HP 8116A func gen with E51... U401 shaper toast?

[sprok]

.

[Jwalling]

U401 has a differential output. What about pin 6?
The chip is socketed, possible to isolate pin 5 and check with a scope?

Jay

[Kleinstein]

The output amplifier pulling the chip down is possible. AFAIK the outputs are current driving and one is going to a current mirror. So there should be a rather low amplitude (voltage) with that output. The other output should be the more interesting part. To check for loading one might measure the impedance of that node by injecting a small AC current to that node, while powered. This could also check the output amplifier.

There are also quite a few other inputs for amplitude modulation and others, that might have a wrong level. So there are other signal that might be causing the chip to give a small signal.

[Jwalling]

Non-isolated pin 5 on the scope is flat 1mV.   Pin 6 is flat 3.15V.

Not sure about the isolated output. Bend pin 5 out? Heh, not sure I want to go there until I'm reasonably sure the chip is dead. Are you thinking that the output amplifier may be pulling pin 5 down?

I was thinking more along the lines of using a sacrificial socket in-between with the pin removed. Perhaps both pins 5 and 6 in case one can affect the other. Given your DC voltage reading on pin 5, it does sound like the chip may at fault. Does it get nice and toasty? those hybrids run pretty hot normally.  If it runs cold it could indicate that it's dead.  Also would be a good idea to check the IREF circuit on pin 18 though there's not a whole lot that could go wrong there.

Jay

Jay

[Kleinstein]

-2 V at open  pins 5 and 6 might be OK, as this are sources of negative current. A more useful signal should be there with something like 100 Ohm each to GND.

It might also be interesting to check the pulse mode, as this bypasses the sine shaper, but uses the rest of U401.
 

[Jwalling]

Isolated pins 5 and 6 and they both are flat -2V and have no waveform.

OK. My only concern would be whether the outputs could be open collector. Probably not, but there's no detailed data on any of them. How about putting a gentle pull-up of 10K or so on pin 6 to +5V?

Jay

[MarkL]

Seems odd that the 2nd harmonic adjustment input (pin 7) is at the extreme range of its possible input of +5V to -5.2V.

One thing to check with U401 out of socket is to make sure R417 can adjust pin 7 between those voltages.

[Kleinstein]

According to the manual pins 5 and 6 are the collectors of a NPN long tailed pair. So a relatively low impedance to GND or + 3 V should be suitable. I would expect a current of something like 1-5  mA, so I would not use more than about 220 Ohms. For the rest of the circuit I would expect something like 50 Ohms to GND at pin 5 (pin 6 is the current mirror and thus similar to 2 diodes to 5 V).

[MarkL]

Here's the readings on U401 on yours and then a working unit:

1	-550mV	-1.11V
2	-5.2V	-5.19V
3	+5.0V	+5.03V
4	nothing	nothing
5	nothing	1.09Vpp sine @ 0.00V
6	+3.2V	590mVpp sine @ +3.03V
7	-5.2V	-5.20V
8	+800mV	-1.38V
9	-38mV	-46.4mV
10	+3.7V	122mVpp sine @ +2.71V
11	+3.9V	122mVpp sine @ +2.79V
12	-360mV	-360mV
13	-380mV	-576mV
14	-750mV	-761mV
15	+190mV	+187mV
16	+38mV	-45mV
17	+20mV	-84mV
18	150mVpp triangle @ +440mV	131mVpp triangle @ +304mV
19	150mVpp triangle @ +440mV	134mVpp triangle @ +304mV
20	nothing	nothing
21	-1V	-1.09V
22	+4.7V	+4.58V
23	+220mV	+194mV
24	-1.56V	-1.64V

At the moment, nothing is jumping out at me except the level shift inputs (10 & 11) are a volt higher in your unit.  But since we don't know what's inside U401, there's no telling if that's significant or not.

There's also no sine wave on those two pins on your unit, so that may say the problem is either in or prior to that stage (looking at the U401 block diagram).

Have you tried different waveform shapes or tried pulse mode?  Or complement vs. normal?  Any difference fooling around with the AM modulator settings?

EDIT:  There's also a large-ish difference on vernier trim input (pin 8 ).  Again who knows.

You said you did "most" of the checks in the service manual.  Can you tell us which tests you did and the results?

[MarkL]

The manual is wrong.  I agree with your analysis that there's no path for base current in Q203 or Q204 (could have been via R204).

Double checking the unit I have here, there is indeed no current on U201 pin 9 to ground with U201 out.  Iup-ref (pin 5) and Idown-ref (pin 3) are as stated in the table.

I think it's good to work through the troubleshooting procedure to make sure there's not something overlooked, but I would be surprised if it's anything with the VCO since you've already observed a signal from it at the shaper input, and the levels there are not out of line.

Side note: There's been a ton of changes to the design of the 8116A during its life.  If you come across something that's inconsistent, check the backdating pages against your serial number.  (Although there's nothing there about this pin 9 thing.)

[Jwalling]

R227 adjusts the offset on TP1 so I set it to 250mV as the service manual shows. Of course that made no difference as far as fixing anything.

