No EL tube amp - PP film filter caps

[tedmich]

I couldn't help myself, I had to buy these:


four metalized polypropylene 60uF caps, 330VAC 2" x 2".....they were <$5 each....


So now I can make my electrolytic less (ELno) amplifier a reality!

He still has 9 of these left:


3.5" OD! 40uF but almost $15 each...
ebay 170369838377

[Chuck H]

Whoa, those ARE big. No doubt worth the extra chassis space in 30 years when you've never had to recap the amp though. Nice price, great find.

Chuck

[dai h.]

might sound worse possibly guessing from my experience. IMO it'd be a good idea to buy several different values since it seems you might have to tweak the value when going to film from regular alu. electrolytics(might also be good not to come to a final judgement without tweaking values also). This sort of thing putting in technically superior parts does have a seductive appeal but in practice, I think it can go either way and from what I've seen and read sometimes worse (distortion or whatever) can be better in a given situation. Price is most definitely right for experimenting though.

[cbarrow7625]

Dai, I know what you saying. I had thought in the past that it would clean up the sound, noise, whatever in my amp if I bypassed the electrolytic filter caps with larger (1-2uF) poly caps. What I found was that the poly caps worked....too good. The distortion channels became sterile, lacking in upper harmonics and liveliness. The clean channels became boring with little life as well. I just chalked it up to the higher ESR and inductance in electro caps not bypassing upper frequencies as well as lower ones. The flaw in electro caps seem to actually be a benefit in creating our tone (as apposed to just reproducing them). No surprise there right? A lot of the gear we use, especially vintage gear, is terrible "flawed" from an electronic design point of view but sounds terrific.

I still want to build an amp with the giant poly caps just to see if these caps work better by themselves as opposed to paralleled with electros.

[dai h.]

I tried "Y" ing the power supply to V1 in my guinea pig Marshall (basically just an ordinary 1987/59 lead circuit) and making the normal ch. de-coupler a 3uF film, and it sounded okay actually, but possibly might be considered too smooth--maybe not "nasty" enough subjectively.

Definitely it seems a lot more complicated than using the conventionally "superior" part when it comes to sound. Been reading a lot on ceramics (bit of a recent fixation), and things that I've found interesting are that class 2 types can apparently change value with signal, plus this can also apparently be influenced by the voltage rating. Some people say not to use them, but it does seem in practice they can be a plus and some prefer them in say Marshall circuits, certain fuzz pedals, etc. When you look at the original vintage marshalls, it's even a bit more complicated since some used class 1 (temp.compensating which are according to some sources good for audio conventionally) and other ones didn't (judging by markings and size--class 2 are volumetrically efficient so I think size is a clue). And tantalums, which some also say sound bad I recently found out were used in Neve preamps, so I guess things are not so simple as some parts being simply usable or not for audio(and a funny thing seems to be that the Neve designers were trying to achieve technical correctness yet today the vintage units seem to be considered attractive examples of "colo(u)red" sound). If there is any sort of lesson I suppose it seems to be that listening is important (of course not to say the technical stuff isn't).

[Chuck H]

I think dai actually did an experiment some years ago with small value resistors in series with the B+ filters to see what the effects of higher esr were. The overall effect will change with design because it does depend a lot on which stages are sharing a filter as to what's talking to what. When building an amp with big film caps it might be an idea to get a handful of low value resistors and experiment with them to simulate the esr of true electrolytics. You could even use small value caps jumpered between the resistor nodes to simulate the high frequency cross talk. Sounds like a long tedious tweak, but not too tough to design in.

Chuck

[cbarrow7625]

Sounds like a fun, tedious experiment. Inductors bypassed with resistors would probably be more accurate to simulate the electros. As if there weren't enough variables already!

[tedmich]

a few of the audiophile people have tried film filters, as their hate for ELs is very great (cept silk and hemp dielectrics...Elna RFS...)

They report a "tightness" in the sound. This is maybe similar to what Steve Conner reports for high EL filtering, you lose the loosness that most guitar PS designs provide. Smaller film caps might be be the worst of both worlds; hummy AND tight and sterile...I have heard similar arguments for slowing down the rectifiers...

Like ceramics, tantalums are similarly a "red headed stepchild"* for tone connoisseurs, although Mesa uses them a good deal (the exception that proves the rule?)

[cbarrow7625]

I use tantalums in the cathodes in some circuits. Sometimes they work, sometimes they don't. In clean-ish circuits I don't care for them - too sterile. In high gain channels they can clean things up quite a bit, hence probably why Mesa uses them. I also use poly caps in the cathodes of high gain circuits often, and occasionally in clean-ish circuits - depends on what I am going for / what the amp needs (wants?).

Every cap probably has its place if it is in the right place in the circuit. Maybe a poly cap for the main B+ node or B+ & screen nodes and electrolytics for the preamps? I've never tried it. maybe they would shine in that application and make the power tube stage extra tight and exciting, especially in a cathode biases amp. Hmmm, I might have to try that just to see.

[Chuck H]

Sounds like a fun, tedious experiment. Inductors bypassed with resistors would probably be more accurate to simulate the electros. As if there weren't enough variables already!

Yes, I thought of the caps across the resistor nodes as a psuedo, similar effect to avoid the use of inductors and to reduce parts count and circuit complexity.

