General question re: recapping

  • Thread starter Thread starter lo.fi.love
  • Start date Start date
lo.fi.love

lo.fi.love

Functionally obsessed.
So, I'm really curious to know more about replacing electrolytic capacitors.

Granted, I know it should be done on older audio equipment, e.g. equipment older than 10-15 years, because the capacitors do eventually fail. I'm curious to know, though, what the benefits are to replacing electrolytics which haven't yet obviously kicked the can (pun, sorry).

That is, I'm asking, what can I expect, or at least look for, when I've replaced electrolytics on ATR amp cards, mixer channel cards, and audio device power supplies? A gain in performance, and of what kind? I'm looking for tangible (audible?) results here, something almost concrete that can be quantified to some degree.

Examples from your experience, with specific items of equipment and a comparison of before and after, will be greatly appreciated. I want to open up this topic here to get a general consensus of the benefits of this practice.

Ok, let's have at it! :)

Cheers,
Jeffrey
 
On 10~15 year old equipment I'd say it is mostly an insurance policy against failure.

Maybe even on the 20 year old stuff.

More than that? Depends on what's under the hood (does the cap say "Beaver" on it? REPLACE IT...NOW), and how well designed the equipment is.

Seems like the Teac/Tascam stuff was thoughtfully engineered and cap values are well appropriate so they aren't being stressed. That'll shorten the life. Seems my old Ampex stuff were VERY well engineered...good quality caps for the day...40 year old power supplies still producing very clean power. Not so with everything. My Roland DC-30 analog delay/chorus box had undervalued filter caps of a not so great brand. Hard to tell if it made a difference because its such a noisy dirty beast anyway but at least I don't have to worry about the PSU taking a major-league dump on everything downstream.

I have yet to find a cap I've pulled from any Teac/Tascam gear even over 25 years that had drifted outside of its rated value, but the only way to really test an electrolytic capacitor is to measure its Equivalent Series Resistance or "ESR". A perfect electrolytic capacitor would have an ESR of zero. ESR has an effect on, among other things, the HF performance of the cap. ESR increases as the cap ages. I haven't swallowed hard enough yet to cough up for an ESR meter.

So, enough of that garbage...

Assuming you are talking about replacing caps with new ones of the same values, the answer to the question "what will happen" really depends...my personal experience has been more with the results of recapping power supplies and my most notable experience was a very noticeable drop in the noise floor on my M-520...I've also had issues with mic trim pots that were crackly and after liberal cleaning didn't seem much better but issues cleared right up after a recap of the channel card.

That's not really what you're after though is it...

The thing to realize here is that, at best, electrolytic capacitors in the audio path are a compromise...they don't do the best job of passing high frequencies...film caps do but a 100uF film cap would be...gihugic. What I DO know is that newer 'lytics generally do a better job with HF signals because, in general, ESR performance has improved in newer caps and supposedly "audio-grade" electrolytics have had some attention paid to that application...That's why I use the Nichicon KT series where I can. Hi temp and "audio-grade". Does that make a difference? I have NO idea, but the price difference is pretty much nil and they are a cool light blue color and I figure it couldn't hurt.

I haven't done it yet but supposedly replacing the output coupling cap on my Ampex AG-440/MM-1000 amp electronics with a Nichicon PW cap in 470uF/63V flavor (stock is 250uF/50V) sounds like "taking a blanket off the monitors." Dunno...the old Ampex already sounds like blankets came off so...what...blankets will start disappearing in my house if I replace those caps? That would make my 10 year old daughter very upset...she's like a blanket magnet...she hoards them...like a blanket version of The Princess and the Pea she's got a half dozen or more of them on her bed..."What happened to all the blankets?!?" sez dad looking at the empty blanket basket in the front room..."On Maddie's bed." answers the chorus of her siblings...

Listen...I'm not really smart enough to contribute any more and besides that I'm too conservative. GCalo will be on here in a sec telling you to bypass all the 'lytics in the audio path with good quality film caps (in general the idea here is to compensate for the electrolytic HF compromise I mentioned above), others will have suggestions to increase the value where you can to open up the bottom end.

I don't know...I just don't know...not doubting, I just have to understand something before I'm comfortable with it and all that stuff is over my head which is why my recap jobs are generally plain jane.

I'll get out of the way now so I can learn something.
 
Most of the recapping I've done was because of failure in the system. I would find the bad cap and replace it, or replace all of them if they were all old. I have no equipment to measure any of them so, it's sort of hit or miss, or eying the bad cap. I've always had good success and fixed many problems by recapping.

I'm just about done recapping the entire M-35. I DO notice an improvement in the sound. However slightly, it's a little more powerful, clear and upfront. An old TEAC repairman told me that caps of that age would only be operating at about 1/3 of their original value. I don't know how true that is, as Cory said, he has found most to be where they should be. But I suppose even if you hit on a couple weaklings it will help out.