I think at this point U401 looks broken. If I had any more canned air I would try freezing it to see if that might temporarily make it work again.

How about putting it into the freezer for awhile in a sealed bag? 

Jay

[Kleinstein]

It does not look good for U401. The chance of lower temperature changing much is not really high. It might be difficult to get a replacement.

The first place to see the signal seems to be pins 10 and 11. So it does not look like the output stage is the problem.  Overall the internals of U401 could be similar to other f-gen chips like the XR2206 or max038, where internal circuit is known.

One could try to adjust the voltages at a few pins (1, 13,17,21,24) to the same values as in Marks measurement. Many of them are just going to pots - so different values can depend on the pot settings. It might help to see if the voltage at pins 10/11 moves with one of the pots - this would be a good sign. This might include changing the ref. current a little away from the nominal 2.5 mA.

[MarkL]

All the detailed tests show everything around U401 as working.  I'll join the consensus and say that U401 is very likely done for.

Unfortunately, it doesn't look like there's any U401 (1826-0923) on ebay or other scrap dealers right now.  The chip is also used in the 8112A, so that's another salvage possibility.

The 8111A also has a shaper chip that is remarkably similar (same voltages and pinout except for pin 17), but that would be a gamble since it's a different part number (5180-2417) and the 8111A is a 20MHz unit.


To confirm that your U401 is dead I'd be willing to try it in a working 8116A if you want to mail it to me (we're both in the US).  PM me if you'd like to arrange this.

[MarkL]

Interesting!  At least you have a working unit now to do comparisons.

So, the E21 unit with the "broken" U401 has an output on pins 5 and 6 now?  And pins 10 and 11, presumably?

And if U401 is removed do you get E21 back instead of E41?

You could try removing U300 and see if U201 is still hot, and visa-versa.  They could be cooking independently or it could be related.

Back to the beginning and checking voltage levels...

I forgot to ask you... Did the original unit ever work in your possession or did you buy it broken?  Just wondering if someone already did some chip swapping on it.

[siggi]

Another thing I have noticed is that the main output offset is double whatever is set. For example if I set a +7.5V offset then the main output will actually have a +15V DC offset. It seems to do this at any offset setting (eg. 1mV will actually offset 2mV). Is it suppose to work that way?

Sounds like you're not terminating into 50Ohm. If you're not terminating the output into 50Ohm, but measuring with high-impedance DMM or scope, you'd expect to see very close to double amplitude and offset.

[MarkL]

None of these units worked when I got them. The newer custom chips seem to be fully encapsulated (black on the bottom) and have white sticky-on labels with the part numbers. The older custom chips are white on the bottom with exposed metal and have the part number printed (silk-screen?) directly on the chips. The E51 unit's VCO is white bottom so may have been replaced at some point.

And hopefully all of the custom ones (U401, U201, U300) have heatsinks on them!

Since you purchased these units broken, there's a possibility that there's more than one thing wrong.  Sometimes people combine units to get one working unit and then sell off the broken units re-stuffed with broken parts as non-working.  Just something to keep in mind.

For the "Low Amplitude & Offset" settings I am unable to get the offset correct. I can get the amplitude correct by adjusting A1R450 (square), R445 (sine), but the offset adjustment on R416 won't go high enough and I have about a -64mV offset error when the main output amplitude is set to 1V. I have tested pin 8 on U401 without the chip installed and it moves between +4.3V and -5.2V (my other E21 unit does the same). However with U401 installed it moves between +1.76V and -5.2V. On the E21 unit it goes between +2.3V and -5.2V. This difference is apparently not caused by the U401 chip itself because even if I swap between my two U401 chips the adjustment range is the same. On the other voltage ranges I was able to set the offset correctly.

Is the 64mV offset showing up on TP4 also?  Perhaps there is a problem with the main output amp now?

If TP4 is not ok, are U401 pin 18 and 19 balanced?

What section of the adjustment procedure are you performing?  A1R416 and A1R450 are mentioned in the Power Supply section, and then again in the Shaper Adjustments section.  There are more detailed adjustments in the latter.  If you haven't done so already, perhaps going through the shaper adjustments will tweak them and other pots to bring the offset into spec.

Not to be discouraging, but these things are a royal pain to adjust.  The pots have way too much range and make the right setting sometimes hard to achieve, and there a several areas where there's a lot of interaction between adjustments.  I think it took me a day.

[MarkL]

...
Hmmm, the manual says R439 can change the shaper offset. On the newer unit this resistor is not installed. So "open" which I assume is the lowest offset. I wonder if I install a resistor there maybe I can increase the offset which is what I need.

The older unit (which is where this shaper came from) is using a 12.1k 1% resistor there. The manual says 7.5k is the max... 

Table 9-1 says that increasing the value of R439 (going towards open) increases the offset.  But it's already open (if I have these two units right), so that would say offset is already at max at -64mV.  You could try holding a 12k in R439's pads anyway to see what it does.

Also take a look at R428, range: 10k to open.  That affects the 1V offset.  "Increasing the value decreases negative offset."  I'm not sure how to interpret that - too many negatives.  Playing with some different values there might also help.

[MarkL]

Great - glad you got it working!

Have fun with the calibration!