Converting EL38 amp to 300b Amp

[Paul Barker]

Baby steps in amp building. Too soon to build powerful GM70 amp.

You may see repurposing of components from the EL38 amplifier. The doner amp.

I spent a long time trying to make the srpp driver with el84/6v6/ 6550. But only the EL38 has better gain better power suply sharing with the 300b better transconductance. Its much more appropriate in all respects and sounds great as an SE amp.



upload pictures online

[Paul Barker]

Happy that I have an ideal transformer. Those caps are 80uF, I’ll order more.

4 new PL38 from Langrex £25 each.

Doesnt seem worth making the amp dual purpose. Ive opted for GM70. The R core OPT’s should fit inside, and the two grid chokes. OTL amp which was gutted out had a stepped attenuator. That is OK.

[Paul Barker]

Happy that I have an ideal transformer. Those caps are 80uF, I’ll order more.

4 new PL38 from Langrex £25 each.

Doesnt seem worth making the amp dual purpose. Ive opted for GM70. The R core OPT’s should fit inside, and the two grid chokes. OTL amp which was gutted out had a stepped attenuator. That is OK.

It only requires one ECC88 1/2 per channel and its 115k of resisters due to the quite high voltage available. A lesson from Morgan long ago.

[Paul Barker]

Actually.



On 10k distortion better and more current allowed through theTributes, which also lessens distortion.

[Paul Barker]

Wait a minute? Lets go the full mile. 20 Watts Class A1. Thd 0.82. Those four are the 1500v caps. There are four mor, which I’ll fit in. Glad I didnt go 300b.That choke is 30 henry 250mA. So this supply is for GM70’s. The original supply can do everything else and take it in it’stride.

[Paul Barker]

So anyway, I cant make any of it fit into the cases that I have. So I’m going with monoblock power supplie and monoblock signal section.

I have my eye on four Hammond Chassis. But I shant order them yet until more funds available. but Ive measured up and everything fits. I have ordered the resistors I need.

I think I’ll mount the Sowter transformer attentuators in each signal section.

The wider shorter one is the signal section, the longer narrower one is the power supply.

Two of each. They seem very good value, but to keep out of overdraft, I'll order them when more money has landed.

I recall Jono telling me in the old days. When you build monoblocks complete one side first, you may learn something you need to know. So I’ll build one channel and enjoy it first.

[Paul Barker]

[Paul Barker]

Anyway Major Tom to Ground Control. I have this Chassis, which the 300b amp is small enough to fit in. The R core OPT,s should fit inside. For the off board power supply I have some 400v power transformers which as they arent centre tapped I’ll use two hexfred bridges matching chokes matching caps, so to all intents and purposes monoblocks in a single chassis. The signal section will also be monoblocks sat side by side.





The power supply chassis to get perspective. It should be enough.

[Paul Barker]

The circuit now Ive landed in Appolo crash and burn.



Oops I fogot that circuit requires 4 PL/EL 38’s.

Queue another picture.

[Paul Barker]

As its monoblocks in one box every filtering section is duplicated.
Plenty of room above the plimsole line for these chokes on top.

And to show the duplicares.



The caps paired with the chokes probably polypropylene as the voltage up there is up to 565v and the small Russian paper in oil 26uF aren't rated high enough voltage.

[Paul Barker]

What can I say? I need a simpler stepping stone. Building a 300b amplifier using my bench supply. I’ll spell out the limits.

1/ Black Gate Limit b+ with the two Black Gates not stacked. I’m stretching it a tiny bit but I can control it to 462 V max at the Black Gates.

2/ Anode Volts Limit. The WE data sheet limits to 400v. The b+ limit is the dominant factor so due to limit 1/Im limited to 377v Anode Volts

3/ My first actually built amp since the stroke.

4/ Current through 300b. WE show a operating point at 77 mA to give 7W at 5% distortion. BUT I want to try the 75C1 Cathode Bias, which sets the current limit at 60 mA.

5/ grid resistance, WE for fixed bias limit is 50kohm. I see diode bias as fixed bias erring on the side of safety. I bet many of you never seen that limitation? Now I see where the figure suggested on a forum for fixed bias GM70 is 50kohm . I would suggest that may be too high.

6/ varying topology at every stage, no low pass built in limit at any stage, and no breaking of WE tolerances. Though Im not using WE Bottles. Lets use their rules!

7/ Make it look presentable to take it to Victor’s Vintage in a form if Environmental Health walked in it wouldnt look out of place next to Quad II’s and Stereo 20’s.

[Paul Barker]

Note WE suggest 5% distortion at a similar operating point: I’m not suggesting VTADIY is a better predictor of distortion. Just showing what is stated by VTADIY.

I have noticed one foible of my cognitive mind since the stroke. The top load resistor i meant as 6K8, I wrote 6Kh.

There you are, an interisting insight. The memory the cognition and the drawing skill are all affected. Then the elephant in the room is the motor skills when building it.

[Paul Barker]

Morgan was saying about bridging transformers with a small cap. The Hammond is bifilar so both windings side by side, surely is equivalent of bridging with a capacitor? But I am prepared to try small caps.

[Nick]

5/ grid resistance, WE for fixed bias limit is 50kohm. I see diode bias as fixed bias erring on the side of safety. I bet many of you never seen that limitation? Now I see where the figure suggested on a forum for fixed bias GM70 is 50kohm . I would suggest that may be too high.

There are two commercial valve amps out there that over the years I must have repaired every one of them in the UK (or at least 90% of them). Two very different amps, one push pull, one single ended. Both had a tenancy to fail with run away as the valves got old. The problem in both cases was fixed bias and too high a resistance in the grid circuit.

Worked fine when designed and tested, 18 months down the line, not so much. I suspect the problem was young designers who had learnt with solid state reading about valves, seeing how simple they were and "what could go wrong?" Missing little lines in some of the docs that seemed innocent enough, and "hey, its works, must be ok".

Back to Paul's comment, anything other than a resistor in the cathode is fixed bias. If increased current doesn't cause increased bias, its fixed.

[Nick]

Morgan was saying about bridging transformers with a small cap. The Hammond is bifilar so both windings side by side, surely is equivalent of bridging with a capacitor? But I am prepared to try small caps.

Yep, I have always believed that bifilar is just a way of smuggling a distributed cap into the circuit, and as caps are cheaper than transformers, why not just use a cap anyway. But (waves hands magically) transformers...