Rectifier Tubes Substitutes?
Any further thoughts on my final question?
Finally, I have read many times that the quality of the rectifier tube can have a big impact on sound quality. Our BEWITCH currently runs the original Chinese 5Z4P rectifier tubes - and gives no problems. But I do have a couple of nice 1940s Westinghouse Canada VU71/5U4G (NOS) tubes that I'm hoping might suit this amp - in substitution for the current 5Z4P tubes. Can anyone see any merit or problems with this substitution?
Any further thoughts on my final question?
Finally, I have read many times that the quality of the rectifier tube can have a big impact on sound quality. Our BEWITCH currently runs the original Chinese 5Z4P rectifier tubes - and gives no problems. But I do have a couple of nice 1940s Westinghouse Canada VU71/5U4G (NOS) tubes that I'm hoping might suit this amp - in substitution for the current 5Z4P tubes. Can anyone see any merit or problems with this substitution?
The VU71/5U4G takes 3A filament current against the 5Z4P taking 2A. I don't think your power transformer can handle the 2A extra.
Don't believe all you read on the Net. Use your own ears.
As for various rectifiers - save your money!
Interesting circuit. Looks like a balanced circuit for balanced input... But the top first triode has been grounded... It therefore does nothing useful for signal amplification.
Looking at the rest of the circuit, the bottom first triode is the input stage. This stage then feeds a long-tail phase splitter. Then the output tubes. This arrangement is sometimes referred to as a "Mullard circuit": https://www.sowter.co.uk/pdf/Mullard 5-10.pdf
If I had to redesign the amp, I would run the first double triode as SRPP, then direct couple the lower plate to the long tail. The other triode's grid the long tail would be capacitor coupled to ground. The output stage remains the same.
I have built a few amps with this configuration and they sound very good - especially with the output tubes in triode mode.
Good luck.
Looking at the rest of the circuit, the bottom first triode is the input stage. This stage then feeds a long-tail phase splitter. Then the output tubes. This arrangement is sometimes referred to as a "Mullard circuit": https://www.sowter.co.uk/pdf/Mullard 5-10.pdf
If I had to redesign the amp, I would run the first double triode as SRPP, then direct couple the lower plate to the long tail. The other triode's grid the long tail would be capacitor coupled to ground. The output stage remains the same.
I have built a few amps with this configuration and they sound very good - especially with the output tubes in triode mode.
Good luck.
Thanks petertub.
I already own the 1940s Westinghouse Canada VU71/5U4G (NOS) tubes, so no money to be saved.
I was planning to have a listen - "use my own ears" as you say - and decide for myself if different (correct) rectifier tubes impact the sound.
But it seems the VU71/5U4G tubes will draw too much current and potentially overheat the BEWITCH transformer; so not possible.
What would this achieve please?
(Most of our speakers dip to 4-ohms or a are 4-ohm nominal load; but I suspect this fact is irrelevant, for the purposes of the above comment...)
Thanks for these thoughts and encouragement.
I will look through the link you provided.
I'm still just learning about circuits, so your change suggestions are a bit above my pay-grade.
1. Use the proper output tap versus the loudspeakers you have (their real impedance versus frequency, and also according to the damping factor they need).
You may, or may not, get more or less power that way. But we are seeking Hi Fidelity Finesse, not Eardrum Splitting levels, Right?
(Well guitar amps are a different thing).
2. Then, use the tap you select above, and design the negative feedback part values to work from that tap.
That is True Global Negative Feedback from the Operating Output Tap that you are using.
Does that sound reasonable?
Comments Please?
You may, or may not, get more or less power that way. But we are seeking Hi Fidelity Finesse, not Eardrum Splitting levels, Right?
(Well guitar amps are a different thing).
2. Then, use the tap you select above, and design the negative feedback part values to work from that tap.
That is True Global Negative Feedback from the Operating Output Tap that you are using.
Does that sound reasonable?
Comments Please?
Last edited:
Yup. We always use the 4-Ohm tap for our 4-Ohm speakers. They are around 95dB efficient and get to ear-splitting levels on just 10-watts, so power is not an issue for us, with this BEWITCH 6550 amp.
In any event, we don't listen at stadium volumes...
SAFETY FIRST!!! From another thread about (a blown) BEWITCH 6550...
I found the following comment on another site's thread about the BEWITCH 6550 amp. In this instance, this particular (blown) BEWITCH amp had been running KT88 tubes; and a failed tube had caused a melt-down.
