1. Home
  2. Electronics Design
  3. Rotten USB to RS232 adaptors.

Rotten USB to RS232 adaptors.

formatting link
Likely same thing wit hsame conbstruction here:
formatting link

and available from any more places

So I have this cable, it drives my LED disco light show PIC, from Linux. Bought it because the chip in it supports Linux... (prolific chip).

So, anyways, that PIC it drives also has timers, and it wakes me up with lights, sets the color of the lights in the evening etc.. But yesterday morning nothing happened. So first I suspected my PIC. Did a code verify: OK. Cold start everything worked, but stopped working after some time... temperature? checked no. Now I leave that USB to RS232 cable connected all the time.... Finally reasoned the PIC must go into input mode, and expect some data... like setting timers, but I did not send a command for it to go into input mode. So late last night had a look at the Linux USB interface if anything happened to be sending data to that serial port (lots of scripts running here). No. Scoped the input to the PIC, looked OK. Programmed an other PIC, after some hours same effect. But I was now sure it was waiting for data, must have received some command. From WHERE? I did not send any, USB did not see any? Scoped again at the PIC RS232 input..Bingo! RF!!!!!! Now that RF comes out of that USB to serial converter. _sometimes_. So tapped it, now no RF but cracking voltages... Aha, defective or lose filter cap in the charge pump, I could make it go completely open by pressing that adaptor. Now that was late last night, so decided to open it up (guarantee? well I dunno, I need that cable NOW), so this morning took the dremel, no way this thing seems to have a way to open it normally, but I could see a filter cap through the dark semi-transparent plastic but... when I was more then 4 mm deep in that adapter and still no end to the plastic, it dawned on me: It was potted. But not potted with solid stuff, potted with a soft flexible plastic that bends every time you insert the thing (it is a D connector with board inside), and anytime you press it or hold it basically. So... no way I can repair it... .. sigh The moral: NEVER pot your stuff, especially if it is in the form of a connector, and ESPECIALLY not with flexible plastic so that if somebody grabs it it puts stress on the PCB inside.

Reply to
Jan Panteltje
Loading thread data ...

lights,

temperature? checked no.

to be sending

completely open

dunno,

seems to have a way

semi-transparent plastic

plastic, it dawned on me: It was potted.

bends every time you insert the

hold it basically.

connector, and ESPECIALLY not with flexible

They do that to save cost. It's not purposely potted but simply a molded-on "clump". Much cheaper than making shells and snapping them on. You'll see the same thing on VGA cables, the connectors are firmly embedded in black plastic, no shells or shrouds or anything.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
Reply to
Joerg

On a sunny day (Fri, 27 Mar 2009 09:51:41 -0700) it happened Joerg wrote in :

Hi, I ordered 3 other ones from Germany, much cheaper place, good references too:

formatting link
Probably the same molded design, but this looks a bit more solid. Amazing prices! Now I have 3 for the price of one here.

Reply to
Jan Panteltje

formatting link

Looks at the bright side: You get better prices on good Gouda cheese and Beerenburger :-)

If it's any comfort the type I have is made in Taiwan and also molded. At that price there just isn't any other way.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
Reply to
Joerg

Jan,

formatting link

Those have an FTDI chip inside as well which IMO is somewhat better than the Prolific chips.

---Joel

Reply to
Joel Koltner

formatting link

480Mbps too - amazing what RS232 can go up to nowadays... :) 
--

John Devereux
Reply to
John Devereux

On a sunny day (Fri, 27 Mar 2009 10:38:00 -0700) it happened "Joel Koltner" wrote in :

formatting link

My research shows that it is also the Prolific chip:

formatting link
If you click on the 'driver' icon, bottom right on that page, and download the driver, then the zip file extracts as a driver for Linux, win98, Vista, and Mac. I looked in the Linux driver source, and it is for the Prologic pl2303 chip. This driver has been in the kernel for a long time now, (they talk about Rat Head 7.3). The fact that it is Linux supported was why I bought these, I did the resaerch first :-) It could have a different chip though, but the FTDI chips are also supported by Linux. When the other one worked, it worked very well, so think Prolific is OK.

Reply to
Jan Panteltje

Hi Jan,

Hmm, interesting. I have a few that look absolutely identical, and they're FTDI... I suppose that over time the OEM might have switched suppliers (or perhaps there's more than one OEM making identical-looking products).

So don't you find it annoying that it's up to you to write your own UDEV rules if you want to nail down a particular USBRS-232 adapter (via the adapter's USB serial number) to a particular device? (E.g., /dev/usbtty0 or /dev/modem or whatever.) I like all the power that UDEV rules give you, but the average non-technical person who tries to use more than one such adapter under is probably going to be quite frustrated when, by just unplugging and replugging adapters (and even just by rebooting at times) their /dev assignments keep changing!

Windows keeps a database of USB serial numbers and will keep assigning the same COM port to the same adapter, which strikes me as a reasonable thing to do. It'd be nice if Linux had that option available...

