Batboard

I've got a Panasonic Toughbook CF-18 that's never had a problem programming all kinds of radios until recently. Lately, I haven't been able to read any radio via the serial port... It will usually start reading (put the radio into program mode, ect.) but usually fails within a few seconds. I've tried multiple makes/models of radios, and different programming cables.

As a test, I connected this PC's serial port to a desktop's serial port via a null modem cable, and ran Hyperterminal on both. At pretty much any baud rate, I can type or hold a key down and watch the correct characters display on the other machine. Once I tried a few file transfers, things got interesting.

At every baud rate I tested, I could send a file using the Zmodem protocol from the Toughbook to the desktop without issue. Going from the desktop to the toughbook at 2400 or 9600 bps, things worked ok, too. Going from the desktop to the toughbook at 19200 bps or faster sometimes works, but numerous errors/retrys occur.

The Toughbook runs Windows XP, SP 2. The serial port settings are all at their defaults, which is FIFO buffers enabled, and maximum TX and RX buffer sizes. I tried setting the buffer sizes to minimum, and still get tons of errors/retrys on a file transfer from the desktop to the Toughbook.

I don't think I've added any software or changed anything on the Toughbook since the serial port last worked...

Does this sound like a hardware or software problem? I wonder if the memory used for the receive buffer on the toughbook might be flakey?

Memory *could* be the issue, but I'd look to the level translators first. Every serial port has a set of level translators that take the lower voltage (either TTL 5V or now-a-days more like CMOS 3.3V) to the RS-232 +/-12V levels. It may be the failing port has had its level translator damaged - or it may even be something as simple as the charge-pump capacitors on the device have dried out. That would cause the port to be ok-ish for low speeds, but at higher speeds, where more charge per unit time is needed due to the lines being wiggled faster, the caps aren't storing enough juice, and the voltages sag.

You might look at the signals on a scope, and see if they are sagging.

I played around with them a little more this morning, and made in interesting observation. I think the problem is pretty much independant of baud rate, as I was seeing it even at 300 baud.

When I send several 1's, 2,'s and 3's to the Toughbook (1111111111222222222222222222233333333333333...), I see the following on the Toughbook's Hyperterminal screen: 1111111111222222222222222222233|3|3|3|...

Looking at an ASCII table, it turns out that the binary value for '3' is 00110011, while the binary value for the pipe '|' is 10110011. So somehow that MSB is being toggled after the second occurance of the third uniqe character. It happens with any other characters too. The difference is always that MSB.

The problem does appear to be somewhat intermittent, so I can't always make it happen. But I have seen it with both fast and slow baud rates.

I guess I should try a reformat/re-install of Windows to rule out any kind of software problem... This PC is like 5 years old, so I doubt it's still within warranty. I'd imagine if I sent it in for repair, they'd probably want to replace the whole expensive motherboard, which probably isn't worth it.

Maybe I should just get a PCMCIA serial port card for the thing, and mark the built in serial port as 'BAD'.

You can contact Heartland with your model & serial #'s, they can tell you where your at warranty wise. I would certainly try a reload of the image disc that came with the unit first.

Well, I just found a PCMCIA serial port card and installed it, and it has the same problem too! I guess a reformat and Windows re-install is in order...

I been noticing some of the same problems with my cf-29. I notice if my battery isn't fully charged it starts to act up. When I plugged the power plug in it went away.

I finally got a chance to wipe and re-install XP, and the same problem is still there!

It is plugged in most of the time. All of the above tests were done with it plugged in to power. I think the battery is in pretty good shape.

I didn't think to test it right after the XP install, and instead connected it to the internet to install SP 3 and all the other Windows updates first. But even before, this laptop didn't get connected to the internet very often (it's mostly used for programming radios) and I don't think that Automatic Updates were enabled, so I don't think it was a recent Windows update that caused it. I can't think of anything that changed on it software wise right before the serial port started acting up...

So now I still don't know if this is a software or hardware problem. It still exists after a Windows re-install, and exists on both the built-in serial port and a PCMCIA serial port.

I just wiped and re-installed XP one more time, this time not installing SP3 or any other Windows updates, and the problem is still there. I did notice the problem going from the Toughbook to the desktop PC, so it looks like it can occur both ways.

I guess it has to be some kind of hardware problem... I'm trying to think what piece of hardware it could be--it's not specific to the serial port, as the PCMCIA serial port has the same problem as the built in one. Everything else on the Toughbook seems to work fine.