Are Polarized Capacitors Necessary for Cathode Bypass?

[ululufut]

I assume not, in that I've seen non-polarized caps in positions other than V1. But not having seen anything other than a polarized electrolytic cap in V1, I want to make sure there isn't some factor I haven't taken into account before assuming a non-polarized cap can be used in this stage as well.

[pdf64]

In order to minimise hiss / hum picked up / generated at the cathode, the input stage cathode should be fully bypassed.
So that needs a big cap, or may want to consider fixed bias with an LED etc.

[JayCobie]

Polarized and unpolarized caps both block DC but pass AC, so either can be (and have been) used as bypass capacitors.

As the input stage is usually the last to overdrive, and also is recieving a relatively low signal level (which is susceptible to noise, hiss, etc), it's role is generally just to impart as much gain to the signal as that very first stage can output. You can then worry about coloration of the signal later in the circuit. Fully bypassing the cathode resistor of V1A with a large cap generally gives the highest gain in all frequency ranges.

This is not a rule, but it's a time tested way of preamp design, hence why you have seen so many polarized electrolytics in that stage.

[kagliostro]

Electrolytic capacitors can differ from their value easily also 40%

other capacitors are less prone to this but large capacitance values are expensive

only now is possible to find supplier that sell it at affordable price

some time ago PRR posted a link

--

Here you can see an example of amp whose filter didn't use e-caps


http://el34world.com/Forum/index.php?topic=14263.0





Franco

[Pietro]

it also functions as a shock absorber.  Higher value caps at input stage give more dynamic sound.  small values =< 10uf make sound more tight and less suspended.

[jjasilli]

I like small value bypass caps, in the 1uF range.  I have used non-polarized caps.  Plastic caps are also available. 

[Mike_J]

Limit I have seen to using poly caps is size can't remember seeing anything larger than .68mF.


In my AB763 single channel amp I used a 25mF cathode cap in V1-a and a three way switch for v1-b. Switch had .68 to 15mF cathode caps. Sluckey suggested I lower the cathode cap to eliminate loose bass in the amp. Worked well. I would recommend it to anyone that is having the same problem. Went to the extent of identifying the outside foil of the .68 cap and oriented it to ground. Don't know that it makes a difference but the amp is pretty quiet so I am happy.


Also used a Solen 22mF @ 630vDC poly cap as kagliostro showed in his picture above as the power cap for the same tube. These caps are huge and expensive as I recall. Didn't notice a tonal difference when installing the cap but shouldn't have to replace it when replacing the electrolytics down the road. Had these on hand so I used them don't think I would buy them new versus an Illinois electrolytic.

[shooter]

here's some fun reading, not so much if you have to pay for 'em 

http://aeaaudio.com/audiophile-quality-capacitors-going-beyond-the-simple-re-cap/

[ululufut]

it also functions as a shock absorber.  Higher value caps at input stage give more dynamic sound.  small values =< 10uf make sound more tight and less suspended.

Is it fair to say then, from your perspective, that in a 5F6-A type circuit it is better to cut lows either by decreasing the V1 coupling caps or adding a V2-A low value cathode cap, rather than to use lower µF cathode caps in the V1 stage?

My concern would be that a cathode cap in V2-A would increase gain beyond what is desired - also not sure if it would have a negative effect on the performance of the tone stack.

[PRR]

A non-polar electrolytic costs more than a polar electrolytic.

If there's no steady DC, you want non-polar. Speaker crossovers, motor start, etc.

A cathode has a reliable 1V-2V DC, so you do not need to pay more for the non-polar, polar meets all needs here.

Its like using rust-proofed screws in a dry inside job: you paid more than you needed to but they will work fine.

If you have a shoe-box of NP caps, sure, use them. My screw-bucket has many coated screws and I don't dig for "the cheapest screw for the job".

[ululufut]

here's some fun reading, not so much if you have to pay for 'em 

http://aeaaudio.com/audiophile-quality-capacitors-going-beyond-the-simple-re-cap/

I'm convinced that there is an audible difference between carbon comp resistors and metal film, however I am not as confident regarding capacitors. The reason, aside from reading in various instances that there isn't an appreciable difference, is that descriptions of the difference between various capacitors usually seems to be vague and qualitative. Is my skepticism misguided in your opinion? I have, for example, read several statements online that tantalum capacitors are an unwise choice for coupling caps. Do you have an opinion on this and other materials, construction quality ranges. etc.?

[ululufut]

A non-polar electrolytic costs more than a polar electrolytic.

If there's no steady DC, you want non-polar. Speaker crossovers, motor start, etc.

A cathode has a reliable 1V-2V, so you do not need to pay more for the non-polar, polar meets all needs here.

Its like using rust-proofed screws in a dry inside job: you paid more than you needed to but they will work fine.

If you have a shoe-box of NP caps, sure, use them. My screw-bucket has many coated screws and I don't dig for "the cheapest screw for the job".

The reason I was asking is that I don't seem to be able to find electrolytic caps under 1µF, so I was considering something else. I actually used a Sprague orange drop as a cathode cap and it worked just fine - just didn't know if it's *always* safe to use.

[shooter]

another fine read for those nuanced folk 

I am not as confident regarding capacitors

 http://el34world.com/Forum/index.php?topic=19534.0
bottom line, while there are tried n true engineering in the sport, EVERYONE has their own slant.  Assume everything in the AC path is subjective (within reason :)
the best way to know, change n check.
start at guitar, strings, pickups, cords, bridge.........end at speaker + box
then build another amp 

[Willabe]

The reason I was asking is that I don't seem to be able to find electrolytic caps under 1µF, so I was considering something else. I actually used a Sprague orange drop as a cathode cap and it worked just fine - just didn't know if it's *always* safe to use.

Yes, you can use film caps for a K bypass cap, they work fine.

Problem is you usually don't need over 5dcv, so film caps at 200dcv or more are larger and cost more than an E-cap, even if their a smaller uF value.

Can we find 100dcv or 50dcv film caps? 

Fender and other amp companies used 25dcv E-caps because they were readily available cap at the time, still are.   

[PRR]

> Can we find 100dcv or 50dcv film caps?

9V stomp-pedal suppliers, not tube-amp suppliers.
http://smallbear-electronics.mybigcommerce.com/capacitors-low-voltage-film-1/
http://smallbear-electronics.mybigcommerce.com/greenie-33-f-to-68-f-bag-of-100/
http://smallbear-electronics.mybigcommerce.com/topmay-tmcf07-5mm-box-68-f-1-f/

Above 1uFd the electrolytic is probably cheaper (though at one-unit pricing the difference may be small).

The electrolytic may not last forever, the film cap probably will. If you are old and mellow, it may be moot. If you are young and energetic, you are likely to re-re-re-build over the next decade, so "forever" isn't permanent.

[dude]

.68uf and 1, so close, the human ear most likely could never hear a difference. So, .68 in an E cap is hard to find, just use 1uf

[glass54]

As we all know, price and size has been an issue in the past, but in serious crossovers space was sacrificed for performance. When I build an instrumentation amplifier (SS) I always use multiple caps in Power Supply (eg 2200uF electro with say 10uF electro and a 100nF MKT) The larger capacitor has substantial Inductance but good store of energy and the 100nF MKT has a much lower impedance to spikes and decoupling is important.
In (SS) signal paths I use a bipolar bypassed with a suitable MKT (eg 10uF plus 10nF.) In Tube circuits I like to add a 100nF 630V across Power supply electros. With 25uF 25V cathode caps I don't usually bother unless it's a high spec preamplifier. In interstage coupling, the values usually don't exceed 220nF 630V, so I am reluctant to use a parallel cap of the next order down. Just use a good Orange Drop  (Sprague or something equivalent).
Please find link to "Picking Capacitors" if you want some bedtime reading. Whilst they are dated articles, were written by a couple of very notible guys especially Walter Jung. http://www.reliablecapacitors.com/oldRC/www.reliablecapacitors.com/pickcap.html
In conclusion, they suggest that the lousy sound in earlier Solid State equipment was mainly due to poor electrolytics 
I use a lot of 50V and 100V Film caps up to 1uF only (usually J rated or 5%) Supplied by Element14/Farnell or RS (in Australia)
Kind Regards
Mirek

[Pietro]

it also functions as a shock absorber.  Higher value caps at input stage give more dynamic sound.  small values =< 10uf make sound more tight and less suspended.

Is it fair to say then, from your perspective, that in a 5F6-A type circuit it is better to cut lows either by decreasing the V1 coupling caps or adding a V2-A low value cathode cap, rather than to use lower µF cathode caps in the V1 stage?

My concern would be that a cathode cap in V2-A would increase gain beyond what is desired - also not sure if it would have a negative effect on the performance of the tone stack.

Not sure about the bypassvap on v2. but 4 the rest that is indeed what me is saying.
I found the magic number to be 33uf on V1 to maintain dynamics, but basically letting the full spectrum through. I squeeze it off with the coupling caps.

For the bassman i would take that route.  Based on 22nf couplers, 33uf bypass is great startingpoint,  or maybe take it down to 22uF max.

250uF is standard? What was Leo thinking?

[PRR]

> 250uF is standard? What was Leo thinking?

I don't think 250uFd is mentioned in this thread?

A BIG cathode cap can reduce hum especially if you get a leaky tube. If you have a "bass" speaker (covers 50/60Hz) and sit there with strings muted, it may make a difference, especially if you make many amps and can't take time to Select non-leakers in the first stage.

Also if you build four preamps (like the first Beach Boys' amp) on one cathode network (to save a half-buck and an inch), without a big bypass cap, some bass at one turned-down input will couple through the common cathode network and appear at a turned-up input. Not a problem when all four Boys are wailing, but may seem wrong in testing.

[ululufut]

When I build an instrumentation amplifier (SS) I always use multiple caps in Power Supply (eg 2200uF electro with say 10uF electro and a 100nF MKT) The larger capacitor has substantial Inductance but good store of energy and the 100nF MKT has a much lower impedance to spikes and decoupling is important.

In Tube circuits I like to add a 100nF 630V across Power supply electros. With 25uF 25V cathode caps I don't usually bother unless it's a high spec preamplifier.

I've read about adding a small value capacitor across electrolytic power supply caps. What is the function of this?

Also, regarding the 100nF MKT you mentioned regarding SS circuits, I always assumed that it was necessary to have polarized capacitors in a power supply. Is there an exception with SS circuits, or has my assumption been incorrect, and polarized electrolytic capacitors are used in tube amps purely because of cost/size/availability?

[glass54]

As you may appreciate, when we design circuits, we usually assume "ideal" components and parameters. In real life,  Electrolytics actually have an additional small Resistive and an Inductive component, thus we get an impedance for the electro rather than a perfect capacitance  . (If you go into the physics, Electrolytic Caps are generally made with 2 plates of Aluminium and an insulator material wound appropriately to fit into a "can". The coils of Aluminum form a small inductor and have leads fitted to them). In addition to the main Electros, we can fit smaller Electros and even Film caps to lower the impedance especially since we are trying to remove or reduce broad spectrum noise not just filter the 60/120Hz (50/100 Hz in Aussie) and harmonics. 
As discussed in this and other posts, It is not necessary to use polarized only caps in Power supplies, but in the earlier days, as space was an issue eg 1000uF/25V Electro is a LOT smaller than anything you could achieve in Film Capacitors. Thus Electros are a preferred option
Getting back to the Tube Amps, size and cost of Spragues/Solen and other Non Polar caps will be an issue (unless your grandfather just left his diamond mine to you  ), we mostly follow exactly what Leo (Fender) did and use the most cost/size effective electros for power supplies with the option of a high voltage Film capacitor (630V MKT) for additional decoupling, as these days the cost is minimal.

I trust this has been useful.
Kind regards
Mirek

[jjasilli]

I too bypass the last filter stage cap which feeds the 1st gain stage(s).  Learned this from "Inside Tube Amps" by Dan Torres.  The bypass cap should be a quality plastic cap.  Value is typically 100 times smaller than the electrolytic being bypassed: e.g.: 20uf main cap, bypassed by 0.2uF.  Test it aurally.  Try no bypass, 1/100, 2X that & 1/2 that. 

[tubeswell]

... I don't seem to be able to find electrolytic caps under 1µF, ...

https://www.tedss.com/2020060246

https://www.bentronics.com/63v-axial-electrolytic-caps-c-16_85_308_309_311_785/068uf-63v-axial-lead-electrolytic-capacitor-p-6659.html

[dude]

OK,  I see one axial E cap, .68uf @ 63v for $1.99 each, don't see a lower or cheaper voltage. I know a lot of Marshall's use a .68uf E cap in first stage K but 1uf and .68, who in the world can hear that difference...? Marshall probably use .68uf at that time as they had a cheap outlet...?

[ululufut]

In addition to the main Electros, we can fit smaller Electros and even Film caps to lower the impedance especially since we are trying to remove or reduce broad spectrum noise not just filter the 60/120Hz (50/100 Hz in Aussie) and harmonics. 

Is this what you were referring to in your earlier statement "In Tube circuits I like to add a 100nF 630V across Power supply electros.."?

By "adding", are you saying run a 100nF in parallel to the electrolytic cap? Also, wondering if you add them to each filter stage, or just to the last stage as was suggested elsewhere in this thread.

[shooter]

I like to add a 100nF 630V across Power supply electros.."?

you're entering into the cork-sniffer world

this is a SS amp PS that I built, never seen much of it transposed to tubes, but shouldn't blow anything up 

[ululufut]

I like to add a 100nF 630V across Power supply electros.."?

you're entering into the cork-sniffer world

this is a SS amp PS that I built, never seen much of it transposed to tubes, but shouldn't blow anything up 

Thanks! There should be a sommelier certification equivalent for people that like to sniff the insides of amplifiers.

