Faking a 5AR4/GZ34

[R.G.]

This is an outgrowth of the guitar amp wiring notes.
I came up with a network of diodes, zeners, and resistors that emulates the I-V curve of a 5AR4 tube.

Fake 5AR4 model.png

It matches very well in simulation. I am dealing with some family matters that prevent me from building this up, so I'd appreciate any real world feedback. I think it will work well, as it produces a spectrum of the ripple voltage in simulation nearly identical to a 5AR4 model.
The choice of 1N5408 versus APT - FRED diodes may not matter much. The 1nF/1K snubber may well stop the transformer parasitics from ringing with a normal diode.

[martin manning]

I came up with something similar to mimic obsolete SiC Varistors for a Magnatone vibrato. It did not work in the audio path due to switching noise, which might not be an issue here. What did work pretty well was using modern ZnO varistors with specific turn-on voltages. https://ampgarage.com/forum/viewtopic.p ... 42#p370542
For this application there might not be anything wrong with a simple sag resistor.

[R.G.]

Yeah, I have been happy with just a resistor. As I mentioned, this is a side effect.

I was attracted by a shiny thing (faking a vacuum rectifier) when doing some more on the wiring notes. I decided to see just how the ripple voltage harmonics of a vacuum rectifier and a diode-resistor compared. The diode-resistor has more harmonic content in the ripple voltage - big surprise. Speculating that this might be a way to explain the Golden Ears who can hear the difference in power supply rectifiers beyond sag effects, I fired up the simulator. I tried a diode-zener-resistor added to the diode-resistor, and ...
Well, some hours later, I have this circuit. It is remarkably like a 5AR4 in ripple harmonics, and in time domain current pulses and ripple voltage. So close that it's hard to see the differences.

So ...IF... someone was fussy about hearing rectifiers, this might be an interesting thing to look at. It kind of fell off the truck on the way to other places.

[trobbins]

Yeh I'd also recommend looking further at the use of MOV's in lieu of zeners as Martin suggests, as they should 'blur' the transition between parallel arms and avoid any stepping characteristic. Whilst zeners likely have a tighter tolerance on operating knee, Zeners and MOV's each have their own SOA characteristic. For MOV's, the average power dissipation is important as it marks the operating boundary below which no device degradation occurs in repetitive pulse operation, and Harris indicates that is circa is about 0.25W for 7mm disk, 0.4W for 10mmD; 0.6W for 14mmD, and 1W for 20mmD.

Typical MOV's are bi-directional, although I don't think that matters in your application R.G., and I'm not sure why you inserted a diode in each arm - was it for belts and braces protection of some kind?

A 5AR4/GZ34 has a higher PIV rating than 1kV, so perhaps worth making the proviso of this being limited to circa 330-0-330V applications, or otherwise needing 2x series 1N5408 as well as commenting the cap as a 1kVac rated part (as some could assume it is a 1kVdc part). I'd suggest just using UF5408, or even UF4007 if the power output did not need the higher current capability.

The packaging of all parts may need some comment about keeping zeners sufficiently spaced from resistors that may get significantly hot.

[R.G.]

Hmmm. Yeah, probably ought to use more diodes.

Of course, using all 1kV diodes for all six means that (quite by accident!) it's already good for 2kV as long as leakages balance. The diodes in the four arms were indeed for belt and suspenders purposes to start. I'll think about what happens if they're taken out. It bothers me a bit to do that without thinking more.

[Stevem]

How you folks rate this.

This is from the late Gerald Weber.

He states that this give the feel of a 5AR4 with a 5U4, so still some sag when loaded high enough.

The rectifiers carry most of the load so a 5U4 should last forever.
The added heat could be a concern I suppose if the 25 watters had to be in a tight chassis.

[R.G.]

I posted this whole concept to the Music Electronics Forum, too. That's where some of the comments I referred to come from. Great forum - give it a try.

I have some suggestions for additional sims from there. I'll add this one in. What else?