GZ34 after a pair of diodes and RC filter (Genz Benz Black Pearl guitar amp)?

[capacitorless]

I'm looking at the schematics of a Black Pearl amp. I'll attach a screenshot as well as the full schematic.

It uses a pair of diodes after the B+ secondary winding, into an RC filter, then into the GZ34. I've seen a decent number of tube amp schematics, but this design is new to me. Is there any advantage to putting the GZ34 after the initial filtering like Jeff did? Not questioning his design - the amp is built well and sounds good.

 

[Accelerator]

@capacitorless​

The Ceriatone Amplification does also this mod. to the most of their tube rectified amplifiers, as it is a good way to let the silicon diodes to makes “the hard job” of the rectification and then to use the tube rectifier to makes only the “sag” trick…

 

[capacitorless]

Thanks very much. That's also what I was contemplating, but I thought there might be something more to it.

Just a bit of background. I'm planning a number of mods to this amp. I need some more B+ and I'm thinking of bypassing the GZ34 entirely, as I don't need the sag. In a nutshell, I'm switching to 6973/6CZ5 power tubes, and adding gain stages to the preamp by reusing/repurposing the two original reverb tubes.

I would be cutting the trace between pin 8 and R130 to get rid of the 5VAC supply, removing the tube of course, and jumpering pin 4 to R130. I've got another idea to further increment B+ but I need to think that through.

I'll measure before/after, but I very vaguely remember 320vdc on the center tap of the OT primary ("plate" in the schematic) when working on another one of these. I would like to get to roughly 350v, but I only expect +10vdc when bypassing the rectifier in this circuit. It doesn't have to be perfect, but the closer the better

 

[rock soderstrom]

but I thought there might be something more to it.

I think the rectifier tube also works as a soft start device to slowly ramp up B+ when switching on to protect the power amp tubes.

But the whole thing is also new to me, normally I only know it as a "hybrid" rectifier where the diodes are directly in front of the tube.

but I only expect +10vdc when bypassing the rectifier in this circuit

Why only 10V? Have you measured how many volts are currently being dropped at the internal resistance of the 5AR4?
Shouldn't that be a bit more?

 

[capacitorless]

The soft start isn't an issue in this particular amp (standby switch), but yes I forgot about that function

I've been looking at rectifier voltage dropping charts to get the 10v (ish) value, but haven't done a measurement yet. I'll do that mod and measure before/after voltages, with the existing el84s in place. I know it has the least internal resistance of the commonly used rectifier tubes in those charts; I've seen 10v to 17v mentioned, but it depends upon the current draw, of course. Edit: I'll see if my meter will work to also get a current measurement at that voltage.

 

[merlin]

I've been looking at rectifier voltage dropping charts to get the 10v (ish) value

Those dropping charts are a bit of a red herring. Voltage drop is equal to the peak ripple current multiplied by the tube internal resistance, so voltage drop varies with current draw and cap value (when the tube is actually being used as the rectifier). A chart of resistance is a bit more realistic; about 50 to 60 ohms (per anode) for 5AR4.

 

[capacitorless]

In this case, the anodes are paralleled, so the resistance (and corresponding voltage drop) would be halved, correct?

 

[merlin]

In this case, the anodes are paralleled, so the resistance (and corresponding voltage drop) would be halved, correct?

Yes

 

[capacitorless]

Ok it looks like I can do this without cutting a trace - and will have more B+ available than expected. Node "A" (decoupled by C114) is marked 320v, so there is more upstream. I'll get the voltages with the GZ34 in place first.

The two PC boards are straightforward to remove, but I'm going to do the least amount of board modification as practical. If I need more juice (target: 350vdc) I'll interrupt the 5vac supply from the power transformer instead, and will jumper pin 4 to pin 8 on the rectifier socket - which conveniently means I can insert a DMM set to current measurement in those holes. Easy.

The power supply circuit board is well labeled. There are test points for B/screen and plate voltages. Nice black board, gold plated holes and test points. I'm comfortable in this situation; I have experience with much more challenging PC board based tube amps (Mesa Mk IIB, anyone?).

Edit: 332vdc coming in on pins 4 and 6 to the (now) Mullard GZ34. 323vdc on pin 8 going out to the center tap of the output transformer, so -9v dissipated by the GZ34. 317 at the plates, so the OT drops about 5v of B+. 318 at the input of the preamp section. Just to reason check the meter: it's an older Metec (Radio Shack era) but it has been solid, and double checking a few other measurements were as expected. So this looks normal and like the other one I wrenched on a few years back.