[ululufut]

On the subject of bypassing electrolytic capacitors with low value film caps, what is the technical and audible effect of doing this?

[shooter]

to big to post
but you can read all about it
http://buildaudioamps.com/wp-content/uploads/2015/02/AN1849%20PS%20by%20TI.pdf

sommelier certification equivalent

talk with a psychologist, they're real good at certifying cork sniffers   

[Pietro]

> 250uF is standard? What was Leo thinking?

I don't think 250uFd is mentioned in this thread?

i think i saw 250uF in the bassman schematic. OP was reffering to bassman with his question.

[PRR]

> Electrolytic Caps are generally made with 2 plates of Aluminium and an insulator material wound appropriately to fit into a "can".

That sounds like a Film Cap.

Electrolytics are funny. There's one "plate" of Aluminum foil. The surface is oxidized, AlOx is good insulation and easily formed with electric current. The other electrode is a damp paste. This all is put into a can which forms the negative electrode for connection, but does not have the area of the foil/paste interface, is just an ohmic connection 'cause we can't solder to paste.

Making of bipolar caps has changed in my lifetime. It seems they do now use two foils, each making a foil-oxide-paste capacitor, in opposition so they stand reverse voltage.

I stumbled across a moderately informative paper from Nichicon:
https://www.nichicon.co.jp/english/products/pdf/aluminum.pdf

[jjasilli]

On the subject of bypassing electrolytic capacitors with low value film caps, what is the technical and audible effect of doing this?

Real world caps are not ideal.  They have resistance which "should not be there" called ESR.  ESR is the sum of in-phase AC resistance. It includes resistance of the dielectric, plate material, electrolytic solution, and terminal leads at a particular frequency. ESR acts like a resistor in series with a capacitor (thus the name Equivalent Series Resistance).  ESR increases as caps get larger in value, which typically means larger in physical size.  Electrolytic caps have higher ESR than other cap types, I think, probably because of the added R of the electrolytic solution, and because they are physically large.  ESR impairs the ability of the cap to pass (or bleed away) desired frequencies.  Hence, bypassing a large  el. cap, like a B+ filter cap, with a smaller quality plastic cap helps bring actual performance closer to the ideal.

How does it sound? "Talking about music is like dancing about architecture".  Frank Zappa.  You can hear it for yourself - checkout Reply 21 above.

[ululufut]

That answers something I was wondering, which is why ESR, which - relative differences aside, is always a low figure in a functioning capacitor - would have any real perceivable effect. I figured that, say a 1.3Ω ESR on a 30µF filter cap would be negligible. But it would make sense to me - if I'm thinking applicably - that a small resistance could alter the cutoff frequency - ? Either that, or I'm guessing that maybe the internal resistance somehow either changes the bandpass slope, or makes it irregular.

---

Having just read the Jung & Marsh article, it does appear that ESR and voltage capacity are relevant regarding THD, so it makes sense to me that, for anything other than an ideal capacitor, there will be phase disturbances and degradations of harmonic content.

As famous as that Zappa quote is, while it's figuratively revealing, in reality you can't dance about anything. Here though, we're talking out perception of audio signals, not the interpretation of music. There is some overlap between the two, however whether it's talking about the relative functionality of capacitors or whether vitamins are worth taking, while the subjective experience may be difficult to express verbally, the factors - if they exist - that contribute to the usefulness of various materials can at least in theory be expressed.

[glass54]

Just adding a little more bedtime reading  https://www.avnet.com/wps/portal/abacus/resources/engineers-insight/article/understanding-esr-in-electrolytic-capacitors/ to take ESR, ESL, C and Rp a little further.
Some of us make a fuss over Solen, Sprague, etc capacitors to achieve transparency as well as quality Non-Polarised caps in SS HiFi components. It is important to choose correct capacitor types for quality audio and in tone control stages to achieve stability in sound especially with temperature variation/rise of tube based equipment. 
After all, we go to great trouble to build our toys so its important that we use good/stable/reliable components for years of enjoyment  (This does also include how and where we run our wire for optimum low noise.)
Kind reagrds
Mirek

[jjasilli]

I'm not good at verbally expressing musical sounds or the tastes of food or wine,  e.g.  It's worth your own listening test. 


On another note people dance about stuff all the time, with the glaring exception of architecture. 

[ululufut]

On another note people dance about stuff all the time, with the glaring exception of architecture. 

LoL!

[shooter]

After all, we go to great trouble to build our toys so

... we can dime them, crank out sounds that cause design engineers all manner of grief, blow all theoretical ideas out the speaker...