Scoping an Ampeg Gemini VI

[RWood]

I am taking the opportunity of a noisy Ampeg to learn how to use my scope, a Philips PM 3212. I have two 10X probes, and tonight I connected one of them to the filter caps in hopes of evaluating ripple current. What I'd like is for you scope users to comment on whether I have done this procedure right, and also help in reading the results.
After doing ripple I would like to use the scope to look at an audio signal.

This amp came to me having been recapped, it looks like 11 years ago with Sprague Atoms replacing the original can. There are two 600V caps and three 450V, strapped like dynamite. I hooked my probe to the + end of the second cap, clipped the ground to the - end. Set the scope for AC volts, and a range of 50mV.



Here's what the results look like:



Here's what I get when I clip the probe to the first cap:




These are the settings I used:



So did I do this correctly? If not let me know what I need to do. And is my conclusion correct that the second cap is allowing much more ripple current than it should and should be replaced? Also if there are any other observations you have, please let me know.

Thanks!

[Tom Phillips]

The scope setup appears to be correct.
With the 10X probe and the 50mV per division vertical amplitude setting your display sensitivity is 500 mV per division. Note however, that the center knob on the amplitude control is not set to the "CAL" (Calibrated) position so your vertical sensitivity is really something less than indicated. (Normally you would set up your scope with both channels set to the "CAL" position.)

The ripple looks pretty low. If we guess that the actual ripple is about 250mV and the supply voltage is ~400V then the ripple would be ~0.06% which seems fine for a second stage. Have you measured the ripple with your DVM? If so, what do you calculate?
Ripple is only one parameter however. Other things could be bad with those caps such as high ESR due to their age. The date codes on the caps look like 1982 to me so I'd say that they are 30 years old rather than 11. The other age tip off is the yellow color. 11 years ago Sprague had already switched to blue.

Good job posting the photos.

Cheers,
Tom

[RWood]

Thanks, Tom. I retried it with the amplitude center knob set to CAL. This is weird because now I get much more ripple on the first cap (which is what I was originally expecting) and I had to change the amplitude to 1V to get capture the whole wave on the screen. Here's what I get:



I don't have an ESR meter - I was hopeful that looking at caps with the scope would give me an idea of which ones were starting to fail. Maybe that logic is not correct. Measuring ripple with the DVM...... I either can't do that with my el cheapo meters or I am doing it wrong. I have two AC settings, 200V and 700V, and when I probe + and - on the cap I get twice the DC voltage as my reading. Please enlighten me here.

I was reading the date code on the Atoms wrong - thanks for pointing that out. I think this Ampeg will get a new can from Fliptops. That will give me a lot more room ,too. This amp has also been modded to run cathode bias, and the 7199 tube has been re-done to take a 6U8. As you can see, it is an early pcb amp. This one takes 6L6GC power tubes rather than 7591s.



I have a signal generator but not a dummy load. When I finish with the power supply I would like to probe each stage, from the input.

I appreciate the help,

RWood

[Tom Phillips]

... I had to change the amplitude to 1V to get capture the whole wave on the screen. Here's what I get: ...

The scope trace is showing ~35Vpp which is ~12 Vrms. Assuming that the B+ at that point is ~450 V dc the ripple is ~ 2.6% which seems normal for that amp. This needs to be adjusted for your actual DC power supply voltage. The formula is (Vac/Vdc) x 100 = percent ripple.

... …Measuring ripple with the DVM...... I either can't do that with my el cheapo meters or I am doing it wrong. I have two AC settings, 200V and 700V, and when I probe + and - on the cap I get twice the DC voltage as my reading. Please enlighten me here...

I’m not sure what you mean by “I get twice the DC voltage as my reading.” With a good meter you should read 12 V on the AC setting (roughly based on your scope measurement) and 450 (or whatever your B+ is) on the DC setting.

[RWood]

The scope trace is showing ~35Vpp which is ~12 Vrms.

It looks like you are measuring the entire wave from top to bottom, so pp = peak to peak. I get that. You were right on with the 450V, btw.

I’m not sure what you mean by “I get twice the DC voltage as my reading.”

Here is what I get when set to DC, then changing nothing except switching the meter to AC:





Maybe it is time for me to get a Big Boy meter!

Thanks for the help,

RWood

[Tom Phillips]

...Here is what I get when set to DC, then changing nothing except switching the meter to AC:...Maybe it is time for me to get a Big Boy meter!...

OK. Got it and I have heard of that problem. Somewhere on the forum there is an explanation telling why inexpensive meters give erroneous readings when they are asked to measure a small AC signal riding on a relatively large DC voltage. So, yes, you need a better meter if you want to accurately make such measurements. It's good you have the scope.

[Tom Phillips]

RWood,
I found the thread I referred to here http://music-electronics-forum.com/t28323/ .
They are discussing the exact problem you are having trying to read the AC portion of the power supply rail.
It's from last February and with the title "Why Does.." no wonder it was so hard to find.
Tom

[J M Fahey]

Sef quoting what I wrote there (and which speaks about your exact type of meter):

It happens with all the cheap multimeters which have just 2 AC scales: 200V and typically 750V (may also be 500 or 1000).
Only meant for tyros to measure Wall voltage ot transformer secondaries, neither of which has a DC component.
It also happened on old needle multimeters, which to read AC on plates, had an extra input labelled, ironically, "Output" which had a .1x400V capacitor in series.
Now, if you have a multimeter which can typically read down to 200mV AC, they have added an amplification stage or at least an active "perfect rectifier" so they *can* afford an input capacitor.
But in a $10 multimeter, forget it.

[Enzo]

Those are great little meters, I have a bunch of them. They are real handy to have sitting around the warehouse. I can pick one up most anywhere and use it to see if an outlet is powered or if a speaker voice coil is open. I don;t need to walk back into the shop era to get "my" meter. But it is a very cheap meter. Harbor freight sells them for under $10 - and that is list price. They often have a coupon in the paper with that meter selling for only $2 or sometimes $3. Not only that, they regularly have them as their FREE item. That is when I go get them. For $2, or $3, or even the $10, it is a great little deal, but limit its use to the most rudimentary of checking stuff.

You don't need to spend $300 on a meter, though that does get you a darn nice one. Once you get past maybe $50, they probably all work pretty well. In our work, the difference between 1% accuracy and 0.001% accuracy is irrellivant. My nice FLuke meter has been droppd to the floor so many times I can;t count, and it has served me for decades. A basic Radio Shack meter might not have survived.

[Enzo]

Those are great little meters, I have a bunch of them. They are real handy to have sitting around the warehouse. I can pick one up most anywhere and use it to see if an outlet is powered or if a speaker voice coil is open. I don;t need to walk back into the shop era to get "my" meter. But it is a very cheap meter. Harbor freight sells them for under $10 - and that is list price. They often have a coupon in the paper with that meter selling for only $2 or sometimes $3. Not only that, they regularly have them as their FREE item. That is when I go get them. For $2, or $3, or even the $10, it is a great little deal, but limit its use to the most rudimentary of checking stuff.

You don't need to spend $300 on a meter, though that does get you a darn nice one. Once you get past maybe $50, they probably all work pretty well. In our work, the difference between 1% accuracy and 0.001% accuracy is irrellivant. My nice FLuke meter has been droppd to the floor so many times I can;t count, and it has served me for decades. A basic Radio Shack meter might not have survived.