Polycarbonate Film Caps

Help Support GroupDIY:


Well-known member
Oct 10, 2016
Brighton Sussex UK
I've often gone for big Solen PP for local H.T. de-coupling near a particularly hot tube.

Yes - the two things that cause problems for electrolytics are ripple current (and hence power dissipation) and ambient heat.
In power applications it's not unusual that the capacitance value used is larger than necessary wrt pre regulator ripple but the capacitors used are required in order to realise a low enough ESR to avoid problems. eg Parallel Output Capacitors in PFC application.
I've substituted with large film caps in that application for an instrument (analytical not musical !) that inherently gets hot although the design didn't get into production. Not an audio application but using similar voltages that you might see in a audio valve application (c 400V) I believe - I know little about the specifics of valves.
Yes - the film cap did cost a bit more than the electrolytics (and would have required a pcb revision for production) but I reckoned the savings in terms of in warranty repairs and reputation made it worth looking at.


Well-known member
Nov 18, 2015
Quite a lot of previous discussions here on the Federal AM-864 and a few people made reproductions , I have a half finished one built onto a quad II chassis but havent got round to finishing yet . It can be a bit of a pig for making farting and burping sounds though , many people seemed to have trouble . I guess you could be exceptionally lucky or select 6SK7's very carefully , either way chances of each valve matching the other across the entire range is unlikely , what I'd planned to do with mine was use a dual concentric volume pot as the input attenuator , ordinarily both controls would be set the same , but it allows the possibility of trimming the drive balance to correct for any misbehaviour while under compression . In any case a single gang dual pot is unlikely to be very well balanced , maybe those stepped DACT pots I pointed out before could give better matching along with repeatability at the input . I never got that far with the project to find out . Id thought about using 6V6's in place of the 6SN7's as I had ample reserves in the power supply .
In the end if you look at the 864 its very much like other varimu comps of the time except theres no dedicated high voltage detection amplifier generating your clamp voltages . The attack is frightfully slow .
I'd also toyed with the idea of using an 1176 type detector/control voltage amp hooked to the grids of the 6SK7's , but in the end I left the idea hanging , PRR(Paul) was none too complimentary about the 864 , might be worth reviewing his posts , or not :D

If it makes rude noises when you try and compress a source you can always switch to no gain reduction and youve got a fairly good sounding balanced preamp that you can drive to any degree of colour you like , maybe EQing out the bass on the inputs and restoring it post compression might help it behave better on some sources with big low end .

If its spaced out over easy Rhodes piano is your thing Permo , the slow attack and harmonics of the Federal could be just what the doctor ordered .

I 'm shifting towards the idea of solid state Cv amp for speed with 864 style audio path , look at the mighty Fairchild , the CV amp is basically a very big tube power amp , where a simpler transistorised arrangement might accomplish the same thing faster and with less hassle . Course you could always re-configure your Fairchild comp as a combi slow cooker if you wanted be kind to the planet :p


Well-known member
Mar 16, 2011
Hilversum, Netherlands
I am familiar with the 864 quirky behaviour, I have a Dizengoff D864 (mono) compressor, it was really cheap, the compression (the way it handles the gain reduction) is absolutely wonderfull, but I don't like the overall sound of this unit and I think that is due to component selection, especially the cheap transformers in this thing.
It's also slightly different from the original as I can't find the 10W 15K resistors in the Dizengoff ?

My thought was "this could be made better"

I always like to have some spare tubes for my gear, so at some point I ran into a batch of NOS 6sk7 real cheap, so I got a bunch of those old metal tubes.
Wich made me think "I could do a stereo version" and I found the schematics here... so I got going.

While doing some more research and getting my bits together I ran into a nice quad of NOS Sylvania 6sk7GT glass bottles, so I got those and tested them in my D864.... I was pretty wowed !
(They did over 20dB of gain reduction without acting funny, (in that circuit) they only fit with the top lid off wich was fine for testing)
Outputtube will also be NOS Sylvania 7n7's and got some glass bottles for the 6sq7 as well, I think these are NOS Tung Sol.
Trow in the Chicago output transformers and my Jorgen Schou input transformers, all polycarbonate caps in the audio path,.... I think I got a recipe for something really vibey.

I took your advice from CJ's 864 topic and went for 50K input pots instead of the 500K (I have both options) I also decided to make those lin instead of log as they track better.
I like the idea of stepped controls, (these things are expensive !) I allready put some stepped controls on the ouput as I found some really nice 11 step switches in my parts bin... I do have a bunch of enormous 4 wafer EBE switches from an old broadcast installation, but I doubt these will fit.
I have used DACT in my hifi tube preamp, these are really nice indeed but also quite expensive.

I've build most of this thing from old bits that I have been collecting (got for free) for years and I got a really nice deal on those ouput transformers, the meters and pushbuttons and tubesockets are from aliexpress, so I did not spend a lot of money on this, just invested a lot of time...

I'm not shure what this will bring me, I hope it will be sweet on stereobus doing 3-5 dB gain reduction max and glue the mix together in that sweet buttery vari-mu way, that only a real vari-mu can do.

Let's hope so :cool:

Latest posts