---Joel

Reply to
Joel Koltner

Yeah, Windows does that part of its job nicely. It's not just those devices, many USB-connected accessories are built the same way inside. There is a FTDI or similar chip and it shows up as a COM port connection. The end user sees nothing other than a regular USB cable. Usually I have several hanging on a laptop, all via a $5-$10 USB splitter from Walmart or somewhere.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
Reply to
Joerg

On a sunny day (Fri, 27 Mar 2009 14:03:55 -0700) it happened "Joel Koltner" wrote in :

Yes, possible, there are hundreds of similar adaptors around.

Yes, well, I have both real serial ports ttyS0 and ttyS1 in use, for PIC projects, so they are fixed, and this one was on ttyUSB0... The other one is for the laptop, and one spare, so nothing gets confused.

I use my own very nice and simple communication program:

formatting link
ptlrc -d /dev/ttyUSB0 (or whatever device) does it.

The only problem you have, is if you disconnect USB while one of the programs that use it is still running, then it will register itself as the next USB device /dev/ttyUSB1 when you plug it in again. I know this... I have not bothered with udev rules ...

Yes, Xp assigns COM14 to that USB adapter all the time... Why 14? I dunno.

Reply to
Jan Panteltje

rules

adapter's

projects,

Because that's the day Manchester United won against Wigan Athletic? If not happy:

-> Device Manager -> Ports (COM&LPT) -> pick from list -> double click

-> Port Settings -> Advanced -> Assign other COM port.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
Reply to
Joerg

Well, this is the downside of Window's database of USB IDs... you've probably had upwards of 14 different RS-232 to USB adapters over time, so there are reservations for all the ones you've ever used, just waiting to fire up COM8: or whatever if you just plug that long-since-forgotten adapter in. :-)

You can see all the "phantom" devices by firing up a command line, and executing:

set devmgr_show_nonpresent_devices=1 (the magic) set devmgr_show_details=1 (not actually needed, but interesting) devmgmt.msc (run device manager)

...and then selecting View->Show Hidden Devices.

---Joel

Reply to
Joel Koltner

Interesting and thanks. Works fine. What a LIST!?!? I have. I feel like deleting the entire lot of them just to start fresh.

Jon

Reply to
Jon Kirwan

supported

rules

adapter's

/dev/modem

average

replugging

projects,

Now I need to find out why COM2 through COM9 are already "used."

Jon

Reply to
Jon Kirwan

In principle it does, in a way, through the UDEV "persistent rules" feature. I don't think anybody has written a persistence rule to handle UDEV serial devices, but it ought not to be difficult...

... EXCEPT that many of the inexpensive USB/RS-232 adapters (especially those build with the PL2303 chip) have no unique identifiers. With the FTDI-based adapters I think you get the serial number as a part of the basic chip, but many of the Prolific chip variants require the use of an external EEPROM, which the OEM is permitted to omit. Such devices show us as generic adapters, and you can't tell one from the other.

Not a problem if you use only one per system at a time, I suppose, but if you have two or more attached, it's a pain, and I don't think there's an easy way around it. Could do it according to which USB bus/port it's plugged into, I suppose.

--
Dave Platt                                    AE6EO
Friends of Jade Warrior home page:  http://www.radagast.org/jade-warrior
  I do _not_ wish to receive unsolicited commercial email, and I will
     boycott any company which has the gall to send me such ads!
Reply to
Dave Platt

That's what I did when I was experimenting with my PIC USB device. If no serial number is assigned, a new COM port will be enumerated each time you plug into a different hub. I removed all USB COM ports. Some of them were from adapters I tried long ago. I have a very old machine (ca 2000).

Now I use the same serial number in all my devices, so once it is enumerated on a machine, it won't care if a different device is plugged in. There could be a problem if someone plugged two of these into one computer, however.

My device now includes a USB hub using the TI TUSB2046B, and my COM port is a PIC18F2450.

Paul

Reply to
Paul E. Schoen

Hi Dave,

Yeah... in that case Windows has the same problem as Linux, of course.

I guess the manufacturers figure it's not worth an extra $0.25 for the relatively small percentage (

Reply to
Joel Koltner

It appears that Windows assigns a pseudo-serial number to devices without serial numbers by using a number which identifies which hub and port it is plugged into. So you always get the same COMnn number if you always have your adapter plugged into a certain hub port, but a different COMnn for each of the other hub ports. So there should be no problem with multiple adapters connected simultaneously. I would expect an error, however, if two exactly identical devices were plugged into two ports at the same time.

Paul

Reply to
Paul E. Schoen

formatting link

Head 7.3).

first :-)

by Linux.

The SiLabs chips are also supported by linux.

Reply to
Jasen Betts

What about those cheap 4:1 USB splitters? I use them here in the lab and it still all works fine after power cycling. (Windows XP).

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
Reply to
Joerg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.