DIY Audio Projects Forum

This experiment has taken an unexpected turn from: Looking for GM70 driver tube suggestions so I have started this new thread.

It's time to get serious. Having satisfied my lust for an SET amp with the 300B, this former GM70 SET Stereo Amp project has now become a "GM70 Triode Stereo Oddwatt", the highest power, most outrageous version I can come up with of a self inverting push pull output stage coupled with a SRPP driver.

I tested the LM317HV with a dropping resistor and it works, or at least appears to. So biasing the cathodes with a CCS shouldn't be a problem.

It's always a happy day when the postman brings a box with those backward letters and numbers!



These GM70 tubes are HUGE! And $44 a pair (about $110 for the quad set with shipping)! 6E5P are also pictured as probable drivers at $2 a pop. The toroid power tranny is big too, 500VA. Those are 18" floor tiles for perspective. I picked up some high voltage PIO Russian coupling caps on ebay.

I purchased (ordered) Edcor CXPP100-MS-10K trannys. By using the 4 ohm tap for my 8 ohm speakers I can increase the primary impedance. The GM70 has a high plate resistance and likes to see a 10K load, which for a push pull class A amp would be 1/2 the impedance of the center tapped primary of a 20K transformer. At least this is what I've come up with so far from reading.

Any help or suggestions would be appreciated. This seems to me like a logical progression in the evolution of the SIPP output driver. Somebody had to have done this before, unless it just doesn't work. I don't want to screw this up, so any help to keep me from doing so is welcome.

B+ will be 850 volts, the tubes are biased at 125 ma each, and the input driver will be a shunt regulated push pull stage (what else?). I'll post a schematic soon.

This should be a fun project.

_________________
Stephen

Heaters:

The GM70 requires 20 volts at about 3 amps on the cathode filament. 240 watts will be dedicated to simply heating the cathodes!

Fortunately for me, and very coincidentally, laptop computers run on 20 volts (or many do) and the SMPS power adapters are relatively cheap and small, and sized exactly right for a single tube.

Without debating the musicality of this method, I got one for each tube (rated conservatively at 4.5 amps).

http://www.laptopz.com/laptop-ac-adapte ... -4491.html

If it weren't for laptops I think a 20 volt SMPS would be hard or impossible to find. A normal transformer for the heaters alone would have to be big and heavy to supply 240 watts of power, or about 12 amps at 20 volts. That's a lot of extra weight and real estate.

_________________
Stephen

6E5P SRPP driver schematic and output trace:




It appears I'm getting over 180 volts p-p swing before any clipping occurs - more than enough to drive the GM70. The 6E5P are connected as triodes.

Thanks to Magnus for the 6E5P suggestion and the awesome book on Russian tubes.

Now if only I could find a PSPICE model for the GM70, although it's not really needed for the design, it would be cool to model the entire amp.

_________________
Stephen

1.6E5P in triod mode has u=50 , in SRPP u will be 30% more, and I doubt that is enough
for driving GM70 with bias 80-100 v, maybe more.
2.850 v is small +B for good sound. My friend Manakov- tube russian guru advice me use
+B= 1100-1200 v, with bias 100-120 v. And SRPP 6E5P in triode doesn't enoudgh drive GM70
/I'll prefer SE GM70, my IMO is that SE is better PP for HiEnd sounding....in russian forum I sow heating 20 v from lap-top PSU - 4A, this is a good idea!/.
6E5P is exelente tube for driving power triodes and I'll advice You use it in penthod mode.
Sound is very detail and superior! /I used it in my 2 projects SE 6S33S like driver - only in penthode mode to drive 6S33S with bias 75-80 v. I tried use it like triode and u really!! didn't enough with input signal 1 v.from CD /.
3. I think that power triodes can use only in SE, IMO....

