Zero Feedback Impedance Amplifiers

[Susan-Parker]

Hi, thanks for the discussion.

Originally Posted by imfree707

Decades ago, I have used power transformer cores for audio, not super hi-fi, in PA amps, car stereo power amps, and other general purpose audio power amps, with fairly good results. I'd be interested in some discussion regarding performance limitations expected by use of power transformer iron in audio use. Thanks.

I would note in addition to my other posts that M6 is a mains/line frequency power transformer material. It's just a better quality than the grunt iron many manufactures use for the cheapest possible cost of build materials.

Nowadays with energy efficiency and being green a more significant factor a lot of this is going away (being replaced usually with switchers).

So basic but good quality M6 is fine, particularly in my circuits with the power mosfet followers where the driving impedance is very low.

M6 is about 0.015" or 0.37mm (ish), measuring with my vernier calliper.

Also I would mention that my very first experiments were with a mains double c-core transformer. I only got a watt or two out of that and the big issue was that the windings between nominally the same voltages were consecutively wound, so differed in total wire length, therefore DC resistance and also to some degree inductance were also different.

The core material itself was (in hindsight) fine. But the DC offset from the above became difficult to balance.

Also the transformer was heavily varnished, so appeared to me to be neigh on impossible to disassemble into usable components.

I then tried a pair of big Audio Note valve transformers. Better sound, still only a couple of watts.

Then found 200VA transformer kits from RS, which I used for my Zeus 35 amps.

Then moved up to my current EI120 transformer design.

Thanks again for you posting.

Best wishes,
Susan.

[imfree707]

Originally Posted by Susan-Parker

Hi, thanks for the discussion.



I would note in addition to my other posts that M6 is a mains/line frequency power transformer material. It's just a better quality than the grunt iron many manufactures use for the cheapest possible cost of build materials.

...snipped...

Thanks again for you posting.

Best wishes,
Susan.

Thanks Susan. I would usually select a transformer with thin laminations from variable load service, such as a 200 VA or so transformer from an old Centronics 301 printer and then completely rewind it for my desired ratios, with center tapped windings done bifilar, as I did in a 100 Watt PA amp I built over 30 years ago.

The amp had decent sound and would do full power from a 13.8 volt supply and drive 2,4,8 ohm and 70 Volt loads, thanks to a real output transformer! It's on the back burner for a 21st century upgrade in the next year or two.

Photos: 100 Watt PA amp and 120 W + 120 W car stereo amp, both using power transformer cores in audio service.

[Sandor]

Originally Posted by Susan-Parker

Dear All,

An up to date schematic using depletion mosfets as a quick overview of the full audio path from line level input to speaker drive:

Hello Susan,
I am going to build your Zeus amp and line driver.
Are those transformers the Sowter models suggested in your web site?
Are those depletion mosfets the best choice for your amps?
Thanks a lot.

Paul

[Susan-Parker]

Hi Paul,

Originally Posted by Sandor

Hello Susan,
I am going to build your Zeus amp and line driver.

Are those transformers the Sowter models suggested in your web site?

Yes, the stepped attenuator is recommended to be a Sowter 1035 Attenuator (with 6 dB Gain and Balance control. 48 dB in 2 dB steps plus 4 x 0.5 dB primary balance taps).

The SE line driver is the same size as the PP output, however it is gapped to allow for the 500mA biasing needed for SE operation.

Are those depletion mosfets the best choice for your amps?
Thanks a lot, Paul

I personally prefer the depletion mosfets, since they auto bias and result in the most minimalist design (if that slight tautology makes sense).

However most mosfet types work fine, one does need to take care of the extra bias supplies though.

The Silicon Carbide types are seemingly popular in some quarters, however they are "reassuringly" expensive. I would recommend using "industrial" strength mosfets to start with whilst you get everything up and running. That way, if there are any Oops moments, there is hopefully not too much pain.

My recommendation here is to use an extra mosfet of the same type to set the bias base level (which will be only just starting to conduct for the other mosfet(s) then uses diodes or a zener to get the needed extra voltage) and this will temperature track keeping the overall bias to the same(ish) point (the bias setting is quite sensitive down to the few mV level so it is worth doing).

The PP power stage will work well even with quite a lot of ripple on the supply, although it's nice to have something relatively quiet.

The SE line driver will need separate good quality voltage regulation. I use a mosfet with a zener to pre-regulate to a nominal voltage, then a proper regulator to finesse it off. That way the mosfet dumps most of the dropping watts, and the voltage reg can get on with giving a nice output without having to deal with a big differential voltage.

Please let us know how you get on. Always nice to hear from people who are building (or have built) one of these systems.

Best wishes,
Susan.

[Sandor]

Hi Susan,
thank you for answering, very helpful.
Given the difficulty in finding a source for those IXTH20N50D mosfets, should I look at some other depletion devices requiring low biasing? Could you provide any hints to select a device allowing auto-biasing?

I read the whole story in this amazing thread. Still, I did not well understand how to trim the resistor across the secondary of the input transformer of the power unit. Should it be placed in a specific point?

The Zeus monoblocks I'm going to build are meant to power my beloved Magneplanar MMG. Do you recommend a 4:1 or 2:1 configuration for the output transformer when connected to MMGs? As MMGs may be easily modded for biwiring, do you recommend feeding separately the tweeter panel and the midbass panel with the two secondaries?
The same experiment could be performed on my (also beloved) LS3/5A with AB1 subwoofers... Bass units require much more power than tweeter units. May this cause any problems?

Too much questions?
Thanks a lot.

Paul

[franzm]

What thermal behavior (Id, Vp vs temperature) have these depletion MOSFETs? NTC, PTC?

[Sandor]

I think I've found some (hopefully correct) answers by myself.

The correct ratio to feed a Magneplanar MMG (4 ohms) is 4:1, right?

The termination resistor on the secondary of the input transformer may be trimmed by injecting a square wave at the input and observing the output with a scope, right?

A possible substitute for IXTH20N50D must have an Id of about 1 A at Vgs=0, right? Sadly, I have found none, that depletion mosfet seems to be unique in that respect...
Susan, which negative bias could we expect from the DC resistance of the output transformer?
May we consider dual mosfet modules? Probably the do not need matching, a remarkable cost saving for a diy'er.