Testing Bad Filter Cap with a scope for 120HZ hum

[wsscott]

I've traced 120hz hum down to the reverb section of a ReVibe Mod that seems to be coming from the B+ C node, 30uf cap.  I think it may be in the cap because other components in the circuit pick up the hum at that point.

I hooked up my scope's probe to the terminals on that cap, and set the scope for 100mv/div with AC coupling.

Although I sporadically got an image of a sine wave of around 120hz, I did get voltage on the scope that reads about 750mVrms.

Is this methodology valid, and is this indicative of a bad cap?

[shooter]

Although I sporadically got an image of a sine wave of around 120hz, I did get voltage on the scope that reads about 750mVrms.

the voltage constant or sporadic also??


you can also use a Meter, set it to VAC, start at the 1st filter cap and work your way "in" to the last node.


typical would be something like;
2-4VAC at 1st tap, 100mV at 2nd tap, 50mV, 20mV.....
each tap should "kill" the AC more n more.


If.. the VAC "spikes up" at a tap, that node should be suspect.


a quick test, gator-clip a bigger 50-100uF cap in parallel with your 30uF (MAKE SURE voltage rating is equal or greater than what's there)


you should "hear" a decrease in background hum

[wsscott]

Interesting results:

ESR meter shows all 5 caps, Nodes A-E, have proper readings.

I don't have a 50-100uf cap to test, just a 20uf.

Node A:  This Node does not connect to any points in the circuit.  It just receives the power off the full wave rectifier, and passes the signal along to Node B, and on.

1.81 vRms--STEADY
120Hz Big sine wave skewed at top

Node B:
1.01 vRms-STEADY
120Hz large sine wave-pretty smooth-not as big as in Node A

Node C:
638-655 vRms- fairly steady
0 Hz--sometimes tiny sine wave appears-scope reads 0 Hz

Node D:
395 vRms-STEADY
0 Hz-no sine wave

Node E:
243 vRms-STEADY
0 Hz-no sine wave

[shooter]

Node E:243 vRms-STEADY0 Hz-no sine wave

most scopes should display 243mV, most nodes that far down the chain should be ~~~20-40mVrms
time to repeat with a good DVM

[tubeswell]

I don't have a 50-100uf cap to test, just a 20uf.

20uF will work fine for a test.

[wsscott]

Sorry, nodes C, D, and E are in mV.

[shooter]

figured they were, my scope will display down to ~~~ 20mV  so your scope might be sketchy, that's why I recommended a redo with a meter.

[BrianS]

I use a volt meter in the way Shooter described.

[wsscott]

I tried checking the VAC with my DMM on the filter caps, and got screwy numbers.  I don't have a DVM.  Maybe that's the issue.

The DMM has an AUTO and a Manual function for voltages.  I tried both methods and got these numbers that don't make any sense.

24.18, 23.65, 21.38, 16.47 and 12.24 for A-E respectively.

The DMM's manual shows the following AC ranges:  400mv, 4V, 40V, 400V, and 750V.

[wsscott]

I guess my DMM isn't sensitive enough.

[shooter]

i use DVM n DMM interchangeably


are those numbers millivolts, volts, kilovolts...?


on the 400mV scale that is the MAX reading for that setting, guessing IF...the volts are stable it should be able to see 40-50mVac fine.

[wsscott]

This seems crazy, but the DMM is in Manual mode and these numbers are all in Volts.

Its in the Manual 40V range since lower range gave me OL.

Doesn't make any sense.  Now these numbers probably are not true VRMS.  My scope's numbers were in VRMS.

Maybe that's the explanation.  But even if you multiply them by .7, they're still much higher than you predicted Shooter.

The DC voltages on the filter caps are 339, 337, 320, 288, and 262 respectively, and are right on with the schematic.

[shooter]

This seems crazy,

 


what I thought with your scope readings, hence meter, which still yields "crazy" so I suspect you either have equipment issues or ya got a gremlin that's all excited about being chased!!!

