6CB5A/EL84 amp

PakProtector

AK member
hey-Hey!!!,
So here is the schematic for the 6CB5A/EL84 amp I'm building. There is some missing bits; the heater transformer is missing along with the 6NO45T delay for the plate transformer primary( to pre-heat the 866's ), and the bias supply transformer and filter. Each monoblock is going onto a 24x10" plate chassis. Output is a 20-20 Plus S-258-Q. E-Linear taps in this OEM configuration are at 50%, and will be reduced to 25-30% if/when the transformers are taken apart at Heyboer. The screen reg is adjustable from ~80V to 125V and will be tuned after the amp is running. I suspect the preferred OP will not be that for maximum power, but slightly lower so that the load line hits g1=0V well away from the knee in the upper left part of the curves.
cheers,
Douglas
 

Attachments

  • 6CB5A amp.pdf
    165.9 KB · Views: 466
glad you know what you're doing. sounds very confusing.

The funny thing is it could only be a little bit simpler. 866's and their delay circuit could be swapped for damper diodes for sure...but that's about it. It shows a bit why few commercial amps ran the TV sweeps; high heater current, complexity for g2 regulation top the list. On the simple side, blaanced, local FB is a plus; no compensation required for a loop of global needs attention.
cheers,
Douglas
 
The funny thing is it could only be a little bit simpler. 866's and their delay circuit could be swapped for damper diodes for sure...but that's about it.

Just curious... what advantage do you figure for using 866's over damper diodes? I understand a little regulation advantage with the MV tubes, but on several other scores the dampers look a bunch better to me.
 
Just curious... what advantage do you figure for using 866's over damper diodes? I understand a little regulation advantage with the MV tubes, but on several other scores the dampers look a bunch better to me.

I think they sound better; why exactly that is I have no idea. I had built a linestage a while ago and ran both dampers and 816's in it and far preferred the 816's. It seemed just a shade pale with the high vacuum diodes. Ov course, keep in mind the time required to make the switch...but I did try it a few times. No change to the operating point was made; B+ was nearly the same( IIRC it was 5V higher with 6DE4's v. the 816's ). I wish I could say why, but since I can't I'll just go to the trouble of using the big Hg tubes.
cheers,
Douglas
 
This design confuses me completely:headscrat. I'm sorry if this question is too basic (or just plain dumb) but do I understand correctly that the Plates of V3 and V4 are loaded with 8KΩ resistors and fed from taps in the output transformers?

If so - why?

Thanks

Shelly_D
 
This design confuses me completely:headscrat. I'm sorry if this question is too basic (or just plain dumb) but do I understand correctly that the Plates of V3 and V4 are loaded with 8KΩ resistors and fed from taps in the output transformers?

If so - why?

Thanks

Shelly_D

That is the E-Linear local FB topology. It is a variant of the sort of plate-to-grid FB developed by Schade. Half the power tube plate voltage( this for a 50% tap ) is fed back to the driver tube's plate...which is also the power tube's grid. The E-Linear connection also allows simplifying the PS; only need one node to feed the OPT's center tap.
cheers,
Douglas
 
Hi Douglas,

In this schematic you show a PP grid choke which you mentioned to me recently in another thread. Have you been able to find the p/n for it?

Thanks,
John
 
That is the E-Linear local FB topology. It is a variant of the sort of plate-to-grid FB developed by Schade. Half the power tube plate voltage( this for a 50% tap ) is fed back to the driver tube's plate...which is also the power tube's grid. The E-Linear connection also allows simplifying the PS; only need one node to feed the OPT's center tap.
cheers,
Douglas

Thank you. Please forgive me being a pest about this but I have never run into this before. Are there any websites you know of off hand that have good explanations? My Google of e-linear feedback returned a confusing array of sites, mostly not related to what I wanted.

Thank you

Shelly_D
 
hey-Hey!!!,
So here is the schematic for the 6CB5A/EL84 amp I'm building. There is some missing bits; the heater transformer is missing along with the 6NO45T delay for the plate transformer primary( to pre-heat the 866's ), and the bias supply transformer and filter.

