Products > Test Equipment
8116A Code 42
MarkL:
Ok, there's a few things we should look at.
The first, as you've pointed out, is U33.6 being close to 0V. Because this pin was changing slightly before, the pulldown for that node might be open (R62.5). The easiest way to check is to use a 470 ohm resistor and hold it to pin 6 and Vee (pin 8 ) at the same time while watching the voltage on pin 6. If you get a normal ECL logic level, R62.5 is open. You can either replace resistor network R62, or just solder in a stand-alone 470R resistor underneath. I would probably do the latter since it's easier and you can move on to other areas.
If pin 6 stays near 0V, it's likely the output on U32.5 is toast.
What is the voltage at U33.5?
On U37, three of the four outputs, pins 2, 3, and 15 are unusually low. The fourth, pin 14, is inconclusive because it is wire-OR'd with two other gate outputs. I am highly suspicious of U37. The reference voltage generator in the chip is shared between the gates and it might be dead.
A possible way to check is to make the unused output, pin 9, live by connecting it to Vee through a 470R resistor. See if you get a valid ECL logic level on pin 9. If not, it may be time to pull U37.
If you remove U37, I'd recommend 1) put a socket in its place, and 2) check the voltages on all the pins while it's out because the isolated readings may provide clues to other problems.
I'm also noting that U37.2 is wire-OR'd with U33.9, but other sections of U33 appear to be working, which goes against the dead U37 reference generator theory. I'm not sure what to think of that yet, but it may also indicate an issue with that gate on U33. Measuring the pins on U37 with U37 out can tell us more.
And as with placing a 470R on U37.9, it may also be useful to make U33.15 live and get a voltage reading on that too. (U37 and U33 pins 9 and 15 should be complements of each other.)
Another approach when there are many logic gates hanging off a single node, like in this circuit, is to use a logic pulser and current tracer combo, such as the HP 546A (adapted for ECL) and HP 547A. These allow you to determine where the current is going. They can be very useful with ECL with its wire-OR'd architecture, and can often lead directly to the culprit. If you have these, let me know and we can make use of them, but they are not common. They can be had on ebay, but they command a high premium and not worth it at the going rate, in my opinion.
Mick B:
Hay there, I can't believe I left off pin 5 :palm: I went back and rechecked pin 6 again and lost where i was.
U33 Pin 5 is -0.814
I will do all the things you ask. I made an offer on one of those logic probes, between my son and I we have several HP devices. of this era. If he excepts, I should have it by June 6 at the latest. I would like to take this time to go through the manual and see if I can make heads or tails of these ECL circuit's I have a basic understanding, but I can do better. I an hoping by the time this HP8116A gets worked out I will have a complete understanding of this unit and how it does what it does at the circuit/component level. Thank you, Mark.
Mick B:
Things to report. The logic probe will be here on 6/1. The voltage is up to 58mv on pin 6. and holding.
ALSO you said "On U37, three of the four outputs, pins 2, 3, and 15 are unusually low." did you meant high?
Anyway, using 470Ω as requested produced.
U33.6 -0.085
U33.15 -0.758
--------------------
U37.9 -3.43
MarkL:
--- Quote from: Mick B on May 28, 2024, 05:30:22 pm ---Things to report. The logic probe will be here on 6/1.
--- End quote ---
I had mentioned the HP 546A Logic Pulser and 547A Current Tracer. There is also a thing called the HP 545A Logic Probe which is often sold with the other two, but not as useful if you already have a scope. What is it that's on the way?
If you have not used any of these tools before, you can familiarize yourself:
https://www.hpmemoryproject.org/an/pdf/an_163-2.pdf
--- Quote ---The voltage is up to 58mv on pin 6. and holding.
--- End quote ---
Is that +58mV or -58mV? If +58mV, I'm not sure where it's getting a positive voltage. Maybe take a look at U32.7 since that signal is coming from a TTL IC with positive supplies. And if -58mV, it's just wrong anyway since it should be between -1.75V and -0.90V (again, typical values).
--- Quote ---ALSO you said "On U37, three of the four outputs, pins 2, 3, and 15 are unusually low." did you meant high?
--- End quote ---
No, I meant low. Those outputs are between -3.70V and -3.13V and a typical ECL logic low is -1.75V. I'm saying those outputs are very low compared to -1.75V (-3.13V < -1.75V).
--- Quote ---Anyway, using 470Ω as requested produced.
U33.6 -0.085
U33.15 -0.758
--------------------
U37.9 -3.43
--- End quote ---
U33.6 is stubbornly staying near 0V and it shouldn't be. The output is an emitter follower, so should be one diode drop below 0V (approx. -0.7V) at a maximum. It could be the output driver on U32.5 is shorted. The other half of the output, U35.2 U32.2, is connected to U33.5 which you report at -0.814V and is ok.
U33.15 is ok.
U37.9 is very low with the only thing on it being the 470R test resistor. Again it should be in ECL logic level range. Both of the outputs on this gate are very low and I'm not coming up with a good model (yet) of how this could be happening.
Below is a diagram of a MECL 10k OR gate, for reference on what's going on inside the chip. Some of the other gate types are structured differently to reduce propagation time, and unfortunately those diagrams are not available so we can't make too many assumptions when considering what might have failed internally.
The diagram is from the Motorola 1989 MECL System Design Handbook, which has a detailed discussion of the circuit's bias levels and transfer characteristics. Chapter 1 is worth reading to understand how ECL works, and what could be happening internally when things go wrong.
http://bitsavers.trailing-edge.com/components/motorola/_dataBooks/1989_Motorola_MECL_System_Design_Handbook_4ed.pdf
EDIT: Typo - fixed reference to U32.2.
Mick B:
Wow, Information overload I love it!
This is what I received today HP 546A Logic Pulser. I will no longer call this Logic Pulser a probe. So much reading, I downloaded both of these publications and the service manual for the HP 546A.
--- Quote ---Is that +58mV or -58mV?
--- End quote ---
It is -0.058
--- Quote ---No, I meant low. Those outputs are between -3.70V and -3.13V and a typical ECL logic low is -1.75V. I'm saying those outputs are very low compared to -1.75V (-3.13V < -1.75V).
--- End quote ---
DOH! I was looking at my numbers being higher than yours. I get it, a lower negative voltage referencing ground. I won't make that mistake again.
OK! I'm going to do a lot of reading. I will get back to you in a couple of days, also I have a couple of other things that need my attention. Thanks again.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version