GU-50 Push-Pull AB OT Impedance/Load Line Questions

[defaced]

Thank you for the lead es345. That's actually not a bad price, it works out to about 155 with shipping according to their online calculator. If I'm looking at it correctly, there's only one secondary, so that's a bit of a bummer.

I am getting quotes from Heyboer, Edcor and Hammond. Right now Hammond about twice the Heyboer in both cost and lead time. I'm not surprised by that based on what I've heard regarding Heyboer's prices. I'm curious to see what Edcor comes back with. Some old posts on DIY Audio are not very kind about their replies and lead time, but that was many years ago. So I'll give them till early next week then proceed from there.

Thinking about the test setup, I'm going to see how an Antek AS-2T300 works with the Hammond (1650T) or Mag Comp OT (40-18072) I have with mismatched loads. I just want to get something "close enough" to see how it performs in a general sense.

[tedmich]

Thank you for the lead es345. That's actually not a bad price, it works out to about 155 with shipping according to their online calculator. If I'm looking at it correctly, there's only one secondary, so that's a bit of a bummer.

Another page from this company has a pull down menu which allows the choice of seven different secondaries for this OT including: 2,4,8,16 singles and 2/4, 4/8 and 8/16 duals, might cost a bit more though

[es345]

If I'm looking at it correctly, there's only one secondary, so that's a bit of a bummer.

There is a drop down menue at right side ("Ausgang") where you can choose 2 secondaries in their standard offer. If want more you can ask per mail.

edit: Tedmich was faster...

[defaced]

Minor update:

I have this thing prototyped and am doing some output power testing, but I'm only getting about 90 watts from the pair. Unloaded voltage is about 850-840v. Idle conditions are 800-790v B+ at 110-100 ma. Full load conditions are about 700-690v B+. With the 8.8K load (2.2K mismatched 4:1). I tried dropping the load to 4.4k just as a quick check and saw a decrease in output power (along with a corresponding decrease in full load B+). So this indicated to me that the B+ sagging to 700v under full load is the reason I'm not getting 120w per pair. I have another power transformer on order which should arrive next week. That'll give me 640v AC and 650ma, so I should be much closer to 800v under load.

One thing I've noticed is that with this setup, the power tube bias point seems to be very narrow. With -50v I get between ~45 and 55 ma per tube (the tubes I have are not matched). If I go much colder I get cross over distortion, and if I go much warmer the output power decreases. Thinking about this, it makes sense considering how close to Class B these are running. I was hoping that biasing colder would allow me to give more input swing for more power, but no such luck. So it looks like the most I can cram into the grid of these things is ~100 V P-P.

Another thing I noiticed is that these are sensitive to screen voltage. With the screens at 275 I lost about 20w output as compared to 300v. 5K screen stoppers are too much (those are what gave me the 275v off the regulated 300v screen node). 470R produces less drop (295v at the screen pins), but I'm not sure if they're enough protection.

I'll post a schematic some time soon, but the current test setup is (it's all sand until the power tubes, LND150 gain stages and PI, ZVN0545 buffers with BJT constant current sources)

Signal generator - gain stage - volume control - gain stage - cathodyne PI - buffers DC coupled to power tube grids- 8.8K OT - 16R load.

[g1]

If I go much colder I get cross over distortion,

Big enough to be audible? You probably know just how much visible on the scope it takes to have a negative effect on the sound but I thought I'd mention it anyway. It's not like SS amps where a bit of crossover is readily apparent to the ear.
And have you played with the screen R values?

[defaced]

I haven't listened to it yet. I probably won't get put a preamp and speaker on it until this weekend at the earliest.

I have not played with screen resistor values extensively. In my impatience I omitted them initially, and I get the same output with 470R as I did without them. At 5K I see a 20w reduction in output. I didn't see much convergence on screen resistor values when I looked at various schems online, so I just got some resistors of the values I saw.