PCGU1000 Died. Resurrected. Some info sharing

Hello all,

It just refused to start up. Went though a bit of troubleshooting (gate arrays make that kinda difficult and found good power most places, but no ±12VDC from the voltage multipliers. No “pump” signals from the Xilinx GA. Both oscillators were running, meaning that the smaller GA was at least partially alive. However, every other output from that chip was static. I was about to appeal for help, but it occurred to me that the small GA (IC8) looked like a key player in getting everything else downstream to wake up. So as a desperation move, and because the outputs didn’t look right (according to my intuition, that is) I thought “What the heck, I’ll reseat this thing.” Those DIP sockets don’t have the best reliability on the planet, for sure. Mirable Dictu. Sent it a command and it’s now working. The “pump” signals look nice and healthy. Voltages everywhere are good, and the red LED is on.

So take this FWIW. This could happen to you, and as fixes go, it’s a pretty easy one.

And BTW, those DIP sockets have always been a nightmare.

Oh yes, and one other thing: A HUGE Thank-You to Velleman for supplying the schematics for these products. Without it, I’d have faced either an expensive repair, replacement, or shopping for something else.

Looks like I spoke too soon.

It WAS working when open on the bench, but after reassembling it into the case, it’s dead again. Symptoms as above. It looks to me like there’s a problem with IC8 being flaky. The strange thing is that the PCGU1000 is talking to the driver and user interface. When it wasn’t working before, the program would report that there was no communication with the device. Now, it starts up normally and appears to program a waveform correctly. Only thing is, the generator isn’t turning on now.

So Velleman… can I get a replacement IC8 somehow? It looks like the problem is there, but without a lot more information on what’s in those GAs, it’s hard to say.

When it worked on the bench a few minutes ago, the “pump” signals looked good, and the 12V supplies were functioning. Now there’s no pump and no 12V. But I’m getting no commands out of any of the pins of IC8 in response to USB inputs. That’s why it’s looking like the problem is there. I tried re-seating it again, but it didn’t help. This kind of looks like a bad package with a broken bond wire in a critical location.

More information: I’m very suspicious of IC8. Here’s what I see on the pins when a command comes over the USB:

1 Normally high, ~1/2 sec of data activity
2 Normally low, ~1/2 sec of data activity
3 Low
4 Low
5 (gnd)
6 Low
7 Normally high, ratty looking data from opto
8 Normally low, clean data from IC8 to opto
9 High
10 Low
11 Low
12 Low, but occasionally there’s a 1/2usec train of positive pulses, but not for all command inputs
13 Low
14 Vcc 3.3VDC
15, 16 10MHz oscillator – always running
17 Low
18 Low

The behavior I saw at pin 12 seems pretty strange. During a brief period of time (10-15 sec) any command sent from the GUI would cause activity at this pin. Then it stopped. I’ve played with it a bit more and haven’t been able to see the pulse again. Unfortunately, I didn’t notice whether the red LED illuminated when this pulse was visible.

With this information can someone who knows how this is supposed to work make a diagnosis? And if you agree that the problem is in IC8 as it appears, how can I get a replacement?

Well well…

I got it working again. The key was the ratty signal at IC8 pin 7, from the opto. It was really ratty looking and didn’t seem to be pulling up as high as it should have. So I put a 3.92K RN50C metal film resistor across R5 and waalaa… red light comes on as soon as GUI loads on the next power-up. The signal is better, but still ratty looking. It must have been right at the threshold of IC8 not seeing it.

So… the learning here is:

  1. Those optos may degrade. I tried swapping them and the generator wasn’t recognized at all, leading me to suspect that they are MARGINAL at best. Think I’m going to buy a handful of new ones.
  2. The 2.2K R5 pullup is too high a value for the 3.3VDC supply.

So, if your PCGU1000 has mysteriously gone Tango Uniform, this is a Really Good Place to start looking for the problem. Now before I declare victory and go home, I’m going to let this operate for a while to see if it keeps working. I’ll post back with any updates.

Hope this episode can help someone else.

Thanks for sharing this info.

You’re welcome.

I should add a couple of details: 1) The resistor in parallel with R5 is very easy to add on the back of the board, as long as it’s 1/4W size or less. The one I added is an RN50C 1/8watt, so it has very thin leads which makes the rework fairly easy. 2) The value I ended up with is 2.71K which appeared to reduce the rattiness of the waveform by pulling it up to a level closer to the 3.3V supply. The resulting current in the low state didn’t affect the low voltage level at room temperature. 3) I don’t have a warm fuzzy feeling for the quality and performance of the 6N137 isolators on the board. I can’t think of any good reason why the waveform on the open collector output should be as corrupted as it is. I’m going to shop around for some new ones. If anyone has recommendations for known-good optoisolators, I’d be most appreciative.

And questions for Velleman: My PCGU1000 is from 2007, s/n 675 I believe. Are there any running changes since then that I ought to consider incorporating? Is there an updated schematic? The schematic I have is from the original support CD. In the intervening 6 years, there must have been a few changes. Can I get a copy of the latest version of the schematic?

Finally: regarding the value of R5 – I think the design value of 2.2K is too high under any circumstances. You should probably re-evaluate this choice and consider incorporating a lower value to improve instrument reliability.

[quote]And questions for Velleman: My PCGU1000 is from 2007, s/n 675 I believe. Are there any running changes since then that I ought to consider incorporating? Is there an updated schematic?[/quote]Yes. Modified the connection of R3 and added a LED in series with it.

[quote]The schematic I have is from the original support CD. In the intervening 6 years, there must have been a few changes. Can I get a copy of the latest version of the schematic?[/quote]Here is the link to download the latest document: box.net/shared/chq5q2c85o

Please see this thread how to “update” your PCGU1000: viewtopic.php?f=11&t=2514