Weber PT in new Marshall 2203 build behaving badly

[jamesmafyew]

I just finished building a Marshall 2203 clone with a few tweaks -- added a 12AX7 FX loop, bootstrapped master volume and elevated heaters a la TUT3, individual per-tube bias controls, and 150k bias feeds so I could use 6550s.

I'm using Weber transformers and a Weber chassis, but this is my own build and not one of their kits. PT is the Weber WPT100 (spec sheet here), OT is their WOT100HHR. I like Weber transformers, used them in a couple of 50-watt builds and I think they're generally nice and inexpensive and I've never had a problem with them before.

Anyway, fired it up and all is well, except that my unloaded B+ (no tubes in the amp) was 445V. I thought that was low so I switched from the 640VAC (center tapped) taps to the higher 700VAC taps. Unloaded B+ went to 500V, cool. PT measured 716VAC totally unloaded/disconnected across the winding (656VAC at the lower voltage taps).

Put tubes in the amp, fired it up, biased them to 35mA per tube, just to check it and see if it works. B+ sinks to 425V at idle! This is with the higher-voltage taps on the PT. Put a sine wave through it and get barely 60W into a resistive load, at which point b+ has sunk to 358V. Output wave is ugly and has a crossover notch, with one side clipping before the other.

I disabled the bootstrapped MV, returned it to a "normal" configuration. No change. Tried other tubes, no change. All supply voltages are sinking like a rock when the power amp is asked to deliver anything. I've used the VARIAC and monitored the current from the wall, and it doesn't go much above 1A when the amp is cranked up (oughta be a lot more than this, right?). So I don't think there's something in the amp dragging the supply down, and the load on the PT is not excessive. PT is not at all warm. I've been over the wiring a dozen times and (I know you've heard this one before) I don't think I made any mistakes. I've built a lot of amps, never had an issue like this.

Preamp is totally spot-on when I look at the FX send signal. So it's a power amp/supply issue. Happens when I inject into the FX return, which goes through a triode recovery stage (clean, scopes good). MV wiper signal is clean, phase inverter input grid is clean.

Phase inverter and power amp are clean until I get up to about 40W, then it goes south very rapidly. + going half of wave clips like I expect, short of 33VACrms, - going half rounds off in a weird way and clips sooner. It looks the same on both the phase inverter plates, like the phase inverter is breaking up before the power tubes do. The asymmetry is reversed on the two plates, so it's like it's happening at the grid...? But I scoped the grid + input and it's clean. Phase inverter plates are sitting at 169V (inverting) and 180V (non-inverting) at idle. Phase inverter supply node drops from 336V idle to 275V at full power.

No oscillation anywhere to be found, build seems totally stable. NFB looks right on the scope, presence control works normally.

Anyway, I'm a bit at a loss. It's like there's some invisible 300-ohm resistor in series with the B+ winding! Is the horrific voltage sag to blame for all of the ugly clipping and crossover notch? Should I just call Weber and send this PT back?

[jamesmafyew]

Quick update to this:

Disconnected NFB and ran the power amp open loop. Phase inverter is now clean right up to power tube control grids, power amp output wave is unchanged. So the phase inverter looked like it did because it was trying to correct the distortion happening in the power amp.

So the issue is definitely PT/OT/B+/power tubes related.

[big_teee]

Got any gut shot pics?
Sounds like a cool amp when you get it running.
My 2204 with Classic Tone Iron, runs 492v Loaded with 2-EL34s.
It is a very loud amp, I can imagine what the 2203 Would be like.
T

[loudthud]

I would measure the resistance of all the windings and give Weber a call. Have them compare them to a known good unit. No sense sending it back if you are just going to get the same thing.

[jamesmafyew]

Forgot to mention that I'm also using Weber multi-section cans and the Weber 9H choke. I usually use discrete caps with distributed filtering but I was trying to build this semi-quickly, mostly just for fun on the holiday break. It's for me, not for a client. The caps test good for ESR and no leakage. I would think if the caps were bad I'd be getting hum, but there's basically no hum to be found anywhere in this amp. Supply is working downstream of the B+ and all the nodes sink together. Bias supply comes off the B+ before the rectifier and it sinks too, which is part of why I suspect the PT.

Here are some gut shots. MV area is messy because I've set it up a few different ways trying to see if the bootstrapped circuit was causing my problems. It definitely wasn't, so I'll put it back later.

[jamesmafyew]

Here are all the stats, just in case someone has anything to compare them to:

WPT100 power transformer

Primary: 2.6 ohms
B+ winding: 128 ohms cold, 140 ohms when warm

(Mains voltage 122VAC)
Heater winding loaded: 6.33VAC
B+ taps unloaded: 361/361VAC
B+ taps idle: 349/348VAC
B+ taps full power: 322/323VAC

B+: 441V idle, 364V full power
Top of bias supply: -72V idle, -65V full power (tubes pin 5 are at -38V at idle)

Mains draw: 1.05A idle / 2.1A full power

[Steve Conner]

The level of sag you're seeing doesn't seem that excessive, considering the HT winding resistance is 140 ohms and the reflected primary resistance will add a fair bit again. Even with the supply sagging to 360V, you should be able to get almost 100W. Check your output stage wiring.

