| Electronics > Repair |
| Series defect on agilent 167xx boards? |
| << < (5/50) > >> |
| DocBen:
@MarkL: Great stuff, definitly keep it coming! :popcorn: I'm a little short on time right now, but I did have a look at the Logic Analyzer Software (which you can download and play with even without an Analyzer) and there doesn't seem to be anything quite like the Tools you have. Maybe thats only activated in the main application on the actual Analyzer when the module is in place. Have to test that sometime. I do wish they had linux and cli on this thing though. Also I think the 16902a has two busses: SVY and RIO (and I think they are both named after bus stops in California, Scotts Valley and Rio Dell) SVY seems to be the "old" 16700 type bus and RIO is an high speed addon I guess (something like LVDS with an FB2040 BTL driver) so memory transfer from the cards is likely significantly faster. But it isnt mentioned anywhere. I can only infer from an Bus tuner application for Rio (sets the voltage swing) and pci to svy / rio drivers the application installs. Right now the first thing I want to do is to replace the electrolytic capacitors on the Mainboard and then try to get the 16740a to be a 16752b. |
| DoricLoon:
I wonder if anyone could shine some light on an issue I have with a 16750A card? Purchased 2X16741A and 2x16750A cards from Ebay advertised as not working. Price was pretty low and the cards came with cables. I was hoping I'd be able to repair at least one of them. The two 16741A's are pretty bad. Large number of self test fails and one of them looks like the whole bank of ram has seen high voltage? (They also show many vias with corrision) The 16750A's are a bit more promising, one in particular. When I ran the self test it failed with a comparator error and a calibration error. Couldn't see much sign of the corrosion but took the stiffeners off anyway and had a good look (looked healthy) I then connected a power supply linked to each of the bits of a pod and checked the thresholds based on the service manual. Pod 4 low order bits were considerably out. Since I had plenty of spares I thought I'd steal a comparator out one of the 16741A's. At least I think I have changed the comparator? The signals enter the card and pass a couple of resistors before hitting a 44pin QFP which I assume is the comparator? (Agilent numbering so I can't search for a data sheet to verify) I checked continuity through from pod4 bit 0 to find the correct chip. After changing the said device I still have the same outcome. Buzzed out the traces crossing under the runners without any breaks. I did however find a small section of corrosion on one of the traces about 5mm away from one of the runners. The trace was open under the corrosion. Soldered a wire bridge across it and touched up the soldering on a resistor network which looked a bit ropey. This time when I carry out the self test the card passes with flying colours. Unfortunatley the thresholds on pod4 low order are still out of tolerance. The error with the threshold set to -5 and 5V is around 700mV (-4.3 and 4.3V) the error reduces linearly heading toward zero. With the threshold set to zero the error is only 35mV (within the specified 65mV). I have tested all the resitors and capacitors in the vaccinity of the comparitor. Have cross referenced with Pod 2 as well as another card and values all seem very close. Does anyone know which devices are responsible for setting the threshold voltages? Service manual states: "The threshold circuit includes a precision octal DAC and precision op amp drivers. Each of the eight channels of the DAC is individually programmable which allows the user to set the thresholds of the individual pods. The 16 data channels and the clock/data channel of each pod are all set to the same threshold voltage." If I could identify the dac and opamp it would be a start? The card is probably useable enough as it is. Still 56 channels working perfectly and thanks to Mark L, and the removal of one resistor and I have 32M memory depth. :-+ |
| MarkL:
--- Quote from: DoricLoon on October 01, 2019, 08:44:43 pm ---... Does anyone know which devices are responsible for setting the threshold voltages? ... --- End quote --- I haven't done any debugging on the comparator or threshold circuits, but I would start with identifying the DAC (U39) outputs for the 8 pods. Then compare the out-of-spec output with some of the working ones all the way to the comparators (U29 U38 U46 U53 U72 U79 U85 U88). The DAC (U39) is an AD7841AS in a MQFP-44 and is near the backplane connector. When you change the threshold voltage in the "Format" tab, the voltages on the DAC should change immediately, so it should be fairly easy to compare what's happening. With a quick continuity check, it looks like the DAC output enters the comparator on pin 39, after going through a resistor divider of 7500R on top and 536R on the bottom. Make sure the arriving DAC output and the output of the voltage divider into the comparator make sense. Compare with the circuitry on a good pod. I don't know what's going on inside the comparator, but keep in mind the the incoming voltages from the tip of the probe pods is arranged as a 10x probe (90k + 10k), so I would expect the threshold input of the comparator to also be in this range (just a guess - I didn't verify it). It's difficult to probe these boards in the chassis, and you can do some checking with continuity. But for real verification, you'll need to solder in some jumpers and run them out of the chassis to probe the voltages I'm suggesting while it's running. If the bad pod comparator is on the bottom, you can also move the card to the bottom chassis slot, take the bottom off the chassis, and relocate the mouse/keyboard interface board (just let it hang in the air but insulate it so it doesn't short against anything). That clears the way to probe the bottom of the board. Please let us know what you discover! |
| DoricLoon:
--- Quote from: MarkL on October 02, 2019, 01:48:48 pm --- I haven't done any debugging on the comparator or threshold circuits, but I would start with identifying the DAC (U39) outputs for the 8 pods. Then compare the out-of-spec output with some of the working ones all the way to the comparators (U29 U38 U46 U53 U72 U79 U85 U88). The DAC (U39) is an AD7841AS in a MQFP-44 and is near the backplane connector. --- End quote --- I did notice an Analog devices chip near the backplane last night but dismissed it thinking it would be nearer the comparators. Thanks for the info. I have downloaded the datasheet for the DAC. If the reference was off then more than one 8 bit section would be affected. There are three reference connections available on the DAC but minimum of two outputs using the same ref. I guess the reference will common to all of them? So carry out continuity checks between the DAC outputs and the resistor divider you mention. Check resistor values and connections to comparators. Thank you Mark, will get a look at it in the evening and will update. |
| MarkL:
--- Quote from: DoricLoon on October 02, 2019, 02:30:49 pm ---... If the reference was off then more than one 8 bit section would be affected. There are three reference connections available on the DAC but minimum of two outputs using the same ref. I guess the reference will common to all of them? --- End quote --- Yes, the DAC reference inputs are all tied together, so your assumption is correct. --- Quote ---So carry out continuity checks between the DAC outputs and the resistor divider you mention. Check resistor values and connections to comparators. --- End quote --- Right. It could be a partially corroded trace coming from the DAC that is causing a high impedance connection to the comparator. So it works, kind-of. Of course it could also be the DAC itself. Or it could also be the comparator, but you've changed that. It's time to measure the actual voltages and compare with a known good half-pod. Here's the mapping for the DAC outputs, which you might find useful: DAC (pin) Comparator Pod/Bits (Clock) ---------- ---------- ---------------- VoutA (2) U85 1/7:0 VoutB (44) U88 1/15:8 (J) VoutC (43) U46 2/7:0 VoutD (41) U53 2/15:8 (K) VoutE (37) U72 3/7:0 VoutF (35) U79 3/15:8 (L) VoutG (34) U29 4/7:0 VoutH (32) U38 4/15:8 (M) Since you're seeing problems with pod 4/7:0, I would focus your search on the path from DAC output VoutG, through the resistor divider, to U29. Photo below of the comparator area. They're all the same as far as the threshold input is concerned. I'm not sure why they assigned two DAC channels per pod since the software will only set the entire pod to the same voltage. It might be they wanted to have the option to split the pod into two different voltages in the future, or maybe there's some small differences in the comparators which are calibrated out during boot. Dunno. |
| Navigation |
| Message Index |
| Next page |
| Previous page |