845s or GM-70 amp.

[Polisois]

I intend to build a high voltage amplifier using the 845s or the Russiam GM-70.
I have 4 each of the above valves.
This would be the first time I use such tubes and a high voltage power supply.
A suitable transformer is available from my old stock that delivers between 1100 and 1200 Volts and it looks like it can handle 500-600mA.
My plans are a parallel SE layout, either in two monoblocks or on a common large chassis.
The heavy power supply would be in a separate box and I will make sure the insulation of the wires from said box to the main chassis to be adequate.
I would be grateful to experienced builders to let me have their advices, such as :-
- which tube would they recommend ?
- suggested HI-Tension voltage (after the filter).
- Corresponding bias ( fixed).
- Load impedance (parallel SE) based on 6 ohm secondary load.
As I would like to stay in Class A operation, what driving swing would you recommend ? ( I am not looking for the maximum power, 20 W per tube would be enough).
The driver unit I designed would be directly coupled to the output tubes. I have successful experiences with 300B and 6C33.
A peak to peak signal of 200 Volts would not be a problem.
With regard to the output transformers, I have the necessary know-how to build them myself and possess a couple of insulation testers (up to 3kV) to check their suitability.
Thanks
Polisois

[Toppsy]

Polisois,

I think this posting would be best placed in the DIY projects - Amplifiers section this forum.

I think this place is for WD designs and kits.

However, you have a PM that may help you with your design.

Are the gm70 tubes you have the copper plate anodes or graphite?
The graphites give quite a dark sound but are cheap. The coppers though more expensive give far superior sound quailty and well worth paying the extra. I use coppers in my gm70 amp.

Colin

[Polisois]

Thank you Colin,
Unfortunately I have graphite plates GM-60.
Another problem is the socket. How did you manage ?
Some suggest to use the 6C33 sockets and widen the diameter corresponding to the larger pins, with a drill for glass.
Anyway, I'll post again the request on the section you mentioned.
BR
Ari.

[Toppsy]

Sockets for GM70 tubes are readily available on ebay and other web sites.

I got mine off ebay.

I'll have a look on some sites and post you the links.

Colin

Try these two links:

http://cgi.ebay.com/NEW-SUPER-Cerami...742.m153.l1262

http://cgi.ebay.com/NEW-GOLD-Ceramic...742.m153.l1262

or if you feel rich and want to splash out try these:

http://www.jacmusic.com/sockets/html/Yamomoto.html

[Polisois]

Thank you Colin.
I am very pleased to find something ready made.
All the best.
Ari.

[Paul Barker]

Originally Posted by Polisois

My plans are a parallel SE layout, either in two monoblocks or on a common large chassis.

If you read RCA manual it recommends a 100 ohm resistor on the anode or the two transmitting valves oscilate to one another. There have always been a lot of oponents of parallel output stages. The Audionote parallel 845 amp was obviously not a great seller as Peter Qvortrup was selling off the output transformers cheap at the end. He may still have some.

Originally Posted by Polisois

- which tube would they recommend ?

drive a single gm70 very hard they are cheap consumables, and driven very hard make an acceptable amount of reasonable sound. 845 is quite different presentation, very polite, not as silky in the midrange as the 211, and should have more bass but actual experience does not correspond with electrical characteristic predictions. You will have to try both and make up your own mind at the end of the day. The best I can say is both valves are quitedifferent and further different to the 211 which I prefer (but not the Chinease one).

Originally Posted by Polisois

- suggested HI-Tension voltage (after the filter).

950 to 1250.

Originally Posted by Polisois

- Corresponding bias ( fixed).

Start with data sheat spec adjust on test.

Originally Posted by Polisois

- Load impedance (parallel SE) based on 6 ohm secondary load.

As I would like to stay in Class A operation, what driving swing would you recommend ? ( I am not looking for the maximum power, 20 W per tube would be enough).

Neither valve particularly fussy there are bigger issues in amp design than load impedance, use what most people use, don't get hung up if you are 1/2 or twice as much.

Originally Posted by Polisois

The driver unit I designed would be directly coupled to the output tubes. I have successful experiences with 300B and 6C33.

I imagine 300b is nicest.

Originally Posted by Polisois

A peak to peak signal of 200 Volts would not be a problem.

then particularly for the 845 I would go with that the gm70 would tolerate less voltage swing for the same power out.

Originally Posted by Polisois

With regard to the output transformers, I have the necessary know-how to build them myself and possess a couple of insulation testers (up to 3kV) to check their suitability.
Thanks
Polisois

The difficulty you face ion opt design is the large amount of dc current your plans necessitate, or go parafeed.

In my experience there are deminishing returns with too many interleaved layers. Easier to wind say 1/2 secondary primary full secondary primary half secondary. Very simple to wind good result on test and good sound. Minimise iprimary inductance for best HF response as leakage inductance which worstens HF performance increases in multiples of 4 as no. of turn increases. So the deminishing returns follow a cubic rel;ationship and very fast run away.

You can't beat physincs it is not possible to get perfect performance across the frequency range, select an area of compromise.

[Polisois]

Thanks Paul for the detailed answer.
a) It certainly is a possibility that the two parallel valves oscillate, if no precautions, such as the one you mentioned, are taken.
However the valves will be in parallel indirectly, as per paper 7151 presented at the AES (Audio Engineering Society - www.AES.org ) at the convention held in Vienna in May 2007.
In two words, the OPT consists in 4 separate primaries, each pair on a branch of a C core or a UI core. Two primaries (one from each branch) are connected in series and the current flows from the B+ to the anode of one of the valves. The same occurs with the other pair of primaries. This is represented as an "X" . This layout ensures a self compensation of the idle currents, so that you do not need to have matched valves.
Honestly, due to the size of the GM-70s, I do not know if self oscillation will take place or not ( the prototypes were using EL34s triode connected).
More details can be found in the above paper, or I can post them to the Forum (provided the XLS files are accepted).
b) It is true that lots of people prefer the 211, but I tend to believe that, if
you give the valves what they need, and if their curves are "regular" you can obtain a good sound as well.
c) I will try around 1000 Volts.
d) Load impedance : OK - I will see practically what number of primary and secondary turns I can fit, and I will stay between 7k and 10k
e) Surely the 300B behaves very well, but I got excellent results also from the 6C33s, in spite of what many people say about this valve. One point on which they are 100% right : these valves are very dissimilar in characteristics, but, if you use one per channel, you can always find a pair that works nicely, and you get 15-16Watts of power (enough for most diy audiophiles).
f) Please refer to point a) above. the self compensation of the idle currents enables you to work with as many milliamps you can physically expect from the windings ( wire diameter, space).
---
I would be very pleased to supply more details if requested, and if the amplifier will work satisfactorily, I will quote all the necessary information to duplicate it, including the OPT.
Ari

[Polisois]

Regarding point a) of my post on the 31th Mrch, I am enclosing two excel files illustrating the operation of the 4x4 SC- transformer mentioned.
Kindly note that said novel transformer can be used with SE as well as Push-Pull layouts. In both cases, the pilot signal to the valves must be of opposed phases.
Ari Polisois
I am converting the Excel files into PDF and will send them later.

[Polisois]

Attached the files of the 4x4 self compensated transformer.
Have a nice week end.
Ari Polisois

[Polisois]

Here is the second file