Recap, restore, or modify: What, why, and how (an Engineer's perspective)

SaturationPt

Big Woofer Club
Subscriber
Of all these, recap seems to be the most debated. Some swear that no values should change, some that everything should change, can we at least agree on the terms?

A short bio: I built my first circuits from kits in the late '60s / early '70s: first an Eico "light organ" then an Eico oscilloscope (round CRT, single-trace recurrent trace), was always a stereo-head and spent much of my teens in a local high-end stereo shop (Pioneer, Marantz, Kenwood, McIntosh, Philips, Dual, JBL, Advent, et al), worked as a Tech for a stereo shop during college, electronics Design Engineer (R&D and high-volume production) for much of my career. Today I enjoy the hobby of collecting, listening to, servicing and restoring stereo equipment both vintage and new.

The concept of recap is the most ambiguous to me as it seems to involve replacing some of the capacitors in older equipment under the assumption that the existing capacitors are bad. It is a term often perverted by those who buy and sell to get a higher price, I even see a shop in Ann Arbor that claims to "re-rate" the output of his "serviced" Marantz products at a higher output because he has added additional capacitors to the power supply filters. This is to me neither a restoration nor a recap, it is a modification.

Recap to me means replacing all suspect capacitors (usually this includes all electrolytic and tantalum in older equipment) with the same rating new parts, period.

Restoration (electronic) to me means a recap, and adding replacement of all suspect semiconductors (those known to have problems) with like-specification new, and testing all remaining, testing many critical resistors and replacing as needed, testing and replacing diodes as needed, cleaning and replacing all thermal paste (or pads), checking all solder, a thorough cleaning including cleaning and lubricating potentiometers and switches as needed (some contacts should remain dry, some are designed to have lubrication), and replacing relays. In some units this also includes light-bulb replacement and velum replacement. Then all voltages etc. are checked and adjusted, any calibrations made. Any Factory Service Bulletins should be part of this restoration.

Servicing is harder to define but is at least a thorough cleaning of switches and pots, removal of dust and dirt in the unit, repair of anything not performing as new (need test equipment for this), adjustment and calibration, and addressing any FSBs.

Modification is anything deviating from the original spec. In most older equipment the original spec parts are not all available and substitutions need to be made and I'd categorize necessary (by availability) minor and calculated substitutions as restoration. Substantial changes or the common "bigger is better in caps" is to me a modification.
 
Last edited:
I want to blog a bit about capacitors and why I stick with original values. I know that many people like to increase values (both capacitance and voltage), and if the design is well understood then it can be okay to good. However this is drawing from my perspective and experience.

When we designed circuits, we designed them to perform well as a complete organism. The power supply, preamp, amplifier, filter stages, all have to work well together and changing one has the potential for unintended consequences in another.

I'll focus first on power supply.

There is an assumption that you can increase the size of the PS filter caps and this will increase your amplifier's power output. There are some cases where this is correct, but I'd say that these are rare. Since power output is the most marketable specification on brochures with the second being the THD spec I will say that there is no case where we would have the opportunity to increase the power output by even 10% by up-sizing a couple of filter caps or we would have done it. Again the additional $0.50 for a bigger cap to add maybe 10wpc? Huge return on investment and marketing bragging rights.

Look at the whole amplifier. If you double the size of the capacitor, it has twice the energy storage. That means that the transformer and rectifier bridge needs to supply twice the current on startup, which they might not be able to do for long if at all. If they could, storage wouldn't be a big factor because we have enough supply to keep the caps "full" for LF at least (the most demanding for power usually). So I'll call this a myth, and I feel qualified to say so. If your caps are out of tolerance and you replace them with new in-spec you likely will hear a difference however.

Continuing to talk power supply filter caps, what will the amplifier do with this additional power if the transformer and diodes are able to keep the capacitors charged?

The biggest current path (you have created additional available current in theory) is from the power supply through the output transistors to the speakers. The wires are probably fine, the protection relay is probably fine, even the circuit-board traces are likely fine with this additional available current but what about your output transistors? A transistor is limited (for this example) in three ways: Internal heat (power) dissipation, external heat (power) dissipation, and maximum current capacity.

Heat dissipation is simple, the voltage drop across the junction is .6V, double the current and you double the power (watts/heat, Power = Volts x Amps). Externally you can add a fan or larger heatsink, but internally you can't change the maximum amount of heat dissipation from the junction to the case. If you go over you get thermal runaway and destroy the transistor, or at least significantly shorten its life.