I'm still looking for some new sure ways to pimp my console. :D
 
Ohooo Recapping

Recapping debates could entertain you for days. There are lots of posting here and elsewhere (prodegy pro) on recapping. I'll just break recapping down into a few broad areas.

The first division is repair/prevention and enhancement.

Caps can fail in a number of ways. Some blow up, some ooze and some make noise or other odd things. In any case you decide to replace the caps and may or may not change the value (enhancement).

Power supply caps do a number of things. They take the rectified DC and smooth out ripple. They shunt to ground any noise that is on the supply rail and of course supply the current needed in your circuits.

So if you have old caps in your PS that no longer do a good job (or never did...) of shunting noise to groung then you can improve the noise floor by replacing them with good power supply caps and get a little lower noise on the supply rails by putting some 0.1 or 0.01 uF poly caps in parallel with the PS filter caps. Big caps may not do a great job at high frequencies. That is why many put in the poly caps.

There are local filter caps on the circuit boards. You will see them across the supply rails. Low ESR is desirable. What does that mean? ESR tells us the resistance of the cap when it is conducting. caps block DC and they block AC below their cutoff frequency. This cutoff frequency is related to the caps value. (larger caps conduct lower frequencies) When the cap is conducting it does so as if it were a resistor. A low ESR cap conducts with a lower resistance. And thus is better at shunting the noise to ground (all a very simplified description)

All this is for power supply and bypass caps. we have not even gotten into the signal path yet, let alone EQ circuits....

In the signal path a low ESR cap has a low series resistance impeding the signal (Duh) Bu replacing the input coupling cap or an interstage coupling cap with a larger cap you will pass lower frequency signals (better bass) and if you bypass it with a poly cap (0.01) you will have better highs. One of the mods in the Yamaha PM1000 is to change out the input coupling cap.

So coupling caps go larger with a parallel poly for better low end with great top end.

Lastly there are EQ mods. Changing to low ESR caps without changing the value will give a "crisper" EQ section. More efficient if you will. Changing the values will change the frequency response of the EQ section and shouuld be done carefully to move the cutoff freq to a better place. For example if the high freq EQ corner freq was at 15 kHz you might want to change it to be 12.5 or 10 kHz. (PM1000 again)

No need for an ESR meter in my mind. But good brand high quality low ESR caps and if you pull a cap replace it rather than test to see if it is good.

Also, there has been some debate as to weather "audio" caps are better for audio than low ESR caps. For me I decided that low ESR is just as good as the "audio" caps and will buy whichever costs less and fits the position on the board. OBTW todays caps are smaller than yesterdays caps so don't let that bother you.

Regards, Ethan

PS take it all with a grain of salt, do some research and let you rears tell you.
 
Ethan, that is awesome...all the thinking and work I've done the last few years helps me understand clearly what you are saying. I realize you are presenting things in macro fashion but it really breaks it up in a digestible way.

One question just to confirm...a "coupling" cap is one that is right in the signal path yes? Like for instance between an input jack and the input of an opamp right (or in the case of "interstage" that would be between the output of one side of a dual opamp and the input of the next stage, yes)?

And it was stated here recently that there is a benefit to replacing polar signal-path 'lytics with bipolar (non-polar) versions...here...this post in this thread...

Now...I thought that if a cap was polar you should replace it with polar, and if it was bipolar, you should replace it with bipolar...whaaaaa?
 
sweetbeats said:
And it was stated here recently that there is a benefit to replacing polar signal-path 'lytics with bipolar (non-polar) versions...here...this post in this thread...

Now...I thought that if a cap was polar you should replace it with polar, and if it was bipolar, you should replace it with bipolar...whaaaaa?
Click to expand...

Waiting to hear too!

I've also got some other questions about capping. But I might need to get some pics first.
 
Any recommendations for the poly caps to try out?

Would these do? http://www.mouser.com/ProductDetail/WIMA/FKP2-01-100-25/?qs=sGAEpiMZZMvCt%252bwg%252braTutQnbjK%2fu%252b%252bPgZf0wJUrBU8%3d
Or should I go for a higher value?:confused: I don't think I have anything higher than 50 volts in the board.

I don't want to go nuts, but I might try it out on the power supply and a couple of channels to see what it does.

Oh, Cory, btw, I used the KT caps for all the caps in the M35. Thanks for the recommendation.
 
My $.02.

There's more to a lytic than measuring it's ESR after a certain time period.

Lytics are like a pressure vessel in an airplane. They are only so many safe cycles. Sure it will work beyond the desinated period but at what expense.

Lytics take a beating w/voltage fluctuations and loads. So why live with an old "cap"? Replace and enjoy with grace!!!

In an audio circuit we always bypassed low ESR NP caps waith a 0.47 film. This aided in further blocking DC electrons from passing to the higher freq drivers, and it greatly smoothed the audible signal.

