Replacing a tube rectifier with solid state?

[wyattrg]

Hey all,

A friend and I are working on our second amp build and I was wondering about replacing the tube rectifier with solid state. We're building an AB763 Vibrolux.

Fender_vibrolux_ab763.pdf

I've read the Lazy Amp Tech tube rectifier thread and it suggests that this is ok given I add dropping resistors in series before the diodes I use for the rectifier.

I have a couple of concerns I wanted to ask before trying to attempt the math myself as I haven't done anything this involved since grade 12 physics (to be fair this is probably more involved than that was).

I've been comparing the schematic of the Vibrolux to the AB763 Bandmaster schem and the values are 320VAC which becomes 440V after rectification. There are 2 problems here, one being the transformer for the Vibrolux has a 325VAC secondary and the other being the B+ with the GZ34 is 415V. It looks like I'll have about 455V B+ if I just replace the GZ34 with diodes, so how should I go about dropping this to 415V? The inline dropping resistor was suggested in the forum I mentioned, but I'm not sure what value would make sense in this context (use ohms law?), or even if inline resistors make more sense than using a resistive attenuator circuit...

Resistive_attenuator.pdf

Another thing that concerns me is what to do with the 5V 3A secondary for the filament of the power tube. Should I just leave it disconnected? If anyone has any cool ideas for it that would be appreciated (I've seen some stuff about relays but I haven't read much about it). If not I think I'll just leave them disconnected.

I'd love to hear people's thoughts and advice, as I'm sure your experience will be insightful for my project and helpful in preventing me from lighting my new amp on fire

- Wyatt

[Raja_Kentut]

Hi Wyatt,

in my Bassman 5f6 I have also replaced the rectifier tube.
I used 120 Ohm in series with each rectifier diode.
The inductor TR2 was replaced by a resistor of 220 Ohm.
The first filter capactior after the diodes has 100uF instead of 20uF.
To stabilize the bias voltage I used a Z-diode.

https://ampgarage.com/forum/download/fi ... &mode=view

All voltages are in the allowed tolerance of +-20% (this high tolerance shows that it is not very critical to adjust exactly to the values in the schematic. Another thing to consider is, that modern digital voltmeters have a higher input resistance than the old ones. So the readings tend to be higher with modern measuring equipment. But this can be compensated.)
The 5V wires are isolated and safely fixed.
You have to experiment a bit with the series resistors and the TR2 repalcement resistor to find the optimum for your setup, but I think my values are a good starting piont.

[edit] one idea for the 5V :
Rectify it - that gives you appr. 7V DC.
Connect a 5V relais with a suitable series resistor.
Connect a elco parallel to the relais coil. This makes the relay switch with a short delay.
Connect one input lead of the power transformer via a n.o. contact of the relais (must be able to handle your input voltage and current)
Connect a 27ohm 5W resistor parallel to the n.o. contact of the relais. This reduces the inrush current of the transformer which protects the fuse from being blown by switch on, and also dampens the heater inrush current a little bit.

[Stevem]

First off my question is why are you doing this?

Tube rectifiers have no switching noise hash the gets into the audio stream, while SS diodes unless they are of the fast recovery type do.
You can also place a .01 600 volt ceramic cap across every regular SS diode you use and that will eliminate most of the hash.

[martin manning]

in my Bassman 5f6 I have also replaced the rectifier tube.
I used 120 Ohm in series with each rectifier diode.
The inductor TR2 was replaced by a resistor of 220 Ohm.
The first filter capactior after the diodes has 100uF instead of 20uF.
To stabilize the bias voltage I used a Z-diode.

Are you familiar with this experiment where the 5F6-A power supply was "cleaned up" with the result being that the amp lost its character?
https://www.ampbooks.com/mobile/classic ... AB-ripple/
A large reservoir will also change the transient response and therefore the "feel" of the amp, and I would think that regulating the bias voltage while leaving the B+ unregulated will result in variation of the output tube operating point as mains voltage fluctuates. The +/-20% on voltages quoted on the schematic is pretty wide, and I think that may have had more to do with mains voltage variation in the 1950's than component the tolerances. Here in the North America, the target mains voltage variation is now 120 +/-6%.

