Hi all.
Quick question that has likely been posed before:
If building a push pull DHT amp, e.g. Pete Millett 6A3 mono blocks, and using a single tentlab regulator per power tube, is there any benefit to having separate windings per regulator or can a single high-current winding feed both regulators without any significant penalties?
Thanks!
Quick question that has likely been posed before:
If building a push pull DHT amp, e.g. Pete Millett 6A3 mono blocks, and using a single tentlab regulator per power tube, is there any benefit to having separate windings per regulator or can a single high-current winding feed both regulators without any significant penalties?
Thanks!
Back in the 50s we didn't have nor need filament regulators. Why need them now?
Push pull cathodes can be coupled together and use the same filament supply.
Push pull cathodes can be coupled together and use the same filament supply.
Oops, I hit 'like' instead of 'reply' (is there anyway to reverse this?)
Anyway, the answer to your question is under 'applications' at: http://www.tentlabs.com/Products/Tubeamp/Tubefilament/index.html
For Push Pull:
With DHT, if you want to use Individual Self Bias for each tube, you need separate filament windings, and separate regulators.
Common self bias, and fixed bias, and fixed adjustable bias do should not require individual separate filament secondaries.
But with common self bias, very well matched tubes are required.
With fixed bias and fixed adjustable bias, either individual bias circuits for each tube, or very well matched tubes are required.
Just my opinions
With DHT, if you want to use Individual Self Bias for each tube, you need separate filament windings, and separate regulators.
Common self bias, and fixed bias, and fixed adjustable bias do should not require individual separate filament secondaries.
But with common self bias, very well matched tubes are required.
With fixed bias and fixed adjustable bias, either individual bias circuits for each tube, or very well matched tubes are required.
Just my opinions
You guys get caught up in your own ego - the answer is in the data sheet (equiv) as linked above, specific for the OP application.
All the rest is 'apple polishing', and irrelevant to the OP question.
All the rest is 'apple polishing', and irrelevant to the OP question.
ISCMMS,
Irelevant?
Tent labs filament supplies, Great!
Tent labs also has a regulated bias supply for those who want to use fixed bias supplies. Great!
However . . .
1. I saw no mention of using very well matched tubes for push pull, and I saw no recommendation to use different bias voltages to match the tube currents. (I probably did not read well enough).
2. The feature of the Tent labs regulated bias is that the bias voltage remains constant with varying power mains voltage.
But a new problem occurs when the power mains voltage varies: Both for triodes, and Pentodes/Beam Power tubes, the unregulated plate B+ and unregulated screen B+ will cause the quiescent plate and quiescent screen currents to vary.
If instead, the plate B+, screen B+, and fixed bias are All unregulated, they tend to track very closely (think %), even when the power mains varies widely.
But I did not see any recommendation to use regulated plate B+ and regulated screen B+ when using the Tent labs bias supply.
(I probably did not read well enough).
Failure to deal with all the aspects of an amplifier design, may cause less than optimal performance of the amplifier.
Irrelevant?
Just my experience, and my opinions.
Irelevant?
Tent labs filament supplies, Great!
Tent labs also has a regulated bias supply for those who want to use fixed bias supplies. Great!
However . . .
1. I saw no mention of using very well matched tubes for push pull, and I saw no recommendation to use different bias voltages to match the tube currents. (I probably did not read well enough).
2. The feature of the Tent labs regulated bias is that the bias voltage remains constant with varying power mains voltage.
But a new problem occurs when the power mains voltage varies: Both for triodes, and Pentodes/Beam Power tubes, the unregulated plate B+ and unregulated screen B+ will cause the quiescent plate and quiescent screen currents to vary.
If instead, the plate B+, screen B+, and fixed bias are All unregulated, they tend to track very closely (think %), even when the power mains varies widely.
But I did not see any recommendation to use regulated plate B+ and regulated screen B+ when using the Tent labs bias supply.
(I probably did not read well enough).
Failure to deal with all the aspects of an amplifier design, may cause less than optimal performance of the amplifier.
Irrelevant?
Just my experience, and my opinions.
I know the tentlabs schematics, and it depends on the time constant of the integrator stage whether mains fluctuations ripple through in a different bias operating point, the integrator will compensate for relatively slow variations in mains voltage quite nicely, Most servomechanisms will.
You are on point with regards to stabilized voltages for optimum performance, most pentode or triode mode amps will benefit from some sort of G2 stabilisation, Anode voltage stabilisation is optional as the curves are quite flat.
Topic poster will may want to look for a transformer that has a primary/secondary shield winding as this will reduce transmitted noise by quite a bit.
Also take care not to increase the input voltage of the tentlabs modules over 7.5V~ as this will entail increased dissipation for the pass element. they can work from 6.3VAC to 6.3VDC but i would suggest a 7V transformer instead. As 6.3V~ will give Approx. 7.5V under ripple rectified and the dropout of the FET is about 200mV you only have 1.2V headroom, which can be too low if the mains goes -10%
Given the relatively high buffer capacitance of the tentlabs modules, you may want to spec 2-2.5A per winding to make sure the transformer runs cool. Otherwise the charging peaks on the filter capacitors may cause the transformer to run hot.
You are on point with regards to stabilized voltages for optimum performance, most pentode or triode mode amps will benefit from some sort of G2 stabilisation, Anode voltage stabilisation is optional as the curves are quite flat.
Topic poster will may want to look for a transformer that has a primary/secondary shield winding as this will reduce transmitted noise by quite a bit.
Also take care not to increase the input voltage of the tentlabs modules over 7.5V~ as this will entail increased dissipation for the pass element. they can work from 6.3VAC to 6.3VDC but i would suggest a 7V transformer instead. As 6.3V~ will give Approx. 7.5V under ripple rectified and the dropout of the FET is about 200mV you only have 1.2V headroom, which can be too low if the mains goes -10%
Given the relatively high buffer capacitance of the tentlabs modules, you may want to spec 2-2.5A per winding to make sure the transformer runs cool. Otherwise the charging peaks on the filter capacitors may cause the transformer to run hot.
Thanks for the input all. Very much appreciated. This project will likely take me yonks to get finished. I need to have a thought about whether individual self bias is worth the additional expense and complication. Any suggestions on where to find 7V transformers? Perhaps a high-current 6.3V secondary would run high enough for sufficient headroom if only lightly loaded.
Apologies. Had a bad day and shouldn't have posted.
Regards.
ISCMMS,
It is OK. Keep posting.
I have made many mistakes on the Tubes / Valves forum.
I hope I am getting better at not doing that, but now the older I get, the more unsure I am of the accuracy of my statements.
So many of us learn so much and are entertained as well.
It is OK. Keep posting.
I have made many mistakes on the Tubes / Valves forum.
I hope I am getting better at not doing that, but now the older I get, the more unsure I am of the accuracy of my statements.
So many of us learn so much and are entertained as well.
IMO, no, it's not worth the trouble. I would build it as drawn. It's elegantly simple--the shared filament supply for all tubes raises the 6SN7 heaters by c. 45 volts, which is exactly right for this design. Provided the output tubes are relatively well-matched, filament noise is cancelled at the output. I'd say, trust Pete on this one. ;-)