[wsscott]

I couldn't have said it better myself!

I think I've probably wasted enough of everyone's time on this.  Thanks to all for your input.  As always.

Happy Thanksgiving!

[wsscott]

This was my first project.  The circuit is designed for a full wave rectifier with a pair of diodes on each leg of the transformer's secondary going to the filter caps.  The transformer has a Center Tap.

Not knowing at the time the difference between a full wave rectifier and a bridge full wave rectifier, I bought a 3N259 bridge rectifier.  In the process of wiring it in, I learned that I should not connect the ground lead of the rectifier into the circuit because it would result in too much voltage with the CT also being used.  The CT is wired to the first filter cap.  So I was told to cut that lead off so I wouldn't screw up.  So I did.

And that's the way this has been wired. The voltages at the filter caps are spot on.  Its worked fine, but for this 120hz hum.

So my final thought on this:  Is that the cause for this 120hz hum?  Should I just remove that bridge rectifier and build a full wave with 4 diodes instead, and leave the CT connected as is to the first filter cap?

[shooter]

some lite reading;


take any good working tube thing you have, amp, PA.... and repeat your test of the taps with both your scope n meter.  POST both AC and DC volts.

[wsscott]

Well I'm totally baffled.  The voltages on the tube are spot on.  I've changed tubes, and no change in the hum.  The hum appears only when the 12AX7's pin 6 plate is connected to the 270K resistor coming off the Node C filter cap.  There's no 120hz signal from node C going into the 270K, and there's no 120hz signal after the 270K if it is NOT connected to Pin 6, plate.  There is only 120hz hum when the backside of the 270K resistor is connected to Pin 6.  The resistor reads within specs.

Will there be a Thanksgiving Miracle this year??

[shooter]

 


..."The ways of the Lord are mysterious..."


guessing you rolled a different tube to eliminate the tube
try temp'n in something between 500pF - .01uF cap across the plate R for that tube (Cap in parallel with resistor)

[wsscott]

Yes, the tube was the first thing I tried.

I'll give the other option a try.  Thanks for the input.

[wsscott]

I put a .01uf in parallel across the 270K plate resistor.

It shows over 340VDC with DMM and 1.5 VRMS with AC Coupling on scope, and 0Hz for the signal that was 120hz.

So that's good but the 340VDC is way too high.  It was, and should be, around 204 or so coming out of the resistor and into the plate.  Node C voltage into the resistor should be around 320VDC and it also went up to around 375VDC.

[shooter]

So that's good but the 340VDC is way too high.  It was, and should be, around 204 or so coming out of the resistor and into the plate.  Node C voltage into the resistor should be around 320VDC and it also went up to around 375VDC.

was "normal" before you added the cap???


IF SO....
measure VDC across the cathode R for the tube

[wsscott]

I decided to solder the cap across the resistor since the numbers were so screwy.

Interesting after I did that and turn on the amp, all of the Nodes were way high.  Then after about a minute they all came back down to where they're supposed to be.

As far as adding the cap, it made no difference.  Still have the 120hz signal.  I think the prior reading was because I didn't have good connections.  Also the VDC across the cathode resistor is 1.8VDC where it has always been.

So adding the cap made no difference and didn't drop the voltages or the Hum.

[wsscott]

For some reason the filaments on all the tubes have gotten really slow to warm up.  The power light comes on and measures fine.  Probably from screwing around with all this stuff.  When the tubes aren't lit, then no current is being drawn, and so the voltages on the filter caps and their connections are high--at least I think that's why they're high.  When lit, the voltages are normal.  Hopefully this quits acting up.

[shooter]

so the voltages on the filter caps and their connections are high

yep, I use the "high" voltage to spec the filter caps, I've even noted on the schematic the "high-time" 
the last amp was almost 30 seconds to drop down to normal operating VDC's


30 seconds at 400vdc across a cap rated at 350vdc might make a messy