Princeton reverb clone hum issue

[Randall]

A Magic Amp Vibro Prince came in with filter hum. It's a Princeton clone. I had tacked in a F&T 22/475v about 6 months ago as one section of the CED can had gone bad. It came back with another section bad, so I put in a new CED can. The hum is reduced, but there is still some there. I have gotten it down to this:

V1 and V2 removed, tank not connected, new matched outputs, biased at 20mA. Remaining V3 and 4 swapped out for no change. Amp hums.

Grounding the second grid of V3 quiets it, as does either side of 3.3M/10pF. Grounding the reverb pot side of 470K makes no change. Jumpering a 22uF/475v cap across the D node reduces the hum, but not as much as grounding V3 grid.

What is up here?

Also should mention, I see anywhere from 50 to 120mV on the black fiber board in places.

https://robrobinette.com/AA1164_Princeton_Reverb.htm

[Chuck H]

Well the hum isn't coming from the reverb driver grid through that 500p cap. And it's not coming from V1b plate because it's not in there. Nothing left but the power supply (I do assume you already inspected grounding between V1b and V3b).

It's probable the cap you bridged across the D filter just had higher impedance than straight to ground with a wire. It sounds to me like it's time for a whole new can or a retrofit for individual filters. There could even be some shorting between can sections confusing the matter by now. It's been so long since I had to buy a can that I wouldn't even know what to suggest now. With the small size of individual filters some guys like to cut open the existing can and stuff the new filters in there for a stock look. But that doesn't sound like your MO. What "I'VE" done before is just disconnect the can and leave it in place for the "look" and then build an individual cap power supply filter assembly on one of those five terminal strips that have the center terminal grounded by the screw mount.

[Randall]

Maybe I wasn't clear. I had subbed a cap for a bad section in the past. This time I put in a new CED can cap. Edited post #1 to clarify.

[Chuck H]

Was the new can a fresh one? I've had no end of trouble with NEW caps purchased from Mouser that were "stale" of their expiration date. And I don't do repairs so this isn't the sort of thing that I would have seen repeatedly like a repair tech. And yet it's happened to me. JM2C on the viability of electrolytics in a repair techs antiquated inventory or off the shelf from some money grubbing "guitar amp parts" site.

If the amp worked fine before and only the power supply could be casual now,.. What else is there to suspect? When you've eliminated the impossible, anything else, no matter how improbable, is suspect and probably the cause.

[nosaj]

Shouldn't it be simple enough to check the health of the caps via ripple test?
nosaj

[Jazz P Bass]

Agree.
I have had "new" cans that came through with faulty sections.

[Randall]

New can from CED.

Define "ripple test" please.

Why does grounding the grid of V3 quiet it down?

[Chuck H]

Why does grounding the grid of V3 quiet it down?

The only explanation (if V1 is removed as indicated) is that it's coming through the 100k plate load resistor (from the power supply rail). Otherwise it would have to be coming from ground or through the 500pf cap at the reverb grid driver. Both very unlikely.

[nosaj]

Test each power leg.
nosaj


Measuring the ripple with a tester


Ripple voltage is a small AC voltage, which sits above a DC offset. It can also be measured using a digital multi-meter, even if the operation is inconvenient, inaccurate, and challenging. For this purpose, the probes of the digital multi-meter and the regulator output are connected. The red is attached to the positive terminal and the black to the negative terminal. Select the “AC voltage” mode by turning the tester knob. The meter will only measure the AC component of the signal, the ripple voltage, if present. The display shows the measured value of the ripple voltage amplitude. If it shows the value 0.00 mV it means that, most likely, the ripple is absent. To correctly characterize the ripple voltage, it is necessary to measure its frequency. Pay attention to the type of AC voltage that the tester measures. It could be an effective value, an average value, peak-to-peak, or other. Furthermore, most times, the ripple waveform is not sinusoidal, but, it is always necessary to check the type of measurement of the multi-meter. The simple wiring diagram shown in figure 2 shows a very simple AC/DC power supply. The components proportioned to provide a DC output signal affected by a ripple noise.

[patlaw]

New can from CED.

I have decided that, unless the owner is concerned about looks, I will no longer use CED can caps. They're made from the same equipment that Mallory originally used, supposedly, but the discrete caps, including but not limited to F&T, are modern technology and, from what I read, higher quality than the sections in the CED cans. Don't ask me to prove what I just wrote because I can't, but it's logical, and I've heard others say the same.

[Chuck H]

Even the antique radio guys avoid buying cans. There are several tutorials on how to stuff discreet capacitors into the old cans for the 'look'. As long as it's been since I had to buy a can cap all I can say is that JJ never did me wrong. Anything made with mojo as part of the formula (like using the original Mallory equipment) usually doesn't appeal to me because what they're selling is lore first and capacitors second. I'd rather buy from a capacitor manufacturer that puts capacitors first.

[Randall]

I thought by "ripple test" you had something specific in mind. With V1 and V2 removed, at the nodes I see

A - 4.1 Vac at B+ node
B - 180 mVac at screen node
C - 2 -4 mVac
D - 1-2 mVac

At V3 I see 1 mVac on pin 1, and 6 mVac on pin 6. To address Chuck's point that it must be the plate resistor, it is 1 mVac on the feed side of that resistor.

So, I am still unclear where this hum is coming from. And especially why grounding V3 pin 7 makes it stop. Where is the evidence this new can is bad?

[nosaj]

Pretty sure you shouldn't have 4vAC on a b+node .
nosaj

[Helmholtz]

I just measured my Vibrolux Rev.

Has 4.5Vrms ripple on first node, 100mVrms on screen node and 50mVrms on PI node. No hum at all.

I think measured ripple values are normal.

[nosaj]

Check your reverb cables for a good ground? V3 is part preamp part reverb recovery.
nosaj