HP 3440a digital voltmeter repair/troubleshooting: moving on to PSU ripple

[Armadillo]

I have a question for you.

If I have a signal that is DC biased at somewhere, let's say, -35Vdc carrying "a" pulse, just "ONE" pulse of 200mV Peak to Peak,

What setting of the oscilloscope would you set to capture the pulse clearly?

I thought, the answer could be very educational, in the hope that members who read this post of "yours" would miss "no pulse" in future.

[valley001]

I am not sure I fully understand the question but will attempt a shot at it.

Would it need to be AC coupling with 100mv scale?

I answer this way because when I set to AC coupling it centers the signal on the scope.  I am really showing how green I am I know. 

[Armadillo]

Don't worry about Green, I was also green, everyone are also green once. The important is, you don't miss the important pulse ever.

Your answer is essentially right, but you miss that "One" pulse, it's not a train of pulses. My take is you miss the pulse.

No worry, I see 1328 people read your post as of now. Incredible.

So let's others have a go at it so that we all learn a common something essential from here. I am not sure either.... so I better keep my mouth shut.

[Armadillo]

Imagine a simple posting like on the "Repairing a Logitech Z906 Speaker" can hit 39,541 views, just amazing and incredible. 
May soon even beat the Pin-Up Posts, above. LOL 

[valley001]

C2 replaced, the new one is so much smaller.  It actually fits under the chassis directly under the old cap.  Old cap is out of circuit now.  Now I dont have any problems with the -35v rail. 

I went through the calibration procedure and it seems to be holding linear and within 5 counts or so.  Reading my precision reference it reads 9.996, 7.496, 5.000, and 2.501

I think there is an issue with the binding posts because if turn them, fiddle with the earth ground jumper, etc the count will dither a bunch. 

[coromonadalix]

maybe some deoxit stuff could do some cleaning and remove surface oxydation

[valley001]

Going through all of the adjustments/calibration on this thing I decided to attempt to look for PSU ripple on -35v rail. 

On the scope, I see -34.6ish average with 1.3 volts peak to peak!  That seems excessive. 

I had the scope set to DC coupling with the 20mhz limit on.  If I switch to AC coupling I could detect a small signal in the 8-9mhz range (harmonic of some sort?)  I do have the regular ground lead attached which I have learned is not ideal when trying to measure psu ripple, I may try making a shorter ground probe with a paperclip or something.  I suppose I will check the filtering caps on A9. 

[Armadillo]

There was a guy that asked the same question on his SMPS ripples using full 100 mHz, he didn't come back with his question there after.

But Oh, I say you are very correct with the BW limit and with the short measuring leads. Must have done a lot of homework there.   

Increase ripples = increase resistance implying possible failing capacitors, cold/corrode solder joints, poor contacts, failing components that draw extra currents etc....

But I am sure you are all able to solve the problem.   

Please share your interesting findings. thanks.

[valley001]

Which are the smoothing caps on the -35v side?  Its not C6-8 or is it? 

[Armadillo]

Any capacitors that are tied to the -35V rails are the smoothing capacitors.
So it can be the local capacitors like C2 at Q1, C3 or any remote ones that are tied to the -35V line.

[Edit: Particularly at the Q2 [amplifier] and Q1 Driver where any ripples get amplified to the Series Transistor Q1 there. See A9 Schematic].

Improve the C2 and C3 there.

[valley001]

Another thing I am working on simultaneously is the 400khz oscillator measurement.

I have been using my scope to take this reading, and when I initially checked I saw it was 400Khz +/- but reading the manual I see it should be 400Khz +/- 400hz.  I see it fluctuating between 401.5Khz ish and 399 Khz ish which falls outside the specification.

Is my 70mhz scope sufficient for taking this reading accurately or do I need a proper counter? 

[Armadillo]

Your scope at 70mHz is well adequate to measure the 400khz frequency.

But I think your meter as measured is accurate enough. The wide tolerance of 400 hz indicates that this is NOT the critical parameter because the accuracy of measurement is on the Dual Slope Measurement technique example both the start to count and stop to count are dependent on the slope regardless of the slight variant. Don't be too piety   

But as you progress, you need to change to better scope definitely.

But 70mHz scope and your knowledge technique already can solved 99% of the many practical daily electronics of today.

 

[valley001]

OK, sounds good. 

This brings up another question, but is it possible to visualize the ramp and incoming signal comparison on the scope using two channels? 

[Armadillo]

Channel 1 scope and TRIGGER on the Rising Edge of the Transfer Ampl at Q12 collector

Channel 2 scope on the Ramp Output at the Collector of Q16.


Channel 1 scope on the Ramp, trigger on start pulse of Ramp

Channel 2 scope on the stop pulses

You will see the count portion of the Ramp or T0 and T1 effectively speaking.

As long as you find joy in doing that.