diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Tubes / Valves (http://www.diyaudio.com/forums/tubes-valves/)
-   -   Plitron Specialist transformer with Unity Coupling (http://www.diyaudio.com/forums/tubes-valves/25271-plitron-specialist-transformer-unity-coupling.html)

pedrinho 3rd January 2004 06:52 PM

Plitron Specialist transformer with Unity Coupling
 
Hi, I'm a Brazilian newcomer to this forum, and I'm stumbled with
the following question: Along with its Specialist series of output
transformer for tube amps, Plitron also has another model with
cathode feedback - the VDV1070-UC - that is unity coupled, i.e.,
works with the McIntosh design. I have Menno van der Veen's
book, and it does not say much about this transformer except for frequency range measurements (which, by the way, are the most impressive I've ever seen). Does anybody know about an amp design using McIntosh's unity coupling and the VDV1070-UC?
Has anyone tried to do so? Or shoud I try to contact Menno himself (sorry for the dumb question)?

Thanks everybody...

EC8010 3rd January 2004 09:43 PM

A unity-coupled output transformer allows you to make a superb output stage.

You now need a driver stage capable of supplying a very large voltage with low distortion. You have moved the problems from the output stage to the driver stage. This isn't a complete waste of effort, because the driver stage only needs to provide voltage, rather than power. Nevertheless, you either need a a driver HT of >700V or you do some rather dubious bootstrapping from the output transformer that seriously compromises stability.

Judging by your question, I suspect that you are not an experienced electronics engineer. Further, if this is your first valve project, I would suggest that a more conventional and slightly less ambitious project is more likely to reward you with good results. I have played with unity-coupled amplifiers, and was not convinced that they were worth the trouble. If I am wrong, and you are an experienced engineer, supported by lots of test equipment, give it a go, but expect HF stability or driver stage distortion to be the major problem.

Bandersnatch 3rd January 2004 11:18 PM

Hey-Hey!!!,
If you want to build with that output tx, study McInosh. There is no other reasonable way to achieve the results. I will second the recomendation for another project direction if this is nearly your first.
regards,
Douglas

SY 4th January 2004 12:34 AM

I'll third that- the Mac circuit uses a combination of negative and positive feedback. If you don't know exactly what you're doing, you'll come to grief.

If you're an advanced experimenter, I've had good luck with HV drive using cascoded diff amp, with a HV mosfet on top, tube on the bottom. I haven't used it with a Mac-type output stage, but it certainly can swing enough voltage if you've got a 1kV mosfet and an appropriate rail voltage.

A step up interstage transformer might be another interesting approach, though you'll probably have to restrict feedback to local loops.

EC8010 4th January 2004 12:47 AM

I've had enough...
 
Look, I've looked it up in the RS and Campagnolo catalogues, and I can't find a Tnuctipun, so what the ****ery is it?

Bandersnatch 4th January 2004 01:01 AM

Try Google then...
regards,
Douglas
What in the heck is RS? I do recognize Crapagnolo. D-A all the way!!!

Magura 4th January 2004 01:07 AM

Quote:

Originally posted by Bandersnatch

What in the heck is RS?


RS components.....a member of the syndicate of thieves (RS, Farnell and so on).

Magura

fdegrove 4th January 2004 01:21 AM

Hi,

Quote:

Try Google then...
.
.....To do whot exactly?

Quote:

RS components.....a member of the syndicate of thieves (RS, Farnell and so on).
And as Fela N'Kuti or whatever his name was said: ITT= International Thief Thief....

If you want to go down the socialist road come informed, otherwise show your best side.

In short, RS. Radio Spares is just another company trying to make a buck...Can they or do they get a veto?

Cheers,:rolleyes:..............?

pedrinho 4th January 2004 03:16 PM

Dear folks,

Thanks for the quick replies. I'll put my doubt into details following
each reply:

Quote:

Originally posted by EC8010
A unity-coupled output transformer allows you to make a superb output stage.

You now need a driver stage capable of supplying a very large voltage with low distortion. You have moved the problems from the output stage to the driver stage. This isn't a complete waste of effort, because the driver stage only needs to provide voltage, rather than power. Nevertheless, you either need a a driver HT of >700V or you do some rather dubious bootstrapping from the output transformer that seriously compromises stability.

Yeah, I've already checked in some places that the driver stage
is the real trick in McIntosh designs. The MC schematics I've given
a bird's eye view on are very tricky...