I don't know if the characteristics of an old school Mullard or Amperex GZ34 are worthy of the -9vdc B+, or if it really matters in this context. A 25 ohm power resistor would simulate sag, but I've heard there is something special about the old GZ34s.

Anyways I'm back to looking to add around 18v. maybe more - and not sure that's going to matter either, haha!

Edit2: The JJ broke it's center pin (more likely I fat fingered it). Fortunately there's clearance under the power supply circuit board to get the remnants out if something like that happens. And for that matter, removing the board isn't super difficult, just a handful of screws. Between that and one of the JJ EL84s cooking off, I'm going old stock on power supply and output tubes from this point forward.

 

[merlin]

If you've already got about 330V available, is there any point going to 350V? What are you hoping to acheive with a 6% change? You won't hear it!

 

[Cqwet Dbdfte]

The tube rectifier provides some slow startup.
Plus that tube imparts some magic.
Using a tube rectifier has some better recovery characteristics for AC , a SiC diode may be equal, but handle larger filter caps.

 

[Admiral-dk]

Anode Voltage of more than 340Vdc. on EL84's and they start to sound Hard and unpleasant .... the higher the worse ...!!!

Did quite a lot of experimenting on the subject back in early 1990's ....

Per

 

[capacitorless]

Anode Voltage of more than 340Vdc. on EL84's and they start to sound Hard and unpleasant .... the higher the worse ...!!!
Did quite a lot of experimenting on the subject back in early 1990's ....

Per

I agree completely. I'm going with something different: 6973/6CZ5 and the other tubes in that family. This is an experimental build too - would rather be trying 7591s but that's for another build

 

[capacitorless]

350v is supposedly a sweet spot for 6973 type tubes. I'm cool with something less, but want even more than 350v in part of the preamp section, which is based upon a Soldano SLO. The later stages of that call for around 378v. So just 350v out of the PT is a compromise already. If I have, say, 385v at the appropriate amount of current, I can get my desired preamp voltages and drop some volts for the power tube section. Edit: 395-400ish if going with a GZ34. Also, there are supposedly some cool overdrive characteristics of the 6973 family that I can dial in via biasing and B+.

I'll be adjusting the power tubes and PI via a VVR circuit, and have a separate dropping chain for the preamp. That's the idea, anyways, but we'll see.

 

[capacitorless]

Plus that tube imparts some magic.

@Cqwet Dbdfte do you mean specifically GZ34s? I've heard some comments about this, but it's been ages. I remember someone saying that the sag characteristics are non-linear, in a good way?

 

[Cqwet Dbdfte]

If an amplifier is for a musical instrument, i.e. a sound effects generator, perception is reality.
OTOH, if for sound REproduction any voltage sag would be an anomaly, and create intermodulation.
A perfect power supply would have no noise, or voltage variations regardless of load, reactive, or resistive.
I prefer output pentodes have their screen grids very well regulated, not just hang a resistance to +B, as to act like another little "plate". Screen regulation would make it more tolerant of +B variations.
In classic amplifiers you never see very tightly regulated power supplies because it would increase the size/cost of the product massively. Today with HV semiconductors it is a snap.

 

[merlin]

350v is supposedly a sweet spot for 6973 type tubes.

Don't believe everything you read.

I want even more than 350v in part of the preamp section, which is based upon a Soldano SLO.

The preamp is even less sensitive to changes in supply voltage than the power amp, I wouldn't waste my time fussing over it. Tweak the preamp bias instead for the sound you like, you'll always get it!

 

[capacitorless]

Any other comments re the GZ34 (specifically old Mullard) and tone/feel in a guitar amp would be appreciated. I know current rectifier tech is way better, and there are other ways to do sag (if I even end up wanting it) but I'm keeping my options open

That said, I'll probably reuse some organ iron that has both 5v and 6.3v windings, and use VVR to tune the output section B+ to taste. The VVR circuit will have a sag control, so that can be dialed in/out as well.

I'll be doing a straight-up SLO using the documented voltages and plate/cathode resistors, and I don't plan on tuning it. I'll likely be doing an alternate pentode stage that I will be tuning to taste, however.

 

[merlin]

Any other comments re the GZ34 (specifically old Mullard) and tone/feel in a guitar amp would be appreciated.

IMO the sound of tube recto sag is usually overstated. It's marginally noticeable with an old recto like the 5R4, but basically negligible with a GZ34 and definitely indistinguishable from a 50 ohm resistor. Heck the transformers have more resistance than that.

 

[capacitorless]

Thanks, that's great to know - those Mullards are pricey!