A 12dB/octave network used an inductor (choke) to aid in passing the DC electrons to ground to protect the HF drivers. Not available in electronic circuits on cards.

So as Dr. Ethan indicated there are values to bypassing lytics in electrical circuits.

I am a big believer and it sure works well for me.
 
Greg? Ethan?

Either one of you willing to take on the question about replacing signal path polar 'lytics with non-polar?

Anybody willing to take on a case study if I put up a schematic of an input strip and what you would change out and with what?
 
I would defer to Ethan on this as he has far more board level experince than I.

As a rule of thumb if they are in the signal path you are generally ok but rely on Dr. Ethan.
 
Been out a while ....

Electrolytic caps do not like being reversed biased. Which is to say having the positive lead more negative than the negative lead. If you do this with some power they blow up.

In audio the power levels are >usually< quite small and some designs allow electrolytics to be reversed biased as the audio signal goes negative. This may work fine for a while or even for a long while but I would expect the cap to mis-behave when reverse biased.

In days past bipolar electrolytics were either not available or expensive and thus not used in run of the mill equipment.

Bipolar electros are in essence 2 polar caps back to back and indeed some older equipment (Ampex ATR100 for example) used back to back electros. Of course when you put caps in series the capacitance is cut in half given equal value caps.

But back to the question - in the signal path you should replace polarized electrolytics with bipolar ones where the signal drives the positive lead more negative than the negative lead. (give no holds barred and a bunch of weasle words)

Some designs kept the positive lead positivly biased and thus were fine. THink of having phantom on all the time. The electro would be at +48 and thus never go negative. But turn off phantom and you could be reversed biased.

--Ethan

input caps are as you said, interstage are as you said (between IC for example)
 
Interesting thread. :)

I have an old Sony reel to reel my Grandma bought in the late 50's, never been recapped but sounds fine. Original tubes too.

I have a probably 1970 or so Alamo tube guitar amp and it works with the original caps. If I replaced the caps in that would I expect to hear a difference?

There's one major can near the tubes, would that be the main baby to want to replace for possible sound improvement?
 
I have a cap in my Traynor reverb, 015 (fA) I could be mistaken on the "fA" though, it's probably uF. Can anyone tell me, is the value "015" the same as "15"???
 
If it is marked .015 that's a whole lot different from 15!

Be absolutely certain.
 
GCalo said:
If it is marked .015 that's a whole lot different from 15!

Be absolutely certain.
Click to expand...

Yeah, I know. It would make more sense if it were .015, but I don't see any decimal point. I don't have the schematic so I'm just going off what I see on the cap. I'll have to look at it again, even closer and maybe take a pic.
 
Consider this, Steve, that it is not usual for a 15 mfd cap to be preceeded by a "0".

Be absolutely certain.

By the way, have we answered your questions, Jeff?
 
evm1024 said:
But back to the question - in the signal path you should replace polarized electrolytics with bipolar ones where the signal drives the positive lead more negative than the negative lead. (give no holds barred and a bunch of weasle words)

Some designs kept the positive lead positivly biased and thus were fine. THink of having phantom on all the time. The electro would be at +48 and thus never go negative. But turn off phantom and you could be reversed biased.
Click to expand...

Butbutbut...how do I know if the signal drives the positive lead more negative than the negative lead??
 
I would be quite interested to hear about reccomended/best practices for removal of components form PCBs. While I have quite a bit of soldering experience (building tube guitar amps and mic pres) I'll be darned if I didn't manage to be heavy handed and lift some tracks on the PCB when I unsoldered some ICs on my Soundcraft series 2 board (circa 78).

Jim Williams has mentioned the Hakko desoldering tool as a must - anyone care to toss their hat into the ring as to their best technique??

Maybe I was just too impatient!! I'd love to re-cap a few channels to see the effect... but the IC incident put me off for the time being!!
 
sweetbeats said:
Butbutbut...how do I know if the signal drives the positive lead more negative than the negative lead??
Click to expand...

The easy answer is that you measure the DC across the cap to find the DC bias (harder than it sounds). It may be zero or it may be many volts. If it is zero then any AC signal will drive the cap reversed bias half the time. If there is (oh say) a 1 volt DC bias then any AC signal of 0.707 volts RMS will be 1 volt peak positive and 1 volt peak negative. So on the positive half (assuming sine waves here) you end up with 2 volt across the cap and on the negative half you end up with 0 volts across the cap. (ignoring time constants) Increase the AC voltage just a little and the negative half will bias the cap negative. And then you get whatever bad sounds the cap is "going" to make....
 
Would it not be possible to assume that if it's not "polar" and it's in the signal path that it could be replaced w/an equal value non-polar (bipolar) film?
 
You must log in or register to reply here.
Back
Top