Quote:


Judging by your question, I suspect that you are not an experienced electronics engineer. Further, if this is your first valve project, I would suggest that a more conventional and slightly less ambitious project is more likely to reward you with good results. I have played with unity-coupled amplifiers, and was not convinced that they were worth the trouble. If I am wrong, and you are an experienced engineer, supported by lots of test equipment, give it a go, but expect HF stability or driver stage distortion to be the major problem.

You're right about that - I AM a newcomer, and with my
scarce budget coming from my PhD scholarship it's difficult
to acquire testing equip such as oscilloscope, signal generator,
spectrum analyser and so on... I'm having to rely on a fellow engineer (who, by the way, is not exactly enthusiastic about valves) to do the testing. I'm just evaluating if it's worth the effort
to invest $300 in a VDV1070-UC transformer and stick to a
McIntosh-like design, or to be down-to-earth and use a,
say, van der Veen design such as one of those in his book.
And, yes, it IS my first project.

Quote:

Originally posted by Bandersnatch
Hey-Hey!!!,
If you want to build with that output tx, study McIntosh. There is no other reasonable way to achieve the results. I will second the recomendation for another project direction if this is nearly your first.
regards,
Douglas

Do you mean Frank McIntosh's original articles? I'd like to have
precise references, if possible.

Quote:

Originally posted by SY
I'll third that- the Mac circuit uses a combination of negative and positive feedback. If you don't know exactly what you're doing, you'll come to grief.

If you're an advanced experimenter, I've had good luck with HV drive using cascoded diff amp, with a HV mosfet on top, tube on the bottom. I haven't used it with a Mac-type output stage, but it certainly can swing enough voltage if you've got a 1kV mosfet and an appropriate rail voltage.

A step up interstage transformer might be another interesting approach, though you'll probably have to restrict feedback to local loops.

Actually, this leads me to another aspect of my dilemma exposed above: I'd like to make experiments with three types of global
feedback: voltage, current and power. I don't know if it's possible
to do the last two if I use a McIntosh-like design, again because of the complexity of the schematics I've seen so far - will I have other instability problems besides those coming from bootstrapping (i.e., more LC loops, etc)? I was planning to do such experiments only if I'd chosen a barebone VDV design,
specially regarding power feedback loops such as the one
used in Stephen Fay White's POWRTRON design, which are
considerably more complex than bare voltage or current feedback.

Well, I've stalled long enough... Thanks all!

EC8010 5th January 2004 11:43 AM

Being unfamiliar with the Powrtron circuit, I waded through the rambling and turgid documents at:

http://www.cosmos2000.org/audio/powrtron1.htm

Fortunately, the original article ("Audio Anthology" Vol 3 pp13) was slightly clearer and comes clean about the feedback, confirming that it uses a combination of negative current and voltage feedback to achieve constant power.

Unfortunately, modern loudspeaker designers assume a constant voltage source, so supplying a loudspeaker passive crossover and drivers from a constant power source is likely to disturb the operation of the crossover and the bass alignment of the woofer. However, the Powrtron amplifier was intended for use with an active crossover, and would have provided HF equalisation to its associated drivers due to maintaining constant power across the rising impedance due to voice coil inductance. This might be useful, but that equalisation could just as easily be implemented by adding a step network at the input of the amplifier.

Until Thiele & Small came along in the mid-70s, the dynamics of loudspeaker bass alignment were very poorly understood, so a constant power source might have seemed appropriate. The effect of a constant power source on a modern loudspeaker would be to produce a change in frequency response proportional to the curve of loudspeaker impedance, exacerbating the effects of the LF resonance.

Simply because the loudspeaker is driven constant power does not mean that its acoustic power response has been corrected. The acoustic power response is the summation of responses at all angles, and is primarily affected by the relative size of the source to the wavelength being reproduced. Nevertheless, driving each loudspeaker driver from a dedicated amplifier is a good idea because it allows amplifiers to be designed and optimised for known loads that only have to be driven over a restricted bandwidth.

I see no particular reason why a McIntosh amplifier should be any more unstable using current or power feedback than when using voltage feedback. However, each time you change the feedback, you are going to have to go round the loop of making the amplifier stable again, which is trickier with a McIntosh.

Experiment with different forms of feedback, but be aware of the effect they will have. If you can borrow an oscilloscope and oscillator, you can measure the voltage response at the loudspeaker terminals with each change in feedback to assist in your deliberations. You will probably need to buy your fellow engineer lots of beer to pay for all the testing...

I don't want to be a wet blanket, but this will be quite a lot of work, probably for little gain. Complexity does not automatically equal quality.


All times are GMT. The time now is 02:30 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2