FYI,

with the VLB card set to COM3/4 it is kinda expected that COM1/2 will already be present in the machine.
If COM1/2 are not present, then COM3/4 will become COM1/2 but on the addresses for COM3/4 which will complicate troubleshooting a little bit.

My advice, leave it at COM1/2.

Given that your ISA IO card works, I'm expecting the voltages to be ok, but go ahead anyway and verify to be sure.

I'm wondering if the serial port connector on the VLB IO card might have been replaced by an incompatible one. Such things can happen.
This link https://www.vogonswiki.com/index.php/Serial_port gives some information about the incompatibility.

Looking at the photos I would say you have a 'type A' where each pin# on the board goes to the same pin# on the connector.
You can easily test this using a multimeter in continuity or resistance mode.

Kinda wish the serial port wiki had a layout of the serial port pinout & idc connector pinout. Maybe someone with access can add that.
the IDC, pinheader on the board, has the odds on one row, and the evens on the other, the serial port has 12345 on one row and 6789 on the other.
Anyway, the pins are labeled on your VLB card so it's easily verified.

If you can get SDT to work, and you can measure voltages - toggling DTR and RTS should cause the respective signals on the DB connector to toggle
+12 & -12v (technically RS-232 specs call for a range of -(3<>15) & +(3<>15) but PC cards almost always use -12 & +12 as those voltages are available
directly from the PC power supply and are on the card slots)

I've seen some very cheap equipment use 0v as one of the levels, but that's outside the spec, and not reliable
- shouldn't be an issue on the PC as valid supplies are available.

More common is non-powered peripherals which to TX back to the PC require some of the PC outputs to be available at -12 and +12...
This can be the "driver" writing DTR and RTS to different states, or relying on no data being transmitted from the PC which makes TXD
inactive while DTR and RTS are active. Again, I doubt this is a factor as you havn't changed software, but worth checking that a valid signal
can come back.

Dave ::: https://dunfield.themindfactory.com ::: "Daves Old Computers"->Personal

weedeewee wrote on 2024-12-23, 08:39:

FYI, […]
Show full quote

FYI,

with the VLB card set to COM3/4 it is kinda expected that COM1/2 will already be present in the machine.
If COM1/2 are not present, then COM3/4 will become COM1/2 but on the addresses for COM3/4 which will complicate troubleshooting a little bit.

My advice, leave it at COM1/2.

Given that your ISA IO card works, I'm expecting the voltages to be ok, but go ahead anyway and verify to be sure.

I'm wondering if the serial port connector on the VLB IO card might have been replaced by an incompatible one. Such things can happen.
This link https://www.vogonswiki.com/index.php/Serial_port gives some information about the incompatibility.

Looking at the photos I would say you have a 'type A' where each pin# on the board goes to the same pin# on the connector.
You can easily test this using a multimeter in continuity or resistance mode.

Kinda wish the serial port wiki had a layout of the serial port pinout & idc connector pinout. Maybe someone with access can add that.
the IDC, pinheader on the board, has the odds on one row, and the evens on the other, the serial port has 12345 on one row and 6789 on the other.
Anyway, the pins are labeled on your VLB card so it's easily verified.

This is actually interesting, I got 'type A' here. Let me try and test this with SDT from Dave (thanks, Dave for your instructions)

I'm testing right now and I've got some interesting results.

@Dave: The reason why I've asked for detailed instructions (again, thanks for that) about SDT is that by using the VLB COMs nothing appeared on the screen when typing, even with the loopback connected. On the other hand, COM from Prime2C controller gives characters back when using SDT and loopback.

Here are the voltages on the loopback:

COM1 (Prime2, working):

PIN: VOLTAGE

1 (DCD) : 0 VDC
2 (RX) + 3 (TX): -9.8 VDC
4 (DTR) + 6 (DSR): 0V (even when toggling in SDT)
5 (GND): 0 VDC
7 (RTS) + 8 (CTS) : -9.8VDC or +10VDC (toggling with SDT)
9 (RI): 0 VDC

COM3 (VLB, non working):

PIN: VOLTAGE

1 (DCD) : 2.16 VDC
2 (RX) + 3 (TX): 2.12 VDC
4 (DTR) + 6 (DSR): 0V (even when toggling in SDT)
5 (GND): 0 VDC
7 (RTS) + 8 (CTS) : +0.65 VDC or + 3.8 VDC (toggling with SDT, and yes, no negative voltage here at all)
9 (RI): 0.56 VDC

I've checked and the -12VDC and +12VDC voltages are present at the VLB ISA slot. -12VDC goes straight into pin #32 of the UM82C865F chip (the one on the right). Dead end.

I'm just wondering how this card is actually working on IDE/FDD right now.

For further reference, when idle in DOS here are the COMs voltages from both I/O cards:

COM1 (Prime2, working):

PIN: VOLTAGE

1 (DCD) : 0.08 VDC
2 (RX): 0.08 VDC
3 (TX): -11.25 VDC
4 (DTR): 11.45 VDC
5 (GND): 0 VDC
6 (DSR): 0.08 VDC
7 (RTS): 11.45 VDC
8 (CTS): 0.08 VDC
9 (RI): 0.08 VDC

COM3 (VLB, non working):

PIN: VOLTAGE

1 (DCD) : 3.35 VDC
2 (RX): 3.34 VDC
3 (TX): 3.19 VDC
4 (DTR): 2.73 VDC
5 (GND): 0 VDC
6 (DSR): 2.44 VDC
7 (RTS): 2.73 VDC
8 (CTS): 2.44 VDC
9 (RI): 2.44 VDC

There's definitely something wrong on this card. Seems to me the IC is providing some sort of switched capacitor voltage converter of some sort, and that's fried internally. Caps definitely look good, and as well anything on the board.

Any thoughts?

I've not worked on SDT for years! - but this discussion gave me an idea for
a slight enhancement.

While it is true that the way to verify TD/RD (data) with a loopback is
to send/receive data - It doesn't work with a meter measuring voltage.
The bits go by way to fast to see ... I always used a little scopemeter
if I wanted to physically see the bits - but not everyone has a scope!

So.. I've just updated SDT with the ability to toggle "sendBREAK". When ON,
it forces the TX data line permanently asserted, just like you would do when
testing DTR&RTS - unfortunately the PC uart has no easy way to see this via
a loopback (you will basically get a null character with a framing error),
but it's just as easy/obvious to see via voltage change on the TD output pin!

Again, easiest way to get the updated version is to download DDLINK
from my site (in fact, that's the only download I've updated so far!)

Dave ::: https://dunfield.themindfactory.com ::: "Daves Old Computers"->Personal

UMC chip has blown voltage level translators.

>6 (DSR): 2.44 VDC

You could maybe bodge it to work by adding external MAX3232 and 3V voltage regulator. MAX3232 powered by 3V will recognize logic high from 2V up. That would fix output, but we dont know if receive is working, to test it you would need to connect another PC with crossover Serial cable.

rasz_pl wrote on 2024-12-24, 06:11:

UMC chip has blown voltage level translators.

>6 (DSR): 2.44 VDC

You could maybe bodge it to work by adding external MAX3232 and 3V voltage regulator. MAX3232 powered by 3V will recognize logic high from 2V up. That would fix output, but we dont know if receive is working, to test it you would need to connect another PC with crossover Serial cable.

I'll just assume everything is blown on that side and keep on using the other 16bit ISA Prime2C card for COMs and (if for whatever reason I'll need in the future) LPT.

Thanks for the help so far 😀