5U4GT Rectifier voltages

[simonallaway]

Is this statement true:

"If functioning correctly, there should never be measurable AC voltage on either of pins 8 or 2 of a 5U4GT rectifier"

I'm measuring about 750 volts AC on these pins, which does not seem right at all. The amp has lots of hum, so I started by measuring DC throughout the power supply to see if they were in the ball-park. I noticed that while I did get a DC measurement, it was oscillating +/- 50V. So I switched the meter to measuring AC, and saw a steady AC voltage. Pins 6 and 4 give me a steady 325 volts AC.

This is a Supro (Valco) 6650TR. I've included the power section below.

[John]

If you're using an analog meter, it could be reading the AC voltage riding on top. At least, that's what my Triplett was doing that threw me off.

[simonallaway]

When you say "AC riding on top" do you mean the unfiltered rectified AC?

[John]

Honestly? I don't know.  But what gave me the idea was, when I was measuring the plate voltage with my Triplett, when I'd turn up the volume knob the needle would swing up to ~600v. I thought I had a  problem with the meter (another measured what you'd expect) until PRR told me (paraphrasing) that analog meters *tend* to measure AC and DC together when present, and DMMs *tend* to just measure one or the other.

So anyway, my idea was that's what your meter is doing,  measuring the DC present plus the AC voltage/ripple on top of it. I know that it taught me that if a voltage is severely out of whack, check your meter.

[PRR]

> Is this statement true:

No.

[tubeswell]

There will be the 5VAC between pins 2 and 8 from the rectifier filament winding.

[HotBluePlates]

There will be the 5VAC between pins 2 and 8 from the rectifier filament winding.

And some additional ripple voltage, because filtering is never perfect especially at the first filter cap.

I'm measuring about 750 volts AC on these pins ...

No way for me to tell if your meter is lying, if you had an odd connection into the circuit, or if a filter cap has failed. You have to know 750vac is impossible, because the 450v filter caps would have already exploded at that voltage (especially when the a.c. swings negative).

Have you tried removing tubes while measuring to see if the fluctuating d.c. settles down some? Maybe filter caps need replacing and aren't filtering out ripple.

When you say "AC riding on top" do you mean the unfiltered rectified AC?

The ocean's surface may be feet above the ocean floor or miles above the ocean floor. That average depth is like the d.c. present at your first filter cap connected to the rectifier's cathode.

But the ocean's surface is not flat and smooth like glass; it has waves. The amount of deviation above and below the average ocean depth is like the peaks and troughs of the a.c. ripple "riding on top of the d.c." at the first filter cap.

As Tubeswell pointed out, you have a 5vac winding lighting up your rectifier's heater. At a minimum, there's 5vac causing the d.c. at the rectifier output to bounce up and down 2.5v in each direction about the average d.c. voltage. The first filter cap won't perfectly filter the rectifier output, and there could be an additional several-voltas to tens-of-volts of a.c. due to imperfect filtering.

This isn't a problem for an output tube plate that swings hundreds of volts with a signal. It a bit more of a problem for the screen, which is why there's an extra stage of filtering for the screen node. Ten-of-volts is a really big problem for a phase inverter or preamp stage whose output may be a few volts to a few tens-of-volts. That why they have even more filtering after the already improved filtering at the screen node.

So either the meter is lying to you, filter caps are bad, or there's some arcane problem (I won't suggest those now cause I don't want to tempt you with a goose-chase). Easiest test is remove tubes, re-measure, try tacking in a fresh new filter cap, seeing if things settle down to predictable levels.

I highly recommend making similar measurements on another known-working amp you have on hand, ideally with similarly-sized output tubes, filter caps and/or rectifier. It will help you see what is normal ripple for a working amp at this point in the circuit (and typical ratio of d.c. voltage to a.c. voltage riding on top).

[PRR]

> I switched the meter to measuring AC

DMM, VTVM, or passive needle-meter?

Most pure passive needle-meters won't ignore DC on the AC range.

[simonallaway]

So either the meter is lying to you, filter caps are bad, or there's some arcane problem (I won't suggest those now cause I don't want to tempt you with a goose-chase). Easiest test is remove tubes, re-measure, try tacking in a fresh new filter cap, seeing if things settle down to predictable levels.

I highly recommend making similar measurements on another known-working amp you have on hand, ideally with similarly-sized output tubes, filter caps and/or rectifier. It will help you see what is normal ripple for a working amp at this point in the circuit (and typical ratio of d.c. voltage to a.c. voltage riding on top).

Thank you for all that. This weekend I'll re-measure with tubes out and try new filter caps. Even though I don't have another similar amp (only a SE champ-like amp and a marshall 2204) I'll measure those for AC and DC at various points to see what kinds of things I see. While I do have a schematic for this particular map, I don't have one with voltages. But I suspect the relationships between plate/grid/cathode are going to be within similar ranges to other similar amps that I do have voltages for. Any discrepancies that spring out should be good clues.

> I switched the meter to measuring AC
DMM, VTVM, or passive needle-meter? Most pure passive needle-meters won't ignore DC on the AC range.

It's a digital multimeter...a BK Precision 2703C.