MCT Serial board mystery - Can you crack the pinout code?

Tothwolf tothwolf at concentric.net
Sat Nov 15 15:27:06 CST 2014


On Sat, 15 Nov 2014, drlegendre . wrote:
> On Sat, Nov 15, 2014 at 9:10 AM, Tothwolf <tothwolf at concentric.net> wrote:
>
>> Standard numbering scheme? Surely you jest? ;)
>>
>> Do you mean something like this?
>>
>> 2 4 6 8 10
>> . . . . .
>> . . . . .
>> 1 3 5 7 9
>
> No joke, you got it right. That's the standard scheme for IDS (I said 
> DIP, which is the wrong term I think) headers.. across-and-back, as you 
> wrote it.

I've seen pins labeled all sorts of ways, even though that across-and-back 
really does seem to be the correct way to label them.

>> It sounds like you've just about got it figured out though. If you have 
>> the datasheet for the 1458 (aren't there 1459s on the board too?) then 
>> you can follow the signals through the level converters back to the 
>> UART, and its datasheet should then tell you exactly which signals are 
>> which.
>
> Don't see any 1459s on there, just the pair of 1458s up near the IDS I/O 
> headers

As Tony pointed out, I was thinking 1488/1489, so disregard that.

> But yes, you have a point.. I could probably have continued, tracing 
> back from the 1458 inputs, through the various logic gates, to the 1014 
> / 1015 UARTs.  But I stopped where I did as it was already pretty 
> tedious and I figured I had enough info for you in-the-know types to 
> figure it out at a glance.. ;-)

I dunno, it is getting beyond what I can figure out without the board in 
front of me to trace out too.

> That, and I am confused by a couple of my findings, particularly the Pin 
> 4 - Pin 6 relationship. Why are these pins essentially common to each 
> other, and to the XNOR inputs? One pin has a 47R in series to the XNOR 
> gate, the other has a 150R in series to that same gate. It's as if the 
> line with the 150R could be asserted 'high' by some line and then the 
> line with the 47R could re-assert it 'low' - but not all the way to 
> zero, more like to 1/3 of voltage. See, that makes no sense to me,.
>
> Ditto for the Pin 1 / Pin 7 relationship. Both of these pins are 
> essentially outputs from the same op-amp output - one is basically a 
> direct-out from the op-amp #2, the other is via the collector of a PNP 
> (3906) transistor - the base of which is also tied to op-amp output #2. 
> This is the stuff that's really confusing me.. is the 3906 acting as an 
> inverter? If so, why?

Many serial interfaces are differential so you wouldn't necessarily expect 
a voltage to be at reference to GND for a logical 0. For "real" RS232, 
you'd expect a logical low to be below 0 volts, so maybe that's some of 
what they are doing here since they aren't using something like 1488/1489 
transceivers to do the level conversion?

> To be clear, J1 (10-pin I/O header) has two pins (4 & 6) both tied to 
> the selfsame +input+ point, and two pins (1 & 7) tied to the selfsame 
> +output+ point. J2 is identical, but it uses a different op-amp and a 
> different XNOR input. Otherwise, J1 and J2 are carbon-copies of each 
> other, electronically speaking.

It is looking like the pinout isn't going to be 'standard' RS232 though. 
Just pin 5 being NC wouldn't match up with a RS232 pinned DE9 where you 
would expect TXD on pin 3.

>> It makes sense that pin 10 would be an extra ground too, since it is 
>> quite possible they originally used a 9-conductor ribbon cable and 
>> wouldn't have connected that pin. You may also find than the signals 
>> come out on the correct pins of a DE9 connected to a 10-pin IDS 
>> connector with some 9-conductor ribbon cable.
>
> Or to a 25-pin version of the same. Sure wish I had one! But I tossed / 
> gave all that away years ago, I think.. =/

Those likely wouldn't have worked with this board since it doesn't look to 
be purely RS232. Even the generic boards that used the 10 pin IDS to DE9 
or DB25 cables varied on their pinouts so there were multiple pinouts for 
the cables. With a straight through ribbon cable and an IDC press-on type 
DE9 connector, you'd get a pinout such as this:

IDS  DE9  RS232
1    1    DCD
2    6    DSR
3    2    RXD
4    7    RTS
5    3    TXD
6    8    CTS
7    4    DTR
8    9    RI
9    5    GND
10   NC

...but some companies would instead use solder cup DE9 / DB25 connectors 
and would split or fold the ribbon cable around different ways from 1->9 
or 1->5, 9->6, etc for DE9 and the DB25 cables had even more variations.


More information about the cctalk mailing list