[Raja_Kentut]

yes, changes in the circuit do change something!

But using diodes instead of tubes makes life easier and … less expensive
The question for me is only : Am I or the person I make the Amp for is satisfied with the result ?
To 100% copy isn‘t possible anyway and for me not desirable, I add my personal touch to my builds.
If you want to replicate the model 100%, it's a different approach, which is also completely ok, but I personally lack a bit of personal creativity..
I take the „originals“ as a starting point and I‘m sure that there is still enough of heritage in the sound and feel of the amp.

[martin manning]

I did not mean to imply that replacing a vacuum rectifier with silicon and series resistors is going to make much difference, and there is no denying the cost and reliability advantages!

[Raja_Kentut]

I really didn't mean my comment as an attack or in any way with a negative intention!
Please don't forget that English is not my mother tongue, I don't want to come across as negative, but maybe I don't always have the right intonation.
Sorry if that was the impression!

[xtian]

First off my question is why are you doing this?

Tube rectifiers have no switching noise hash the gets into the audio stream, while SS diodes unless they are of the fast recovery type do.
You can also place a .01 600 volt ceramic cap across every regular SS diode you use and that will eliminate most of the hash.

I've read this, about switching noise from diode rectifiers, but still I used 1N4007s with no snubber caps in many, many builds and never heard any related noise.

[pdf64]

… To stabilize the bias voltage I used a Z-diode.

https://ampgarage.com/forum/download/fi ... &mode=view
…

I can’t see a zener in the bias supply?

Anyway, my thinking is that it’d be a bad idea to regulate the bias voltage.
It seems better to allow the bias and HT supplies to track the mains voltage.
If bias is regulated then the HT, or at least the screen grid supply node, voltage should be too.

[R.G.]

Tube rectifiers have no switching noise hash the gets into the audio stream, while SS diodes unless they are of the fast recovery type do.
You can also place a .01 600 volt ceramic cap across every regular SS diode you use and that will eliminate most of the hash.

I've read this, about switching noise from diode rectifiers, but still I used 1N4007s with no snubber caps in many, many builds and never heard any related noise.

It's one of those things where sometimes you get away with it, sometimes not.

The actual syndrome is, near as I can tell, the abrupt and delayed slam-off of normal silicon rectifiers can excite the parasitic resonances of the wires connected to the rectifiers to ring at RF. The twice-line-frequency bursts of RF can be picked up on the wiring or components inside the amp and be RF detected into a buzzy 2X-mains hum.
As to why you never heard any related noise, well, the common factor there is you -
- maybe your wiring practice is just that good, and your amps don't pick up rectifier RF bursts
- maybe your chassis layout is good enough to keep the RF far enough away from sensitive wires
- maybe some of both of those reduced the problem to inaudability
- maybe you're using a better grade of 1N4007s, or maybe 1N4007s have just gotten much better these days.

[Raja_Kentut]

… To stabilize the bias voltage I used a Z-diode.

https://ampgarage.com/forum/download/fi ... &mode=view
…

I can’t see a zener in the bias supply?

Anyway, my thinking is that it’d be a bad idea to regulate the bias voltage.
It seems better to allow the bias and HT supplies to track the mains voltage.
If bias is regulated then the HT, or at least the screen grid supply node, voltage should be too.

obviously I forgot it in the schematic, its from the right side of the 15k resistor to ground.
In fact it does not really stabilize the voltage, but limits it under no load condition. I noticed that it drifts up over time...I didnt check how high it would go at most, but it may would reduce the kathode currents too much that happens with a long time constant so that the output would become louder when start playing. I should reduce the filter caps for the bias...