice ring to it.. give it that.
>host: It's been a while.. here is 4 bytes of data coming to everyone so they can sync their shit and listen to what I have to say.
>host: nevermind everybody. I just want to talk to client address 0. I'm about to give you a request that you must handle. crc5
>host: I'm not even gonna say it because you guessed it right. Whichever one you slaves who were told to listen, here it is. From now on you are address 0x0c. go cry to your mama if you don't like it. Also here's a crc16 to keep you busy at night.
>client: Master, I wanna say something. please accept my sync garbage data. I just wanted to say message received.
>host: Holy molly.. just look at the time. It's syncing time!! Also, client address 0 listen up. everybody else go fuck yourself. I'm about to tell you something important so listen up. crc5 goes here.
>host: 01010101010101010101010101010100. I was just going to say message received. enjoy your stay. here is the crc16
>client: proceeds to send sync shit. Host sama, your message that said you got my message successfully has been received successfully. thank you.
>client proceeds to change address to 0x0c
tbc
>>54820019
Now here is where it gets interesting. If your OS doesn't have any drivers for the device, then there's that and everything stays the same. poor client forever friendzoned by host and not talked to again.
If you are lucky, your OS may decide to load a driver for your device.
>driver: Please proceed to give the device the following command: "You are to load configuration code 1"
>host: SYNC PARTAY!!!! EVERYBODY INVITED. GET YOUR FRESH SYNCS HERE FAST!
>host: Talking to client address 0x0c. you are about to receive a command. be ready. crc5
>host: now, say with me, a 0. a 1. 0 1 0 1.. comon! 0101 0101 now say it fast 0101010101010101 aaand 00. Whoever I was talking to last time, you are to "load configuration code 1" that's all I fucking can tell you right now. <insert crc16 here>
>client: yay let's sync. message received host thank you.
>client loads configuration 1
>host: Nobody ever dies unsynced. I'm about to tell 0x0c something. stand by. as always, we never forget the crc5.
>host: As I was saying. transaction complete successfully. good job pet. now go match this crc16 for me.
>client: (syncs silently). Message received my lord.
>driver: Now that the device is configured. it knows what the fuck it has to do. Just ask him to piss some bytes.
>host: hey my subjects. drop your water and start syncing. Because I'm a generous god. I'm allowing device 0x0c to give me some bytes. go ahead moose. you are clear. this crc5 guarantees it.
>client: Allow me to become one with the host and sync our clocks with each other. Hi. Your byte is as follows. here is a crc16 that I calculated just for you to ensure it doesn't go corrupted. peace.
>host: Let's mash it up a little. KJKJKJKJKJKJKJKJKJKJKJKJKJKJKJKK. message received whoever that was.
>host: hey driver sama. here is the byte your device gave me. cya later
>driver: O fucking S. Give this fucking byte to the program that wants it so i can leave this thread forever.
the end.
My supermicro dual cpu board has both serial and firewire, and a parralel printer port
I don't remember if it has 1 or 2 though
>>54820003
>Let's just everyone sync our clocks again.
I'm 80% sure that's a reference to something
But I don't know what
>>54813097
I'd like to have a motherboard brand that has RS232 because of some old as fuck biotech hardware I bought off eBay which specifically can't connect to a USB serial adapter (for whatever reason). Doesn't really matter if I ever used it or not
>>54820271
>biotech
Neat
Is it anything interesting or just some junk to play with?
>>54820029
Amusing but suffered from the evil dead syndrome.
>>54820003
>>54820019
>>54820029
I would maybe agree with you if the USB to Serial adapters didn't work in literally every case I have ever tried, in the case of every user of a serial cable who I have ever talked to, and seemingly thousands of reviewers on various tech websites. Why else would laptops cease to be manufactured with serial ports?
I get that you read some wikipedia article about how USB works and now you want to flaunt how cool and intelligent you are on /g/ but everything you said is essentially irrelevant, which just goes to show how little you actually know.
You need some common sense. People like you are literally the problem with technology. How about instead of ranting on and on about how shit USB to serial connectors are while having absolutely zero personal experience with them, you go pick up one such adapter for <$10, or borrow a coworker's.
>>54820003
>>54820019
Try >>>/vr/
never going to be part of any business, please stay in the 80s somewhere else
>>54822963
Personal theory on this - he's fallen for the age meme and thinks serial is sooper cool without ever having used it. Probably read some rant that was relevant in 1997 when USB was new and you were working in sub-700mhz pcs. Seen the same shit for typewriters and then you have 25 year olds who have never used one ever insisting they're superior typing devices. If you were actually forced to use a typewriter now, and actually forced to use serial now, you'd hate it instantly. Also bonus points for explaining how different your port is to everyone at the coffee shop, and how it doesn't work with anything anymore, but if it *did*, it would be so superior in the way it did so
>>54823120
Which is why serial connections are normally used to configure routers and switches that haven't been brought online?
>>54820003
>>54820019
>>54820029
While we're on the topic, ACPI is bad because it's smart and exists. It has to "ask" the CPU to not use as much power, which requires some certain cycles, which will lag the whole system by 0.00000001ms. Batteries and laptops were better when there was no such thing and it just ran at 100% the entire time, because the little amount of work it has to do to keep asking is more than the gain. While we're at it, everything smart/good/made past 1985 is bad too because I personally don't understand it
Industrial systems commonly still use them.
Having a true serial port is essential for maintaining, supporting and repairing industrial control units,PLCs, motor inverters, various sensor systems, and profibus (Plus several of it's sibling protocols), as well as tapping into the kinds of networks that were installed in factories 30 years ago and have never needed to be replaced because they do the job they were designed for and do it so well they only get noticed when the £10k an hour process breaks and you have to get it running again asap. Ripping it all out and replacing it with modern kit would literally bankrupt the company due to downtime, missed deadlines and cost of replacing all the machines that will not under any circumstance talk to USB.
USB is for consumers.
>>54823120
>never going to be part of my hipster startup.js
fixed that for you fampai