[überfuzz]

Off topic: It always interesting to see others builds. I usually keep my signal wires close to the chassis and the heaters up high. You did it quite the opposite here. Nice build!!

[jamesmafyew]

The output stage wiring is definitely correct. If the OT was bad, it wouldn't be clean at any power level, right?

In any case, "almost 100W" is not what I would expect from this rather large PT. I was getting 61W before it broke up in that weird asymmetrical way. If I wasn't using a scope and a dummy load, I might be playing the amp happily at home and not realizing anything was amiss.

The Weber 50W PTs I've used consistently are rock-solid and every "50W" amp I've built with them has done 60-70W before clipping, with idle B+ around 480V and full power around 430V. Weber says that the lower-voltage taps @ 640VAC should be able to get me to 480V or so idle B+, but I'm getting 441V idle with the 700VAC taps.

[Jazz P Bass]

[QUOTE=jamesmafyew;286379]The output stage wiring is definitely correct. If the OT was bad, it wouldn't be clean at any power level, right?
QUOTE]
Not necessarily true.
The OT may fail at any output level.
Monitor your mains current & see if it jumps when the output signal goes goofy.
For that matter, monitor the B+ & see if it dives at that point.
In my experience, if the output tube grids look good at full power & the output looks bad, (with known good tubes) then the OT is suspect.
I like to run each output tube all by itself.
Looking at mains idle draw & output signal.
Then I will switch the tube to the other half of the OT.
Of coarse the signal will look goofy, but I can get a general idea of the health of the tube & the OT.

[loudthud]

Have you tried different taps on the output transformer? Just try all three taps with the same dummy load and compute the power at clipping.

[jamesmafyew]

^ Yes, tried this. I'm using a 16-ohm dummy load. 8-ohm tap produced 50W, 4-ohm tap produced 30-some. Voltages measured on each tap of the secondary are correct in relation to one another -- i.e. the 4-ohm tap is half the RMS voltage of the 16-ohm tap, etc.

Monitor your mains current & see if it jumps when the output signal goes goofy.

It doesn't. Steady mains draw from 1A at idle up to a hair over 2A at full power. No jumps. The B+ sag is very smooth from idle down to where it sits at full power. Nothing sudden about it. It starts sagging as soon as there's any output at all, and drops consistently until it bottoms out. It's pretty much exactly what I would expect to see if there were about a 300-ohm resistor in series with the winding.

I've also clipped in some different filter caps and no change.

The goofy-looking clipping is strange to me. If someone posted the scope trace and that was all I had to look at, I would say that there was one tube of the four with a blown screen resistor, that's what it looks like. Except the output sine is clean and symmetrical up to a point, THEN it's like the negative-going half has its gain cut in half, as though one tube just gave up at that point. The positive-going side clips normally, just at a lower level than I would expect to see (33Vrms, instead of 40ish). There's a lot of crossover distortion that appears at the same time as the lower part starts to round off, even though the tubes are biased to 70% at idle. Of course, the B+ sags so much that they're not at 70% anymore at that lower B+. Because the bias supply comes off of half of the B+ winding, it sags 10% while the B+ itself sags 20%, because each half of the winding is contributing ~10% sag. So the bias point changes significantly as the supply droops.

In any case, I'm now out of town for Xmas, so will return to this after the new year.

[g1]

If someone posted the scope trace and that was all I had to look at, I would say that there was one tube of the four with a blown screen resistor, that's what it looks like. Except the output sine is clean and symmetrical up to a point, THEN it's like the negative-going half has its gain cut in half, as though one tube just gave up at that point.

If this was all I read, the PT would be one of my last suspects. Perhaps this is an unrelated issue?
Aside from that, have you tried biasing colder, and if so, how is the B+ affected?

[jamesmafyew]

It's possible it's unrelated, I've considered that. I've thought that I may actually have a bad OT, and that the power output would be more like 80-90W without the asymmetric clipping issue. Then it could be that the PT sags a lot but is still "normal," though I'd honestly find it disappointing if this much sag is normal for this PT. I shudder to think that both transformers could be bad.

I have tried biasing colder, but the B+ still sags and then the crossover notch is visible more of the time.

[big_teee]

It's possible it's unrelated, I've considered that. I've thought that I may actually have a bad OT, and that the power output would be more like 80-90W without the asymmetric clipping issue. Then it could be that the PT sags a lot but is still "normal," though I'd honestly find it disappointing if this much sag is normal for this PT. I shudder to think that both transformers could be bad.

I have tried biasing colder, but the B+ still sags and then the crossover notch is visible more of the time.

Good Morning Gang:
You might consider buying a set of these.
Marshall JCM800 100W Transformer Bundle w/ choke MADE IN USA
You could use them to test with, Classic tone Iron usually runs on the High side of the voltage scale.
If your Weber stuff is bad then you Get them to make it good.
If you don't fix the problem, use the new Iron for another project!
Happy Holidays.
T