Not really sure what to call this exactly, but its just an attempt at putting some data to speculation. Recently there was a thread where the subject of re-wiring the Dyna ST-70 rectifier to convert it from conventional full wave operation to use as a ballast and what effect that may have on the current capacity of the tube. I don't have an ST-70, but I have a Pilot SA-260 which is close enough to use as a test dummy.
The setup. First, lets see what the SA-260 power supply looks like. This is the exact schematic in use for mine, the values in red represent a change from stock. Its pretty close to as-designed, just with some values shifted for availability reasons.
As you can see its a conventional full wave rectifier setup using parallel 5U4GB tubes. None of the rest of it particularly matters for this discussion, but the rest of the amp is a push-pull EL34 ultralinear amp with 3 12a*7 tubes as the driver. Not grossly dissimilar from various Dyna ST-70 configurations. It also biases the tubes at 50ma per, exactly like the Dynaco does.
First, to set the stage for this I did a slightly modified version of the "yellow sheet mod", pictured here
Instead of a single 4007, I used two in series to make sure I had sufficient PIV available for the next part of this. The UF4007 was chosen on the basis of one unique property that it alone possesses: I had them on hand.
So with the stage set, lets see what we get. Before each test, the output tube bias was confirmed at 50ma per-tube after a 5 minute warm up period. All voltages measured at the junction of R48, L1, and C1A. This is marked as 430v on the schematic. Voltage supply as measured with my meter is 116vac. 5U4GB are old production RCA. The 5AR4 are new production Sovtek.
Stock with the 5U4GB
idle voltage: 400v
22 watts output: 367v
difference: 9%
5U4GB with pins 4 and 6 jumped together
idle voltage: 415v
22w output: 386v
difference 7.5%
5AR4
idle voltage 425v
22w output 399v
difference: 6.5%
5AR4 with 4 and 6 jumped
idle: 432v
22w output 406v
difference: 6.4%
As you can see with both rectifier types a bump in voltage is evident at idle and at power output. 22 watts both channels driven was picked simply because thats where it started clipping with the stock rectifiers. There may be some other issue here since the voltages check in low, but thats a problem for a different time. Regulation showed more difference with the 5U4GB than the 5AR4.
I suppose I could try and see if other pairs of 5U4GB clock in with similar results but I don't know how many pairs I have. I do not have more pairs of 5AR4 though so unless someone wants to send me a box of them for testing, this is all I can do. I'm a little too cowardly to try this with a single rectifier but its possible there would be a more obvious difference there.
comments and further data welcome, but lets not turn this into a which magical rectifier or diode will solve all our problems thread.
The setup. First, lets see what the SA-260 power supply looks like. This is the exact schematic in use for mine, the values in red represent a change from stock. Its pretty close to as-designed, just with some values shifted for availability reasons.
As you can see its a conventional full wave rectifier setup using parallel 5U4GB tubes. None of the rest of it particularly matters for this discussion, but the rest of the amp is a push-pull EL34 ultralinear amp with 3 12a*7 tubes as the driver. Not grossly dissimilar from various Dyna ST-70 configurations. It also biases the tubes at 50ma per, exactly like the Dynaco does.
First, to set the stage for this I did a slightly modified version of the "yellow sheet mod", pictured here
Instead of a single 4007, I used two in series to make sure I had sufficient PIV available for the next part of this. The UF4007 was chosen on the basis of one unique property that it alone possesses: I had them on hand.
So with the stage set, lets see what we get. Before each test, the output tube bias was confirmed at 50ma per-tube after a 5 minute warm up period. All voltages measured at the junction of R48, L1, and C1A. This is marked as 430v on the schematic. Voltage supply as measured with my meter is 116vac. 5U4GB are old production RCA. The 5AR4 are new production Sovtek.
Stock with the 5U4GB
idle voltage: 400v
22 watts output: 367v
difference: 9%
5U4GB with pins 4 and 6 jumped together
idle voltage: 415v
22w output: 386v
difference 7.5%
5AR4
idle voltage 425v
22w output 399v
difference: 6.5%
5AR4 with 4 and 6 jumped
idle: 432v
22w output 406v
difference: 6.4%
As you can see with both rectifier types a bump in voltage is evident at idle and at power output. 22 watts both channels driven was picked simply because thats where it started clipping with the stock rectifiers. There may be some other issue here since the voltages check in low, but thats a problem for a different time. Regulation showed more difference with the 5U4GB than the 5AR4.
I suppose I could try and see if other pairs of 5U4GB clock in with similar results but I don't know how many pairs I have. I do not have more pairs of 5AR4 though so unless someone wants to send me a box of them for testing, this is all I can do. I'm a little too cowardly to try this with a single rectifier but its possible there would be a more obvious difference there.
comments and further data welcome, but lets not turn this into a which magical rectifier or diode will solve all our problems thread.