Hey hey!!! Is the 6CB5A the driver tube? It looks like it could be used as a nice final. It also looks like it consumes 3 times the heater current the EL84 does.

I have a few questions about the schematic and parts selection:

In the phase splitter section it looks like V3.Grid 3, V3.Kathode, V4.G3, and V4.K are all tied together? Are all of those connected to the cathode of the cap that connects to V3.G2 and V4.G2? What is the symbol after the cathode of the cap, I can't make it out. Where is all of that terminated, is it grounded?

In the final section I see that each power tube gets a grid choke from G1 to each tube's bias supply. Any suggestions for what to look for in a grid choke? Are the cathodes of the finals grounded? Or are they tied to G3, and then through a resistor to ground?

Each monoblock is going onto a 24x10" plate chassis. Output is a 20-20 Plus S-258-Q. E-Linear taps in this OEM configuration are at 50%, and will be reduced to 25-30% if/when the transformers are taken apart at Heyboer.

If/when next group buy happens, let me know. Would you want to get multiple percentage taps, so you can experiment around?

The screen reg is adjustable from ~80V to 125V and will be tuned after the amp is running. I suspect the preferred OP will not be that for maximum power, but slightly lower so that the load line hits g1=0V well away from the knee in the upper left part of the curves.

I thought the idea of using a rotary switch to select the zener diode length by grounding out the cathode in different places was simple and brilliant. I am not at all familiar with the solid state devices used, I can only guess that they are a voltage regulator, maybe?

The funny thing is it could only be a little bit simpler. 866's and their delay circuit could be swapped for damper diodes for sure...but that's about it.

I see the B+ is a bit lower than the 300V that I commonly see with EL84 amps. I'm going to guess that's because it's fixed bias and probably Class A? I appreciate the schematic for "best possible sound", but I'm guessing that you could use an existing supply as long as it's got the right B+ and can handle the current? I might start by retrofitting an existing piece to try something like this.

It shows a bit why few commercial amps ran the TV sweeps; high heater current, complexity for g2 regulation top the list. On the simple side, blaanced, local FB is a plus; no compensation required for a loop of global needs attention.

I haven't built an E-Linear amp yet. This might be a good candidate. I like the simple short feedback idea.

I think they sound better; why exactly that is I have no idea. I had built a linestage a while ago and ran both dampers and 816's in it and far preferred the 816's. It seemed just a shade pale with the high vacuum diodes. Ov course, keep in mind the time required to make the switch...but I did try it a few times. No change to the operating point was made; B+ was nearly the same( IIRC it was 5V higher with 6DE4's v. the 816's ). I wish I could say why, but since I can't I'll just go to the trouble of using the big Hg tubes.

Reading this makes me happy. It sort of affirms the subjective quality of my attraction to tubes themselves. Sometimes things just sound better.

That is the E-Linear local FB topology. It is a variant of the sort of plate-to-grid FB developed by Schade. Half the power tube plate voltage( this for a 50% tap ) is fed back to the driver tube's plate...which is also the power tube's grid. The E-Linear connection also allows simplifying the PS; only need one node to feed the OPT's center tap.

The Schade feedback is the type employed in the RH84 amp, no? Neat schematic that you've presented here, gives one a lot to think about! :thmbsp:
 
I think they sound better; why exactly that is I have no idea.

Thanks. I put away or gave away all my mercury tubes when the kids came along -- seemed kinda dumb to wait for some toxic accident. But they are getting old enough now that I could see using some in a power supply -- as long as it has a cage. I still have some 866's around somewhere... I recall someone saying that he had to put resistors in series with the diodes to get any two of them to act at all the same in the circuit. Have you found that to be an issue?
 
----In the phase splitter section it looks like V3.Grid 3, V3.Kathode, V4.G3, and V4.K are all tied together? Are all of those connected to the cathode of the cap that connects to V3.G2 and V4.G2? What is the symbol after the cathode of the cap, I can't make it out. Where is all of that terminated, is it grounded?