In the interests of safety, I am very keen for experienced others to comment on the truth of the following conclusion, drawn from another engineer, who evaluated this same BEWITCH circuit...
"Whilst it's okay for 6V6, EL84 and EL34 tubes to use a common cathode bias resistor, this should never be done with KT66 or KT88 type tubes. (Yes, it is a design fault)
The circuit needs to be modified so each output tube has it's own resistor and bypass capacitor. 560 ohm 10 watt to 25 watt with 100uf to 220uf 100V bypass cap for each will do."
I found the following comment on another site's thread about the BEWITCH 6550 amp. In this instance, this particular (blown) BEWITCH amp had been running KT88 tubes; and a failed tube had caused a melt-down.
In the interests of safety, I am very keen for experienced others to comment on the truth of the following conclusion, drawn from another engineer, who evaluated this same BEWITCH circuit...
"Whilst it's okay for 6V6, EL84 and EL34 tubes to use a common cathode bias resistor, this should never be done with KT66 or KT88 type tubes. (Yes, it is a design fault)
The circuit needs to be modified so each output tube has it's own resistor and bypass capacitor. 560 ohm 10 watt to 25 watt with 100uf to 220uf 100V bypass cap for each will do."
i bet they do....try it and find out, if the ac volts at the socket pins drop below 5 volts, say 4.5 then your power traffo can not hack it...to test is to believe...
a standard practice for me, even the 6bq5's get their own cathode bias resistors...
if anyone is impressed with any tube amp, then it should be because of the topology and the irons that were used....tubes are easy, just get the ones that you fancy....
The circuit needs to be modified so each output tube has it's own resistor and bypass
Thanks for confirming these thoughts, TonyTecson.
Would that be a hard thing for you (or someone else?) to detail on my schematic - for this newby?
If I have clear and accurate details on a schematic - of exactly what I need to change - I have a good chance at completing this task.
Thanks for confirming these thoughts, TonyTecson.
Would that be a hard thing for you (or someone else?) to detail on my schematic - for this newby?
If I have clear and accurate details on a schematic - of exactly what I need to change - I have a good chance at completing this task.
I agree with SONDEKNZ,
Output Tubes Burning Out?
1. 6550 Self Bias, Maximum Rg = 250k
But be conservative and easy on the tubes, consider using Individual Self Bias resistors,
do not use a shared self bias resistor.
The Bewitch 88 uses 220kRg, and a shared self bias resistor.
Use individual self bias resistors.
2. KT88 Self Bias, less than or equal to 35 Watts pa + pg2 dissipation, Maximum Rg = 470k
KT88 Self Bias, greater or equal to 35 Watts pa + pg2 dissipation, Maximum Rg = 220k Rg
But be conservative and easy on the tubes, consider using Individual Self Bias resistors,
do not use a shared self bias resistor.
The Bewitch 88 uses 220kRg, and a shared self bias resistor.
Use individual self bias resistors.
3. Then there is the advantage that using individual self bias resistors, and good tubes, the quiescent currents will closely match.
That is a plus for the push pull output transformer. Un-balanced currents = early saturation of the output transformer core.
Follow that rule, no matter the type of output tube. 2A3, 300B, 7591, 6L6, . . . etc.
Output Tubes Burning Out?
1. 6550 Self Bias, Maximum Rg = 250k
But be conservative and easy on the tubes, consider using Individual Self Bias resistors,
do not use a shared self bias resistor.
The Bewitch 88 uses 220kRg, and a shared self bias resistor.
Use individual self bias resistors.
2. KT88 Self Bias, less than or equal to 35 Watts pa + pg2 dissipation, Maximum Rg = 470k
KT88 Self Bias, greater or equal to 35 Watts pa + pg2 dissipation, Maximum Rg = 220k Rg
But be conservative and easy on the tubes, consider using Individual Self Bias resistors,
do not use a shared self bias resistor.
The Bewitch 88 uses 220kRg, and a shared self bias resistor.
Use individual self bias resistors.
3. Then there is the advantage that using individual self bias resistors, and good tubes, the quiescent currents will closely match.
That is a plus for the push pull output transformer. Un-balanced currents = early saturation of the output transformer core.
Follow that rule, no matter the type of output tube. 2A3, 300B, 7591, 6L6, . . . etc.
Last edited: