200- 500 watt Tube Amp project

[LinuksGuru]

Originally Posted by prairiemystic

I'm not sold on toroids being that much better, based on "flux efficiency".
The 1650W is 0.1lbs/watt and Sowter is about 0.24lbs/watt. The PAT 4141 is 0.0588lbs/watt. I think that's about 24%-58% less and I'd guess less iron is due to running higher flux density.
The Plitron PAT 4141-00 has no specs on distortion and DC imbalance. It's easy to fudge frequency response curves when no operating level and source/load impedance is mentioned. I can get 117kHz out of anything at low levels with low Z drive. So I find it strange to have all this glam yet missing some of the real measurements.

Toroids can be much better and may be not, depends upon construction.
Remember, with push-pull amp higher flux density means higher 3rd harmonic distortions, with 1.5T it can be as high as 5-8% (varies upon particular core material samples), it applies to EI, double C-core and toroids.
For high-quality devices flux density should be kept below 8000 - 9000 G.
DC imbalance is another plague for PP audio transformers, simplified methods of design have no way to evaluate its impact on LF response and THD/IMD. Most trafos sold as "black boxes" with frequency response measured at 1W, which is meaningless at RMS power and DC imbalance always present at some degree.

PS. Due to the specific shape of toroid winding them for audio is quite cumbersome, a Chinese factory I tried to order custom toroids from (of my own design) rejected them, stating that windings/sections layout is too complex.

[Zero Cool]

I am currently building a set of 200 watt 4 chassis mono block circlo-atma-clones. most of the parts have been easy to source with the exception of the monster sized heater transformers. to get 200 watts 20 6AS7G's are required and each has a 2.5amp filament current draw. multiply that by 20 tubes and we need 50+ amps of 6.3V!!!

So building a Big amp is going to require big $$$ no matter which way you slice it. you can use big tubes which require big iron OPT's at a big $$ price or you can use lots of tubes which require big iron for heaters at a big $$ price! in reality the heater transformers are probably cheaper than the pair of big opt's needed.

I have managed to collect all the parts needed for my amps but i am hung up on the cost of the filament transformers at the moment. just don't have the spare cash right now.

zc

[LinuksGuru]

Originally Posted by Michael Koster

It totally bugs me when products are sold without good specs or test reports. That alone keeps me away from certain products...

However, the core material should make a difference.

Indeed.

I believe that the core material used in the toroid may have a higher saturation flux density and may be operated at higher V/turn.

Yes, that's correct. However, higher flux density in trafos used in PP amps means higher 3rd order harmonic distortions.

There also needs to be enough core cross-section to handle the power.

That IS the problem with the almost all toroid cores. Their geometry optimized for 50/60/400 Hz, not 20 Hz.

Originally Posted by Michael Koster

I think it's possible for the toroid to equal the power performance of an EI iron core at less weight if these conditions are met.

40 - 80W toroid, under right design & build circumstances, CAN outperform EI and double C-core, by a wide margin in all aspects (except cost price may be).

I don't know about the small signal performance; could be better due to more flux at lower levels, or could be worse due the core material having higher remanance. Probably pretty good as long as there is zero DC bias.

DC BIAS mostly affects LF performance. Small signal performance depends upon design of course and core permeability at very low flux.

I have tried to get a DC current spec from a few transformer makers and for these toroids the number I get is usually zero; not <1mA, not <1 uA, not <1pA, but zero...

Some Russian transformer design books I have (published back in 195x/196x) stated that output transformer for push-pull amplifier must be able to handle as much as 10% imbalance of idle current imbalance of the output tubes. For typical PP amp it is about 5 - 10 mA.

But there are toroids that handle DC current, but using so-called "distributed gap" toroid cores. Apparently there is a dielectric material compounded in with the core material which has the effect of a magnetic gap. I've never tried one, but you see them in flyback switchers and used as filter chokes.