That is supressor connected to cathode, and they're riding a CCS set for ~38 mA. It is returned to the boas supply; that one has plenty of current capacity to sink a few mA like that. Another RC stage downstream and I'll feed the bias pots.

----Are the cathodes of the finals grounded? Or are they tied to G3, and then through a resistor to ground?

The supressor/beam plates and the cathodes are effectively grounded; I am using 10R resistors to measure idle current.

----If/when next group buy happens, let me know. Would you want to get multiple percentage taps, so you can experiment around?

That is the plan. Tap percentage depends on number of layers, and I suspect these will have either 8 or 10 layers per half of the primary, so 12.5 or 10% increments would be available w/o issue.

----I thought the idea of using a rotary switch to select the zener diode length by grounding out the cathode in different places was simple and brilliant. I am not at all familiar with the solid state devices used, I can only guess that they are a voltage regulator, maybe?

Thanks! These are high-precision/low noise and at least as important low impedance. They're ~1 Ohm from 1 to 20 mA, where good Zeners are about 7 Ohms or about in the neighborhood of a gas VR tube.

----I see the B+ is a bit lower than the 300V that I commonly see with EL84 amps. I'm going to guess that's because it's fixed bias and probably Class A? I appreciate the schematic for "best possible sound", but I'm guessing that you could use an existing supply as long as it's got the right B+ and can handle the current? I might start by retrofitting an existing piece to try something like this.

The B+ is set for delivering a Class A output stage with a 5k a-a load. That is effectively 2k5 per phase, so B+ divided by a-a/2 needs to be close to the idle current, or perhaps smaller to be sure of Class A. I plan on 90 mA idle current per 6CB5 so it is close...depends on g2 voltage on where the load crosses the g1=0V anode curve while the other is still conducting.

In this one the EL84 is a signal tube, plate voltage of ~150V with g2 around 100 delivers a good idle point and load line with the 8k plate loads. It is quite different from a TX loaded power stage running from a 300V B+...:)

Look up Pete MIllett's page and find his AudioXpress article on the KT88 E-Linear Amp.
cheers,
Douglas
 
Thanks. I put away or gave away all my mercury tubes when the kids came along -- seemed kinda dumb to wait for some toxic accident. But they are getting old enough now that I could see using some in a power supply -- as long as it has a cage. I still have some 866's around somewhere... I recall someone saying that he had to put resistors in series with the diodes to get any two of them to act at all the same in the circuit. Have you found that to be an issue?

The series equalizing resistor is only for use in parallel; two or more to increase current capacity v. the singles I drew. A pair of 816 would serve this amp, and 866 a bit of overkill, and if I wanted even more than that I can step up to 872's...:)
cheers,
Douglas
 
Doug, couple questions..

Is this your basic layout?

And the hookup of the RCA/XLR input?
I know on the LTP both grids get a grid stopper if you want one an that both grids
get a grid to ground resistor and set the input impedance, BUT are the resistors in
the XLR connections to ground in parallel with that standard RCA and opposite grid
to ground resistor? Meaning that say you want a 100K input impedance you'd need
a 200k resistor at each grid to ground and on each XLR input to ground?

EDIT: (or leave both grid to grounds and remove both XLR input to grounds then)
 

Attachments

  • doug_amp.JPG
    doug_amp.JPG
    32.5 KB · Views: 106
hey Keg,
You've got it right IFF you wish to be able to run either connection...however, it is more an either, or sort of thing; you can't simultaneously ground the grid and drive it from the balanced connection.
cheers,
Douglas
 
The 816 must have slipped right past me. Do the hg ones glow like the 866, or just around the center like the 83?:scratch2:
 
The 816 must have slipped right past me. Do the hg ones glow like the 866, or just around the center like the 83?:scratch2:

Construction is the same as 866; and so they glow well. It is dependant on cathode sheild to anode spacing. I wish I could source some Taylor globe 866's...the ones without the cathode shield that is. On those the glow is absolutely brilliant.
cheers,
Douglas
 
Back
Top Bottom