| Products > Test Equipment |
| Agilent 16717A Comparator and ZoomChipSelTest failures |
| << < (3/6) > >> |
| MarkL:
A minor update... At some point during my probing, I killed one or both of my U69 and U70. I didn't notice the board had started failing at some point since the PC driving the test was happily sitting in a loop on a different bench. I'm sure it was the result of a slipped probe shorting out two pins somewhere. I also noticed that U69.Test_Clk_CmpOut_zoom.png was not showing a nice sharp edge for the comparator output (pin 7). It's a super fast comparator, and this would not do for the signal speeds this board is capable of. Oh well, I've done worse. Both comparators replaced. Maybe only one of them was dead, but whatever it was it affected all of them. No doubt it was interfering with one of the test signals common to all. Below is a re-capture of the Test/Vref/CmpOut pins, U69.Test_CmpOut_Vref_fixed.png. Nice sharp edges now with a fair amount of ringing on the rising edge. Note the swing is now between 1.7V and 2.5V. With the dead comparator and previous screen capture it was not at these levels. U69.Test_CmpOut_Vref_start.png shows that the CmpOut duty cycle does not quite track the 50% test clock. This is because Vref has just entered the trip point for the comparator. As Vref gets closer to 0V, the output approaches 50%, as it should, as shown in U69.Test_CmpOut_Vref_after_20us.png. This is more info than you need, but since I had to troubleshoot my board, I thought I'd post it for future reference in case anyone else lands on this thread looking for LA comparator info. Also attached is a shot of the empty footprint for a comparator. They are all seem to be the same, at least on the bottom. In my 1NB4-5036 pinout description, I wasn't sure where pin 19 goes. Under the chip I now see it goes into a via, but I can't find the other end of the via anywhere on the board. It's still an unknown. So... Be careful probing, and use the little slotted protector caps that keep the probe tip on one leg at a time. |
| MarkL:
We seem to be done with this thread, but I have a couple of minor updates before I forget. Below is my updated notes on the 1NB4-5036 comparator pins after I could see the traces underneath. It also occurred to me that the comparator pins I couldn't trace are probably not a N/C, but are likely going to one of the large BGA ASICs on an inner layer. I also forgot to post my analysis of what's inside the 16-pin yellow resistor packs that's next to the comparators. Attached. Just puttin' the info out there. Reverse engineering always involves some amount of guessing. Corrections welcome. 1NB4-5036 comparator pins based on use in 16717A module --------- 1 CmpIn 2 -5.1V (decoupled to GND) 3 CmpIn 4 -3.2V (decoupled to GND) 5 CmpIn (on U69 is clock in from pod 1, input voltage/10) 6 GND 7 CmpOut from pin 5 8 nCmpOut from pin 5 9 +3.4V 10 CmpOut 11 +3.4V (decoupled to GND) --------- 12 CmpOut 13 +3.4V 14 CmpOut 15 +3.4V (decoupled to GND) 16 CmpOut 17 +3.4V (decoupled to GND) 18 Low when test, goes 3.4V to ground, slow rise (probably a pullup) common to all comps 19 via under chip, +2.1V steady not common to other comps can't find other end, comes from big ASIC? 20 +3.4V (decoupled to GND) 21 CmpOut 22 +3.4V (decoupled to GND) --------- 23 CmpOut 24 +3.4V 25 CmpOut 26 +3.4V (decoupled to GND) 27 CmpOut 28 GND 29 GND 30 -3.2V (decoupled to GND) 31 CmpIn 32 -5.1V (decoupled to GND) 33 CmpIn --------- 34 +3.4V (decoupled to GND) 35 CmpIn 36 GND 37 50Mhz clock input, +/-5mV square wave during test NOT common to all comps source clock split 4 ways with cap/res to get 5mV source clock common between left 4 and right 4 can't find source end, each side comes from big ASIC? 38 GND 39 Vref 40 +3.4V (decoupled to GND) 41 CmpIn 42 GND 43 CmpIn 44 +3.4V (decoupled to GND) --------- -5.1V must be power +3.4V probably power, is also Vdd on a couple of LVC08 on the top, U77 U78 -3.2V could be power or bias Vref is 136mV when set to TTL trig level +87mV when set to +1.0V -11mV when set to 0.0V -110mV when set to -1.0V |
