GM70 Amplifier

[Paul Barker]

How much do they matter? Globe 27s look superb . In my way I even rigged up 6p5s in basically the same choke loaded one valve pre-amp. I tried them all and an old 6sn7 sounded as good ( and it's stereo) . Only recently I put 2 and 2 together realising how similar 5687s were to ecc99, 6n6p and yes 6sn7, 12bh7 (lower current heater) hence then related to the gorgeous older vales. Never tried 56 but I have 27, 37, 6j5 etc.

I may not be right but I'd rather get the bugger going and then tweak.

Id rather get the bugger going too.

[Paul Barker]

Other factors than tube rolling are the priority such as how to power the VA, the driver of the Cathode follower and the Cathode follower itself.

Ive found a 530-0-530 @ 190mA which should do all those, leaving 650 volts for the Gm70’s maximum possible capacity.

So the plus rail of the cathode follower can come from the additional transformer. I might have enough black gates for it. The 650 is all black gate. Which just leaves the negative rail. I dont yet know about that. But I can see a variac and another transformer for the negative supply to set the bias of the output stage and replace the variac with a correct transformer probably bought in or might stumble on one browsing the lockup. The 650v is valve rectified no reason the 530v shouldnt get same. So should probably keep that trend for the negative rail.

3 power supplies. Its a lot easier to just stay in A1. But to operate such valves as 801a 211 GM70 212, 833 in A1 one must use higher voltage to obtain the ac voltage swing for power output. My stated aim is to employ the 0.5 amp transformer which voltage is not suitable, but its build and ready to go.

The 801a amplifier was limited by not having sufficient drive for A2 excursion, and the space in the chassis wasn’t able to accomodate the additional requirements. Thats is one reason it’s forsaken. Other reason is my state of mind, my wifes illness has made me very heartbroken. Anyone who has looked after a loved one with demensia of whichever type will understand. Ive always known it would come. So I’m ploding on in familiar ground and I really appreciate the support of you guys supporting my amnesia which is probably stress related not demensia for me. Especialy, though not wanting to single anyone out. But Nick’s contribution about the distortion perspective in A2, and reminding me about the 76. My memory of that was exactly the same as when you meet a friend and the frontal lobe wont release their name because of the stress of searching for it. 2 days slater the name is set free by the stressed portion.

Everyone is much appreciated.occupying my mind is keeping me sane.

[Nick]

27 is a indirectly heated 26 just about.

[andrew Ivimey]

Supposedly yes but the only reason I go for the 27 is that they look ssooooo ggggooooood . 6sn7s don't but they do a good job. You can see why/ how they were developed.

I've only had the courage to use 26s as pre-amp - lovely but hardly practical- oh hell- who cares!

Onwards.

[Paul Barker]

When I was a new recruit to self build valve amps it was great fun turbe rolling. But one was never doing it scientifically. Like Andrew says about the 27 the mesh plate draws you in. The 6/12 sn7 or 7n7 dont have that mojo. But Nick is the stability in this design, pointed to these valves as fit for purpose, and really the whole valve rolling thing may distract from best for purpose. Then I recall leaf rustling noises from said valves, and start thinking what western made supplies I might find in my jumbled lockup. When my mind’s eye reminded me all the years I grabbed 6J5’s, not knowing which were thrashed by Radio Hams and which are good. I never got round to sending my curve tracer off for upgrade and repair. But I guess I can, the door is probably still open. And friend Steve in Hull has good valve tester and generous spirit, and Max has made an offer. So there should be a quality control applied to the use of 6J5’s.

Otoh I may have a single pair of the desired nos version of the USA made 6SN7’s RCA Redbase. Just one pair, I recall buying them and my minds eye sees them in mint boxes. So we shall wait and see if they materialise.

[Paul Barker]

Ive just emailed Ronald to see if his offer to fix my utracer and update it to 3+ is still a service he provides. Then I can test and match the small signal valves. I’ll update when answer comes. I dont thnk my eyesite is up to scratch building a new version.

[Max N]

Another option might be 5687 (or whatever else you have that’s suitable) for first gain stage, and xn7 for second stage (you only need the big voltage swing for the second stage).

Thinking about the cf stage, do you need the negative bias voltage to be there before the GM70 starts operating? Which might imply solid state diodes for the cf rails?
I may be talking rubbish because I am very rusty…..

[Nick]

Yep, I do that on my 211 with a similar driver. SS for front end, damper diode for 211.

[Paul Barker]

Ill go with that. Are hexfred’s still in vogue?

[Nick]

They are still out there. I tend to use UF4007's but they are EOL now as well.

[Paul Barker]

Yes Max probably optimal as -2v bias allows plent of headroom, greater gain and cathode feedback ticks all boxes. Dont think 27 would be full bodied, might sound weedy. Ive not found a wow response to 27 anywhere.

Anyway the 5687 I have experience of and like it. The drive to the 6j5 or 6sn7 will empower it to all voltage potential required. Input
Capacitance between the two isnt different enough and as neither is high mu there should incur good hf response, unlike if 6sl7 was used. The 5687 as a cf would drive the GM70 but the el84 is much better with its transconductance. My 5687’s are matched and nos boxed Ratheon. So looks like a line up. Bags of headroom, a little local cathode feedback good input impedance and a high transconductance sledgeammer to drive the GM70 grid positive.

[Paul Barker]

Nick, dod you recall I had made a bridge for high voltage and you commented I hadnt put resistors to ensure no individual diode went beyond its limit? That bridge is hexfred. I should find that, then I’ll look with a loop to see the number then see what voltages individuals are ok with because the negative supply will probably be between 40and 100, probably 40 to 60, and anyway the gm70’s are so understressed aiming T a bias in the middle of the range wont kil anything.

[Paul Barker]

Nick, dod you recall I had made a bridge for high voltage and you commented I hadnt put resistors to ensure no individual diode went beyond its limit? That bridge is hexfred. I should find that, then I’ll look with a loop to see the number then see what voltages individuals are ok with because the negative supply will probably be between 40and 100, probably 40 to 60, and anyway the gm70’s are so understressed aiming T a bias in the middle of the range wont kil anything.

Those AE interstages you got me capable of 50mA would do the loading, just need to arrange the negative supply from the dcr of that against the quiescent operating point.

[Nick]

FWIW, I used a current source for the CF load.

[Paul Barker]

The hexfreds are 1200v 12A so thats chalked up.