_________________
Hi End is back proportional to the number of composite parts!
Projects: OTL 6AS7 Gen, Electric, SEs 2A3 RCA, 300B JJ, 6S4S, 4P1L, EL11 Telefunken, 6AS7 RCA, 6S33S, 6S41S, 6S19P, PP 6005 Gen. Ellectric , headphone ampl. OTL Loftin White 6AS7 RCA....SE E84L& E80CC Siemens&Tel-n.
http://azazello-sound.blogspot.com/

I certainly hope you're wrong about #3 but all the comments are appreciated.

I calculated with the bias point and B+ that I chose, the GM70 would need at most 140 volts of voltage swing on the grid (which will be biased around -75 volts at the chosen B+ and bias current), and the 6E5P in a SRPP according to the PSPICE simulation will give 160 volts before clipping assuming it's biased just right. I could have made a mistake in my calculations. When I looked at using the 6E5P in pentode mode as a SE driver I couldn't get enough voltage swing and the output impedance was high. However, I'm not very good yet at designing pentode stages.

If I raise B+ as your friend suggests I completely agree I'll probably need to choose a different driver tube that can tolerate a higher B+ and give a wider output voltage swing. It's not the gain I'm concerned about, it's the output swing available and distortion when loaded by the power stage. The 6E5P SRPP (according to PSPICE) is just powerful enough (but barely) to drive a GM70 operated at the bias point I chose. One concern I have is the voltage withstand capability of the output transformers if I go with a very high B+. Based on the power rating and primary impedance they should be able to handle about 1400 volts p-p, but that's across the entire primary.

My plan is to get about 50-60 watts per channel using 100 watt transformers - overvoltaging them slighty due to increasing the primary impedance from 10K to 20K (8 ohm speaker on 4 ohm tap - you have to be careful with the voltage) and asking more than 50 watts from them. The transformers are the limiting factor here. I may have to keep B+ conservative. If I had purchased "real" 20K 100 watt transformers I would indeed raise B+ and shoot for 100 watts. But I bought 10K 100 watt transformers and simply changing the reflected impedance won't increase the breakdown voltage. I could run higher bias currents and use the 8 ohm tap and 10K primary but the operation wouldn't be nearly as linear.

Thanks for the tips. I do realize that the traditional use of these power triodes is in SE amps, but I just can't resist doing something different, and I see no reason it won't work. One thing is for sure. I'm going to find out. Keep in mind too that this will be a completely class A amp, so the tubes might not know they're in a push pull amp if we don't tell them.

My original intention was a GM70 SE amp, but after the 300B I now have an SET and it's loud enough for the environment where it will be used and sounds fantastic. Now I want something different. Stay tuned for success or failure.

It's going to one hell of an expensive mistake if it doesn't work!

Thanks.

_________________
Stephen

azazello,
I have considered your advice and decided to take some of it. I take back the statement about the transformer being the limiting factor. Actually it appears the transformer can handle everything the tubes can throw at it. There are a couple of errors in my previous post about the voltages and power.

I need to increase B+ over 1000 volts. I miscalculated!!!

Here's the load line for one tube across half of a 20K primary with B+ at 1200 and the bias at 100 mA. Ignore the marks on the line.



The 6E5P will not drive the GM70 at this bias and voltage to full power. I need to do some more number crunching.

Thanks for pointing these things out.

EDIT: Crunched some more numbers. I don't think I can get more than 50 watts per channel using a B+ of 1200 volts no matter what I do. I also need a driver that can swing an output of 240 volts! The only way I can see getting any more power is to go with parallel tubes. I don't think I want to go there. This stuff already won't fit in the case I got.

My latest load line puts the grid bias at -120 volts and it can swing as low as -240 and as high as 0. B+ is 1200, and the bias is 100mA per tube.

For all I know I made a mistake. If anybody sees how I can get more than 50 watts per push pull pair please enlighten me.

To get 100 watts at the transformer's design impedance of 10K would require 2830 volts peak to peak across the primary, or about 1400 volts across each side of the primary. However, 10K is too low for the plates of the GM70, so the transformer will be run as a 20K transformer. Assuming I can get that large of a voltage swing, across a 20K primary the same 2830 volts will give me about 50 watts with much improved linearity over a 10K load.