| fisafisa:
hi guys, did you solve the problem? Recently I worked on a similar fault and solved it. The boards were of the 16740 family. The problem turned out to be a resistor of 7.5K that failed open in some channels and in others had a larger value. This issue seems it was common to all comparators on the top of the board. Some failed during testing. I ended up replacing all 4 on both boards. The resistor was used in a voltage partitor to generate the reference voltage in pin 39. When correct, in circuit the resistor measures 7.05k. The last I changed was measuring >8K and was on a pod were I had failures associated to the zoomChip. The pod itself seemed to work fine. After the change all the test passed again. F |
| MarkL:
--- Quote from: fisafisa on May 06, 2021, 09:43:44 am ---hi guys, did you solve the problem? --- End quote --- I think Keith will have to answer that. He sent me a couple of cards for further investigation and repair, but we concluded the above debug info was not from one of them. --- Quote ---Recently I worked on a similar fault and solved it. ... --- End quote --- There was another person who had a bad 7k5 resistor at the comparator input, same as your board: https://www.eevblog.com/forum/repair/series-defect-on-agilent-167xx-boards/msg2719080/#msg2719080 I've found a lot of bad resistors on 165xx/167xx series boards (I'd say on the order of 15 on 20 boards), and in all parts of the circuitry. Some resistors have been off, but most completely open. And a few have been associated with setting the output of a voltage regulator, which was particularly bad. I have a bunch of other test equipment from HP/Agilent from this same era, and I've never seen so many bad resistors. The bad resistors are usually, but not always, near an area of corrosion. I can't say the corrosion caused it, but it is a suspicious coincidence. |
| MarkL:
Here is an update to the 1NB4-5036 pin usage posted above. There are a couple of changes: - Pin 29 is an input (not a GND) - The test clock input is 50mV (not 5mV) Didn't want to leave bad info out there. Thank you to user shakalnokturn. 1NB4-5036 comparator pins 44-pin LQFP, 0.8mm pitch based on use in 16717A module --------- 1 CmpIn 2 -5.1V (decoupled to GND) 3 CmpIn 4 -3.2V (decoupled to GND) 5 CmpIn (on U69 is clock in from pod 1, input voltage/10) 6 GND 7 CmpOut from pin 5 8 nCmpOut from pin 5 9 +3.4V 10 CmpOut 11 +3.4V (decoupled to GND) --------- 12 CmpOut 13 +3.4V 14 CmpOut 15 +3.4V (decoupled to GND) 16 CmpOut 17 +3.4V (decoupled to GND) 18 Low when test, goes 3.4V to ground, slow rise (probably a pullup) common to all comps 19 via under chip, +2.1V steady not common to other comps can't find other end, comes from big ASIC? 20 +3.4V (decoupled to GND) 21 CmpOut 22 +3.4V (decoupled to GND) --------- 23 CmpOut 24 +3.4V 25 CmpOut 26 +3.4V (decoupled to GND) 27 CmpOut 28 GND 29 CmpIn (previously said GND - wrong!) 30 -3.2V (decoupled to GND) 31 CmpIn 32 -5.1V (decoupled to GND) 33 CmpIn --------- 34 +3.4V (decoupled to GND) 35 CmpIn 36 GND 37 50Mhz clock input, +/-50mV (approx) square wave during test NOT common to all comps source clock split 4 ways with cap/res to get 50mV source clock common between left 4 and right 4 can't find source end, each side comes from big ASIC? 38 GND 39 Vref 40 +3.4V (decoupled to GND) 41 CmpIn 42 GND 43 CmpIn 44 +3.4V (decoupled to GND) --------- -5.1V must be power +3.4V probably power, is also Vdd on a couple of LVC08 on the top, U77 U78 -3.2V could be power or bias Vref is 136mV when set to TTL trig level +87mV when set to +1.0V -11mV when set to 0.0V -110mV when set to -1.0V |
| Navigation |
| Message Index |
| Next page |
| Previous page |