Re: Figuring out a serial port

From: Justin <shadow_at_darksideresearch.com>
Date: Mon, 9 Sep 2013 08:29:15 -0400
Message-Id: <B81C36CC-E162-4889-B53F-0D2973040590@darksideresearch.com>
I eventually figured it out with a combination of what you describe and guessing.

The item in question was an original Valentine Research G-Analyst, so I guessed that the baud rate for the serial port would be 8,N,1 with no flow control (particularly since I had no desire to figure out the flow control pins up front).

I cracked open one of the housings for it and discovered that the RJ45 connector it uses for its control unit had some thin and heavy traces that it shared with the 8 pin DIN connector that the serial data cable would connect to.  I checked the heavy traces and determined that one of them was hot 12V and one was ground (apparently at one point they planned to provide 12V power for accessories that were released via this DIN port).  So, I eliminated those as possible TX and RX lines and had 5 left to work with.  I could only see four traces, but I could not see the backside of the board because I was concerned that removing the transducers would mess up their calibration.  So at that point I started to brute force the pins and discovered that pins 7 and 1 were receive and transmit.  The others have voltage on them but it sends a data stream without them connected so Ive left that alone.

The only weird thing Ive run into is that if I connect the shield braid to the shell of the DB9 at the end, the thing stops sending data.  This was very confusing when I first trimmed off the extra wires and permanently soldered the TX/RX pins.

On Sep 9, 2013, at 5:23 AM, davee.roberts@fsmail.net wrote:

> I agree with SD. Tracing the hardware circuit out from the connector is probably going to be your best bet.
> 
> This will tell you which pins are not connected and (more importantly) which ones are connected to ground or power. You should also be 
> able to find this out with a multimeter set to a low continuity range. You can then avoid any power pins!
> 
> If you are lucky, the hardware may have a UART chip inside which will mean that you can track down the data sheet and work out the circuit 
> both forward from the UART and backwards from the connector. If not, you will just have to 'wing it' and trace out the circuitry the hard 
> way.
> 
> Be warned, the serial levels may not be RS232. They may be logic level, 20 mA current loop or (even older) 60 mA current loop. Connecting 
> directly to an RS232 port could spell disaster if the equipment doesn't actually have this type of input/output stage.
> 
> Any chance of posting a photograph of the insides? This may elicit some further suggestions.
> 
> Dave
> 
> 
>> Message Received: Sep 09 2013, 03:25 AM
>> From: silverdr@wfmh.org.pl
>> To: "cbm-hackers@musoftware.de" <cbm-hackers@musoftware.de>
>> Cc: 
>> Subject: Re: Figuring out a serial port 
>> 
>> 
>> 
>> -- 
>> Sent from mobile phone (so please have understanding).
>> 
>> On 8 wrz 2013, at 21:44, Justin <shadow@darksideresearch.com> wrote:
>> 
>>> Is there a trick for figuring out which pin is which on a serial port that can be executed with a multimeter and an ancient no-logging 
> Heathkit oscilloscope?
>> 
>> If this is your only testing equipment then it may be in fact tricky.. 
>> 
>> First, have you checked the levels? Is it really an RS232 with its notorious logic levels? Or something else? This you should be able to 
> check with your scope. 
>> 
>> Next I'd of course connect a logic analyser with or w/o level converters, depending on the outcome of the previous test. But since I 
> understand that you have none, the "get one" suggestion is quite appropriate IMHO :-)
>> 
>> How about looking closer at the hardware itself then? I still remember very well the lesson Ruud once gave me by deducting virtually 
> everything (including error in the silkscreen!) about an unknown to him piece of hardware by merely looking at an (unpopulated!!) PCB of 
> some seven ICs of various sizes. After this experience I know that a lot of information can be gained this way. At least you should be 
> able to eliminate some pins from the equations. 
>> 
>> 
>> -- 
>> SD!
>>       Message was sent through the cbm-hackers mailing list
>> 
> 
>       Message was sent through the cbm-hackers mailing list



       Message was sent through the cbm-hackers mailing list

Received on 2013-09-09 13:00:08

Archive generated by hypermail 2.2.0.