This is my current understanding. That could and is always subject to change.

Thanks.

_________________
Stephen

Sincere I wish you success according Your reason!!
/I understand maybe not all from Yours long posts....Google didn't translate my all clear, but I know now Your
main line, thanks!/.
You can try, I think big mistake not will be! I know, You can manage!
Good luck!

_________________
Hi End is back proportional to the number of composite parts!
Projects: OTL 6AS7 Gen, Electric, SEs 2A3 RCA, 300B JJ, 6S4S, 4P1L, EL11 Telefunken, 6AS7 RCA, 6S33S, 6S41S, 6S19P, PP 6005 Gen. Ellectric , headphone ampl. OTL Loftin White 6AS7 RCA....SE E84L& E80CC Siemens&Tel-n.
http://azazello-sound.blogspot.com/

I think for my criteria, switching from upper to down output tube /acrossing 0 point/ in PP-cl.A is fatal to good and detail sounding and make sound "metal and cold"....I sense in my 96 db Fostex 206E I prefer listen SE 2A3, 6S4S, 6S33S, 4P1L, 30P1S, 6LR8, 6S41S....that I have. /sorry for off/.

_________________
Hi End is back proportional to the number of composite parts!
Projects: OTL 6AS7 Gen, Electric, SEs 2A3 RCA, 300B JJ, 6S4S, 4P1L, EL11 Telefunken, 6AS7 RCA, 6S33S, 6S41S, 6S19P, PP 6005 Gen. Ellectric , headphone ampl. OTL Loftin White 6AS7 RCA....SE E84L& E80CC Siemens&Tel-n.
http://azazello-sound.blogspot.com/

Sorry, but I add and will not write more....Listen to the music is not only listening and have fun by big power, but and try to find the place of every instrument or voice in the scene, when You will close eyes, that is posible with SE and dificult in PP with switching output tubes, that disturbe.
I know everyone haven't this kinds ampl to compare....Sory again.

_________________
Hi End is back proportional to the number of composite parts!
Projects: OTL 6AS7 Gen, Electric, SEs 2A3 RCA, 300B JJ, 6S4S, 4P1L, EL11 Telefunken, 6AS7 RCA, 6S33S, 6S41S, 6S19P, PP 6005 Gen. Ellectric , headphone ampl. OTL Loftin White 6AS7 RCA....SE E84L& E80CC Siemens&Tel-n.
http://azazello-sound.blogspot.com/

I appreciate all of your comments and consider none to be OT. If you're right about some of your observations this will give me an opportunity to hear first hand what you mean.

I'm convinced this will work. I'm not convinced it will sound great, but neither can I find a reason for it not to. I'm not looking for the SET sound. I want something different. I think a 300B SET, a KT88 oddwatt, and a power triode push pull amp are a good mix of amps to compare and see what I like.

There are times I like to listen at low volume with my SET amp, but there are times when I can't live without my 500 watt/channel MOSFET amp and 1000 watt 18" subwoofer. Sometimes I want volume. So I think each amp will have its place depending on my mood and the volume I wish to achieve.

I can always convert this amp to an SET, or even a parallel SET by replacing the output trannys and rewiring the circuit. So if it doesn't sound good in PP I'll turn it into a SE. It's hard to imagine an amp that sounds better than the 300B SE I built other than power output.

If I did build an SE, what do you think about a parallel tube amp using two GM70 each with its own adjustable fixed grid bias? Are the tubes considered to be "fighting" each other in parallel and does the unequal gain affect the sound, or is it a good way to get more output power? I ask because I'll have 4 tubes to work with if I change this later, and paralleling the tubes would allow use with lower impedance transformers from Edcor or Hammond instead of specially wound expensive ones.

Thanks for your comments. I wouldn't have caught my mistake as soon if you hadn't commented on the 6E5P.

_________________
Stephen