Wanting to build a GU-50 SE amp

[DeathRex]

I've had these GU-50s for about a year now, and I need to put them in an amp. I've been looking for a schematic for a while, but with little data on GU-50s, I've been finding alot of variance.

One schematic has a 2.5K to 8 ohm output transformer, another has a 5K, another has a 8K. One has the tube running at 20ma and 300 volts across it, another has 60ma and 350 volts.

I've seen a 6SN7 SRPP driving the GU-50, and even both sides of a 12AX7.

The GU-50 is supposed to be a 40 watt tube, but it just doesn't look as impressive as a KT88.

I was thinking of running them at 500-600 volts and 250 on the screen grid, keeping the anode about at 20-25 watts, with a 6SL7 driving it. I guess I'd like to use an Edcor transformer, the GXSE15-8-5K. I can get a 5K to 8 and if I can, switch from the plate tap to the screen tap for more output. Will that get me about 3K?

[azazello]

I made SE GU50. Edcor GXSE 15..../5 sq. cm/ is very very small for GU50.
My output trani had 10 sq.cm - Hammond 125FSE, with very good sounding.
If You will use it in pentod mode, Ra must be about 5-7 kohm, if use in triode mode /Ri - 1 kohm, 2 and 3 grid to plate/ - Ra - 2-3 kohms. Output trani must be GXSE 25..... +U about 400v, Ia - 90-100 ma. In triode mode power will be about 3-4 watts, in penthode - about 8 watts. Good drive tube - E180F.
Good luck!

[EWBrown]

The 40% screen tap on the GXSE15-8-5k would be closer to about 800 ohms.

40% squared is 16% of 5K, = 800 ohms. Would not be 3K

If you flipped it around the other way, using the plate connection for B+ and the SG tap for the plate, then the effective impedance would be about 1800 ohms (60% squared, etc)

HTH

Thanks for the GU50 info, I have seven of these rather strange looking tubes (and four of the even stranger looking socket assemblies) looking for "gainful" employment.

IIRC, GU-50s are a Russian replication of the infamous WWII German LS-50 "Wehrmacht" transmitter tube.

In operating characteristics, it is fairly similar to a 6146.

/ed B

[DeathRex]

I guess neither the GU-50 or GXSE are very flexable. I thought I could get close to 8 watts out in triode, but I would like to keep the plate below 30 watts. I'd like to be able to switch from one mode to the other, probably need that Hammond to do that, just double the money.

[azazello]

EW Brown!....Ri in triode mode is 1 kohm, Ra must be 3 and more kohms / Ra = 3-5 x Ri/!

[EWBrown]

I generally use between 3 and 5 times the tube's RI to determine the SE OPT primary impedance. For a tube with internal resistance of 1K, use an OPT of 3 to 5K, etc. For a 2A3 or 300B, 2.5 to 4K is a good range, though I have used 5K with god results, slightly lower output power and significantly lower THD.

In some very few cases, as low as 2X Ri will work well, but mostly this just brings up the 2HD and 3HD to some rather intolerable levels.

With some tubes (12B4A is a good example) the OPT primary impedance can be as high as 10-12X the tube's internal resistance. This means less output power, but also a vanishingly low THD. Jef Larson's "Abraxas" 12B4A SET is a good example of using these extreme ratios.

/ed B in NC

[DeathRex]

What is the difference (power out and Ri) in trioding a GU-50, with either tying the control grid and screen grid together, or tying screen grid and anode together?

[kheper]

The GU-50 is supposed to be a 40 watt tube, but it just doesn't look as impressive as a KT88.

I was thinking of running them at 500-600 volts and 250 on the screen grid, keeping the anode about at 20-25 watts, with a 6SL7 driving it.

Many of the SE GU-50 Pentode mode designs are just modded KT-88 schematics.

I do not know what the sweet spot for the tube is, but running it > 600V is too dare-devily.

Here is a 3.5K @ 15W output for cheap. Tomiko makes nice transformers.

http://www.ebay.com/itm/1-output-transf ... 0520309435

For 20-25W plate dissipation @550V would be:

550V x .04A = 22W

GU-50s are 40W tubes, you can push much more current through them than 40ma @550V.


