EF86 amp with reverb and ready to build? Your opininions on this layout pls :)

[pnadora]

Hallo,
again thanks for the help so far in my older posts!! I worked your advice into this new layout. (I still cannot get expresssch to run properly so pls excuse that it is just a picture)
I am building an amp out of an old hammond k-100 that did not work anymore (only the amp worked luckily).

I really like the way an ef86 sounds in v1 especially with pedals in front so I stole the setup for v1 from the old vox layouts. After that I made some changes to my first layout and I am now planning to use a 12AX7 as CF to drive a FMV tonestack. I planned the CF to be at a similar operating point as in the old bassman amps, since I liked Merlins describtions on how it contributes to the sound and I am pretty sure that after an ef86 it should be pretty sweet. After the tonestack I am splitting into the reverb and I am using parallel 6FQ7 Triodes to drive the reverb transformer. (Sadly the old organ reverb tranny is broken, so that I am using a standard fender one now into the old 8ohm gibbs manufacturing reverb pan). I want to use trimpots for the dwell control and the resitor between reverb send and return so that I can adjust the reverb in a way that I have a nice one knob reverb control on the front panel. After the reverb I am going into a 12AX7 to give me one more gain stage and the second 12AX7 half inverts the phase in a paraphase inverter setup for the power amp tubes. The voltage divider brings the signal level down for the phase inverting half of the 12AX7.
The power amp is then taken out of an ampeg reverberrocket that seemed to be the only amp I found that uses the 7591 output tubes that came with the broken organ and since I want to use as many organ parts as possible those are the ones I plan to use. The power amp setub is also pretty much 100% taken from the tubedatasheet as far as I can tell.

For the poewr supply I have those large cylindrical capacitors from the organ that I want to reuse. I worked the current draws of the tubes out with the datasheets and 2 days of load line drawing and chose the power supply resistors in a way that I end up with the desired voltages.

The output transformator of the organ seems to have a winding ratio of 33,5:1. The 7591 datasheet suggest a load of around 6600 to be ideal. I think that with the 2 organ speakers in parallel I end up with a load of 4ohm aka 4500 on the power amp side, but the OT is pretty massive so I don think that should be a problem, right?

What do you guys think about this design? I am really excited so far and cannot wait to start ordering the missing parts so I dont trust myself to much at the moment

[terminalgs]

the power supply isn't illustrated, but based on your voltage supplies, you might be planning on a single power supply node feeding the ef86, the recovery gain stage (you list it as 'mixer') and the phase inverter.  If you are planning this, it might cause problems. A traditional series power supply is a good recommendation:


rectifier->A->B->C->D->E


A supplies the OT, B  for screens, C is for reverb driver, D for the 2nd gain/PI triodes, and E for the ef86.


reverb recovery could be E as well.,   reverb driver could be B if you want to shorten the chain, and your CF can be on E as well.  If I were you,  I would worry about hitting those voltages exactly based on tube datasheets.  12ax7s can take more B+.


also regarding the 500K trim pot between the dry feed to the reverb and the wet return:   500K might be too low a value.  besides preventing a feedback loop from wet back to dry, it also forms a voltage divider (for your primary signal) with that 220K grid leak resistor on the next stage. You are knocking down you signal to 30% of what is is leaving the volume pot.  It might be okay, but you may also find it best with 1M or 2M resistor instead.

[pnadora]

The power supply is my main worry at the moment so thanks for the post! Here is what I will probably change to:
A-OT 450V
B-Screens, Reverb driver 400V
C-Phase inverter reverb recovery 270V
D-EF86, CF 250V
The common 40+20+20+20u can capacitors should be a good choice, right?

I should probably just use the fender values for the resistor between the reverb connecitons and go for 2.2M and a bypass cap. But would that not make my dry signal too low?

[pnadora]

I adjusted the layout and went with this power supply configuration:
A - OT 450V
B - OT Screens 400V
C - Reverb + Phase inverter 300V
D - EF86 +CF 270V

Putting the reverb driver (6FQ7) on 400V didnt seem right looking at the tube data.
It still runs at around 4.5W dissipation if i am not mistaken. But that should be ok for both triodes I think. Maybe I should just buy a 12AT7...
The resistor+cap between the reverb insertion points is still strange to me on the fender layouts 

[terminalgs]

The power supply is my main worry at the moment so thanks for the post! Here is what I will probably change to:
A-OT 450V
B-Screens, Reverb driver 400V
C-Phase inverter reverb recovery 270V
D-EF86, CF 250V
The common 40+20+20+20u can capacitors should be a good choice, right?

Yes, if your rectifier can handle a 40uf as the reservoir.  If SS, no problem, if its a 5Y3,  that's too big.  To go from 400V@B to 270V@C will require a big value resistor, I'd use a 10K between B&C and C&D and see where your C&D voltages land.


If you are using suggested values on the tube datasheets, keep in mind those are only suggestions or 'typical operations'.  You can have 300V on the plate of a 12AX7 in a gain stage.  I don't know about the 6FQ7 but the 12AT7 (like you mention) can handle 400VDC at the plate driving a reverb transformer.

The closer you can get the entire PI/output stage/power supply to look like some known good design, the better.  You mentioned an Ampeg Reverberocket.  I'd clone one of those (RR, Gemini, etc) as closely as possible.  Later RR's might be worth looking at too.

I should probably just use the fender values for the resistor between the reverb connecitons and go for 2.2M and a bypass cap. But would that not make my dry signal too low?

The fenders that use 2.2M or 3.3M also have the signal coming from a 2nd gain stage following the T/S. Yours is not going to have as much gain, so I'm not sure what the best fit will be to get the gain where you want it at grid of the power tubes..,, Your idea of a trim pot is good, I'd just use a higher value, maybe 1M to test with and tweak. be prepared to bump to a 1M pot + 1M R in series..

[pnadora]

The idea of 1M pot and 1M resistor seems the best option! :)

I changed my mind and I will just go for the standard 12AT7 in the reverb driver position. Just to make it more predictable for me and who am I to mess with one of the best reverb circuits ever :P

The power supply I designed works out in the in the duncan PSU designer with the current tabs that I included in the schematic. The voltages are closer to 280V and 308V at tab D and C. Maybe I should up the 20k resistor to 22k to bring it down a little. Instead of one 40+20+20+20u can capacitor I could also go with 40+40 and 16+16u or 32+32 and 16+16u to make the response a bit quicker I think. I am using a GZ34 as rectifier so the 40u at the start should not be a problem.

Here is my current version of the layout with your ideas worked in and the power supply :) :

[ac427v]

I have always followed Merlin's notes when elevating heaters. He has an additional cap to ground from the junction of the two resistors.
http://www.valvewizard.co.uk/heater.html
Craig

[pnadora]

thanks for the comment! Since I take the voltage from the last junction of the power supply it should already be smoothed out pretty much. But I will think about adding a cap just for safety. :)

[sluckey]

The filter cap will ensure that the filament CT is referenced to AC ground. I think that's a good idea.

[pnadora]

Ah ok! Do you know value would be right for 50Hz european power supply with the 220k 47k  voltage divier?

[sluckey]

It's not critical. Anything greater than 10µF and rated to handle the elevation voltage will work. If I had to buy a cap, I'd get 47µF @ 100V, just because I like that value for bias caps too.

[pnadora]

I think I pulled some of those out of the organ :)