I think the point is not to design a better power supply, since the ones we use now seem to work and sound just fine. But rather to build an immortal power supply (relative term) so the only maintenance the amp ever needs is tube replacement. For a personal amp, speaking as the owner of a few amps I built over ten years ago, I think it's a very worthy quest.

Chuck

[tedmich]

Yes, I thought of the caps across the resistor nodes as a psuedo, similar effect to avoid the use of inductors and to reduce parts count and circuit complexity.

I think the point is not to design a better power supply, since the ones we use now seem to work and sound just fine. But rather to build an immortal power supply (relative term) so the only maintenance the amp ever needs is tube replacement. For a personal amp, speaking as the owner of a few amps I built over ten years ago, I think it's a very worthy quest.

Chuck

that's pretty much what I am going for, a bullet proof "forever amp" true stereo 2x (3x12ax7 + 2xEL84 PP) with one voiced high and one lower, both high gain, 2 tuned port cabs.

Supposedly Billy Zoom's personal amp is made with "forever" specs(also UL taps and no volume knob?!) He has some interesting amp clips at Gretsch
The Gretsch Pages: Discussions: Billy Zoom's Jet Set: A Vist to Zoom Studios - and a listen to the prototypes.
talks too much on some...

[Albert Kreuzer]

Wasn't R.G. designing what he called "the immortal amp"?
IIRC it was called the "Workhorse". I think he used motor caps in the power supply. Does Visualsound still make them? I can only find pedals on their website.

Cheers,
Albert

[loudthud]

So what does the forever amp use for cathode bypass? I'd hate having to use something bigger than a 12AX7 to bypass one cathode. I found these. Never tried them but the price is low. Only available in surface mount but you could just solder leads to them.

http://www.avx.com/docs/Catalogs/noj.pdf

[tedmich]

So what does the forever amp use for cathode bypass? I'd hate having to use something bigger than a 12AX7 to bypass one cathode. I found these. Never tried them but the price is low. Only available in surface mount but you could just solder leads to them.

http://www.avx.com/docs/Catalogs/noj.pdf

could work, OsCons are technically not ELs eh?
Wima MKS-2XL come in 16v
10 µF 8.5x14x7.2mm
15 uf 8.5x14x7.2mm
22 uf 11x16x7.2mm

for 4~$8 each....

oh and tants last a long time (if they don't splode) and Mesa uses them in cathode bypass

[dai h.]

I think dai actually did an experiment some years ago with small value resistors in series with the B+ filters to see what the effects of higher esr were. The overall effect will change with design because it does depend a lot on which stages are sharing a filter as to what's talking to what. When building an amp with big film caps it might be an idea to get a handful of low value resistors and experiment with them to simulate the esr of true electrolytics. You could even use small value caps jumpered between the resistor nodes to simulate the high frequency cross talk. Sounds like a long tedious tweak, but not too tough to design in.

Chuck

I forget exactly what I did but one picture I took of my guinea pig amp I can think of tells me I did at one time use very small values in series (something like 0.27ohms). From a previous discussion about this (if anyone wants to look it up), R.G. recommended a range of values (specifics escape me at the moment--possibly 15 to 50ohms??--info was copy/pasted but on another PC which is presently down at the moment).

a few of the audiophile people have tried film filters, as their hate for ELs is very great (cept silk and hemp dielectrics...Elna RFS...)

They report a "tightness" in the sound. This is maybe similar to what Steve Conner reports for high EL filtering, you lose the loosness that most guitar PS designs provide.

from what I've read not all audiophiles take such an extreme position, and some (after trying Muse, etc.) prefer regular (inexpensive) alu electros.

Smaller film caps might be be the worst of both worlds; hummy AND tight and sterile...I have heard similar arguments for slowing down the rectifiers...

when I tried my experiment for V1 it didn't make any difference in noise (even though 3uF is smaller than the usual 8uF-50uF found in that basic 5F6-A/Bassman circuit). The thing is though (I think) by the time the B+ gets to that point, it's been filtered, so mainly the purpose of that cap is more about completing that stage's signal loop to itself more than reducing residual AC on the B+ line (i.e. "ripple") so that's probably why S/N wasn't affected. Also films can spec out differently (that is can act as capacitors better depending on frequencies better than the same value alu electrolytic), so lower values might be okay to use. I remember measuring some 1uF alu electros on my LCR meter and found it interesting since they all seemed to be about 0.68uF at 1kHz, but at 100Hz,120Hz measured 1uF(kinda makes sense as multiples of ripple freq.). (From the same previous discussion ref'ed about series R with filter cap, IIRC R.G. stated the specs could be diff. depending on value--the gist of this seemed to be that you could not always count on say going down a bit when changing from alu elec. to film--memory is a bit foggy so may be misremembering here though).

edit: (some links I found interesting looking for info on caps. Have more but on another PC)

http://www.prodigy-pro.com/diy/index...e;topic=2287.0

http://www.prodigy-pro.com/diy/index.php?topic=11838

Capacitor Type Selection Optimizes PC Sound Quality for Windows Vista Audio Requirements - Maxim

http://www.maxim-ic.com/appnotes.cfm/an_pk/4333