Audion Sterling SE EL34 to 6B4G

[revintage]

I have had this Audion SE UL EL34-eating amp lying around for years. To achieve 9W Erik chose 475V anode voltage. This was far to high for the output tubes to be reliable. As I need a 6B4G SE for my midhorns, this will be a good candidate for modification.

One obvious solution would be to find a 270VAC toroid to replace the 370V transformer. To keep costs down, my idea is to try SMPS´s as these are dirt cheap. I have one 300VDC plus 6.3VDC and one 48VDC to use. The supplies are known for unwanted hf-ripple. By using the original supply including rectifier bridge, the ripple should be neglible.

By feeding the DC through the bridge the hf-ripple will be rectified, anyway in theory. One of the original 100ohm serial resistors will also be replaced by a small 10mH ferrite core inductor. I have never seen this solution before, so it might not work IRL.

So far I have removed the bulky original toroid and begun modifying the circuit board. An E88CC at 5mA with CCS will drive the 6B4G.

As only one 6.3VDC heater supply is available, I will try an extremely simple solution, feeding the driver and both DHT´s from one source. No place inside for regulator boards. Matched tubes will be needed.

IMG_0725.jpeg (197.79 KiB) Viewed 1748 times

IMG_0726.jpeg (238.39 KiB) Viewed 1748 times

PSU.png (23.28 KiB) Viewed 1748 times

[revintage]

Original schematic versus the one to be tried.
R1/R10 to be decided later.
The OPT is 5,6k:8 ohm which means the amp should output max 2,5W with very low H3.

sterling.png (191.57 KiB) Viewed 1739 times

sterling6b4g.png (31.32 KiB) Viewed 1739 times

[Paul Barker]

I think you have too much B+ on the chart below I’ve written in the B+ you require, and the cathode resistor value and wattage based on 4x formula which is safest, and that only if the heat sink for a cadok or space and air movement for a green wirewound is sufficient.

I’ve just put the Max operating point for the 6b4g in a calculator and it came up with this.

The good news is you get 4.5 watts at 3% 2nd h.



It really is a problem only having one 6v filament supply because it makes both 6b4g’s fixed bias as they are combined at the filament and at ground. So the individual valves cannot get the benefit of auto bias.

So your valves must start perfectly matched and end perfectly matched at end of life. Unrealistic expectation.

[revintage]

I have set the working point a little below 15W Pd at ca 280V/52mA, taking the voltage loss over the OPT and cathode resistor in account. Note the bridge is fed with 348V DC. I will get ca 47V at the negative and ca 53V on the positive side of the cathode, so 900 ohm(original green wirewound 510 ohm in series with an added 5W 390 ohm) should be about right, but will be tried out when the amp is up and running. Should give just above 12V ptp over 8 ohm at clipping, so anyway above 2W. About the common cathode resistors it is a compromise that will work, but how it sounds is to early to say. I have a good stock of 6B4G(both used and NOS), will sit down by the AVO and sort a pair out. Now waiting for a few missing components to arrive.

[revintage]

Maybe I should use the second noval socket by parallelling the drivers? This way the S will be 9mA/V and Ri 3.75k with the CCS at 5mA(2.5mA each tubehalf). With a single triode S will be 7mA/V and Ri 4.5k at the same 5mA CCS current, according to Philips datasheet.

The schematic now shows the calculated DC voltages.

para88.png (32.49 KiB) Viewed 1636 times