Micro-gapped core is one of the solutions, and definitely not the best. Any cuts in core material is a source of additional distortions. DC current imbalance in PP amps is very small, in properly working amp <= 10 mA.
So small micro-gap is not a problem to make with EI or double-C core, but very problematic with toroids. No ordinary industrial band-saw will be up to the task.

Maybe someone could test the Plitrons and write a report.

It would be really great! I'm somewhat reluctant to throw away 600 EUROs from my pocketbook just for the sake of cool experiments.

[LinuksGuru]

Originally Posted by Zero Cool

I am currently building a set of 200 watt 4 chassis mono block circlo-atma-clones. most of the parts have been easy to source with the exception of the monster sized heater transformers. to get 200 watts 20 6AS7G's are required and each has a 2.5amp filament current draw. multiply that by 20 tubes and we need 50+ amps of 6.3V!!!
zc

IMHO it would be much better to connect heaters of 4 output tubes in series, thus, output voltage should be 25.2V and amperage requirements will be reduced by a factor of 4. This will still require custom-build transformer, but much simpler and cheaper !

[atmasphere]

I did several OTL prototypes that used a series filament connection and ran them off the AC line with a dropping resistor. As a prototype it worked well and in 30 years had no problems with internal shorts in the tubes that could have been able to do serious damage! Not something I recommend though.

[LinuksGuru]

Originally Posted by atmasphere

..... internal shorts in the tubes that could have been able to do serious damage! Not something I recommend though.

A fuse over each loop will protect against overheating of filaments if one of tube get shorts.

[Zero Cool]

Originally Posted by atmasphere

I did several OTL prototypes that used a series filament connection and ran them off the AC line with a dropping resistor. As a prototype it worked well and in 30 years had no problems with internal shorts in the tubes that could have been able to do serious damage! Not something I recommend though.

I have owned 2 of Ralphs prototypes with the series filaments. My father ran the first one, the 30yr old model every day for a year without issues and he now has the 2nd prototype which is about 20yrs old i think Ralph said and he has run that amp now everyday for i think close to 2 years now without fault! he loves it!

I have considered maybe running tubes in Pairs at 13.2V with separate transformers for each pair as smaller transformers are easier to find and easier to hide under a chassis.

[Allen Wright]

Something no one has mentioned, or at least given the weighting it deserves: HEAT.

Tube amps that you would want to listen to are AT BEST 50% efficient, so consider 2kW into the room anytime they are all running.

So amps of the number and power requested by the OP will COOK him and all his room partners. In mid winter this may be welcome if you live in the Arctic, but come summer and the air con bill will be equal to the power bill for the amps = serious.

Or it will be HOT and very likely VERY uncomfortable.

Regards, Allen

[Zero Cool]

Circlo-Atma-Clone's in winter here in Mn will be NICE!

[tubelab.com]

Something no one has mentioned, or at least given the weighting it deserves: HEAT.

It was discussed (maybe over discussed) in the thread that begat this one.

Monstrous 1kW amp!

Trust me heat is one of my biggest considerations here in south Florida where 30 degrees C is the norm in the winter except for a few cold days. I don't use my 40 WPC SE amp because it heats up the room faster than the air conditioning can cool it. I do plan to build something in the 200 to 500WPC range, but I am in no hurry, and more efficiency improvement experiments will be performed before the final design is chosen. I really liked the sound of the red board at the 150 WPC level, through the big Plitron toroids.

Maybe someone could test the Plitrons and write a report......It would be really great! I'm somewhat reluctant to throw away 600 EUROs from my pocketbook just for the sake of cool experiments.

I have a pair of big Plitron toroids that are rated for "400 watts at 20 Hz" according to the label on the transformer itself. I got them from Plitrons surplus page. I have used them at the 200 WPC power level and can say that they are the best sounding P-P OPT's that I have but my collection isn't that big. I can also state that the THD measurements are almost constant across 20 Hz to 20 KHz at 100 watts too! My FFT analyzer is still dead. I collected a bunch of big sweep tubes at the Dayton hamfest so soon I plan to see how much power the red board will make when wired as a mono block. Should do close to 400 watts.......MAYBE THIS WEEKEND.