Here is a 5K @ 18W Tomiko output for cheap.

http://www.ebay.com/itm/1-output-transf ... 0502785423

[DeathRex]

Yea, I'm liking those Tomiko's, even though they're kinda smallish. I've made many UL amps and I've read GU-40s are nice in pentode. I'd still like to do about 500 B+ and 250-275 screen. I'll have to see what I have in power transformers.

How much power can I get out of a GU-50 SE pentode 5K output xformer, 500V B+ @ 60-70ma?

Kheper look here at the outcome of the oddwatt you helped me with.

[kheper]

Yea, I'm liking those Tomiko's, even though they're kinda smallish. I've made many UL amps and I've read GU-40s are nice in pentode. I'd still like to do about 500 B+ and 250-275 screen. I'll have to see what I have in power transformers.

I would go with a 200-225V toroid for a voltage doubler. You should get about 500V B+ and about 250V off of the center leg to run the screens and volt amps/srpps.

How much power can I get out of a GU-50 SE pentode 5K output xformer, 500V B+ @ 60-70ma?

I'd guess maybe 10-12W continuous class A.

Kheper look here at the outcome of the oddwatt you helped me with.

That is one sweet amp. The chassis is a work of art.

[EWBrown]

The two different LS50 / GU-50 grid connection methods are for selecting Class A SE (or PP) or Class B PP operation, Class A would be "1" with the G2 connected to the plate, and Class B would be "2" with the G2 connected to the control grid G1. The positive grid voltages are the tip-off that this is for Class B operation.

For our purposes the "1" connection for standard triode-strapping would be best. Class A is necessary for SE, and Class B is for PP only, a Class B SE would sound really NASTY

Class B does pull control grid current, and requires some amount of drive power, and an interstage tranny is necessary, and resistance coupling would NOT work.

The old fashioned 46 tube had a similar scheme, the second grid was connected to the plate for Class A low power operation, and the two grids were connected together for Class B PP operation, which could generate some pretty impressive power levels (up to 20W) for that era

/ed B

[kheper]

Here is a schemo for the GU50 in triode mode. 9W in triode and 18W in pentode are claimed at 370V B+. The author recommends pentode mode.



If you are going to run in triode, 500V is out. In pentode @ 500V, > 20W audio output sounds possible. I was way off with my 10-12W estimate.


EDIT: Some specs below.

Ua=300V Ug2=250V Ug1=-24V Ia=130mA Ig2=3.5mA Ug1=~17V Ro=2kOm P=18W

[DeathRex]

I found this one that seems to be the original pentode version from tubeland. But I have a few questions:
1. Why would you have 5.6K cathode resistors on the 6SN7s, that seems way too high.
2. The output transformer seems too low, maybe 5K?
3. The coupling capacitor seems too low at 0.15, maybe needs to be 0.33?
4. The low gain of a 6SN7.
All of which seemed to be addressed in the triode version.
Here is where I found it
http://www.tubeland.de/gu50se.htm

And he is saying 11watts in pentode.

[DeathRex]

Brian Weston at Edcor says "I am sorry, but we do not do multiple secondaries an any of our single ended transformers." They don't do multiple primaries either.

If the secondary is on the outside, I might just be able to add a few turns to add a 10 ohm tap onto a 8 ohm secondary. Hook up a 8 ohm speaker and turn a 5K primary into a 4K.

Has anyone opened up a GXSE?

[kheper]

I found this one that seems to be the original pentode version from tubeland. But I have a few questions:
1. Why would you have 5.6K cathode resistors on the 6SN7s, that seems way too high.
2. The output transformer seems too low, maybe 5K?

People who claim to have experience with the GU50 run it a 2-3.5k in SE and PP (at least at low voltage). They think the low output impedance matches the tube's characteristics.

3. The coupling capacitor seems too low at 0.15, maybe needs to be 0.33?

The charge cap coming off of the srpp in the original schemo is too low. In the newer one, it is 1.0uf.

4. The low gain of a 6SN7.
All of which seemed to be addressed in the triode version.

And he is saying 11watts in pentode.

Someone is lying. The specs are all over the place, especially Va=300, etc. yields Po=18.