Current limitation is like a fuse: once you go over it burns open.

Power supply capacitor changes however are the easiest and have the least risk of change to the sound of the amplifier, minor changes (as available parts aren't always exact) usually have no unintended consequences.
 
Last edited:
Other capacitor changes as well as semiconductor changes:

Every part of your complicated audio equipment was designed, tested, re-designed, tested again, once everything behaved as it was supposed to and the validation lab / test lab could no longer find a way to break it, it was released to production. Some things have changed through this process as theory doesn't always predict real-world results, but we have the tools and equipment to create accelerated aging, push beyond specs, and measure the results.

It isn't just power output, sure more is better. When it is being prototyped and tested, or even on a proper test-bench many measurements are made. I can look at the waveform from stage to stage, component to component, compare the wave shape, amplitude, even phase changes. I can look at a specific frequency, or how white or pink noise input can affect the output now with a computer comparing the entire spectrum and highlighting or graphing deviations.

In the old days we could use the horizontal and vertical input in a scope and get a shape on the screen that shows phase change as we sweep frequency, overlay two waves in the scope to see if they match or where distortion is occurring (and whether it is in spec). Not as accurate as today's methods! Point is, very slight changes to components in the signal path can change how the amplifier behaves and how close the input is to the output. You might think for example that a preamp or driver transistor with a higher frequency capability is better, however you might now pass a frequency downstream that was never there before and was never compensated or filtered. You could cause a harmonic or oscillation that never was a problem in the original design and testing. How will you know? The easy answer is that you probably won't unless it has a significant sonic output. No circuit is perfect by itself, feedback and other design elements are there to compensate from one imperfect component to the next so that the result is as close as possible to perfect.
 
Last edited:
So for those among us who like to DIY, how do you recap or restore your gear?

If you're lucky, you've chosen a piece that has a good following here, and there's a "recap list" or other BOM that has been vetted by one of our many experts on that piece. For me that's my starting point. I like to go through the list piece by piece, comparing to the original and making changes where I feel that a substitution is not the best, but so far I have found these lists to be acceptable and to follow the original designs. Most people are satisfied with using these lists as-is, I recommend this for all but the most obsessive or bored Engineers, Designers, and Technicians.

The other option is to work from the Factory Service Manual (FSM) and build your own BOM. This is a difficult and tedious task if you understand electronics, vintage components, and audio. Impossible if you don't, you will need help.

The changes that you hear after a good restoration will be well worth the effort, typically you will make it more valuable, more reliable, and you will appreciate the sound of your equipment as much as the music itself.
 
Last edited:
I restore old amps as a hobby on nights and weekends. I really enjoy the work and it's better than computer games and tv watching of which I used to do a lot of. I have an inventory of Nichicon and Elna caps and Wima and Kemet film caps. After troubleshooting problems, I just go through and start replacing all the electrolytic caps with new ones of the same value. I may bump up the voltage ratings, but usually I just replace with the same value. Pretty much every amp I do there is at least one cap that shows signs of leakage on the legs or on the board that is not visible when installed. It's fun for me to go through the process of starting with a broken old amp and bringing back to new life. I often can detect improved performance as i'm doing the restore, but not always.

The best experience I've had to date is with a Pioneer SX-550. I had one given to me and I decided to sell it. Everything worked, all the lights worked, but it sounded really bad, Dull and lifeless, like a rug thrown over the speakers. I couldn't accept this, so I thought I would go through and replace all the caps including those little blue tantalum caps. I wasn't crazy about the Idea because it was a 20 watt pioneer and I didn't think it was worth the time. Anyway, I started with the Preamp/Phono section, replaced everything and then listened, no improvement, did the power supply and filter caps, no improvement, did the amp section, no improvement. I was so disappointed. I ended up removing the faceplate and the metal cover under the faceplate for some reason and there was a tantalum hiding behind the tone controls. I replaced that and my gosh, the sound just opened up. I was stunned on how good this receiver sounded. That is why I do the work.
 
I tend to do mostly tube gear.
I'm of the school you can bump PS capacitance some (the factory ratings were +/-20%.)
I've dabbled a little in SS gear. If I can change a tantalum with a film , I will (small values usually work.)
As far as e caps I try and bump voltage a level or 2. (IE...35 v may become a 50v , maybe a 63v)
I generally don't bump value more than 20%.
Sometimes the new smaller caps, it's tough to find a like (V, C, ) and still match physical dimensions.
I readily admit the people who designed and built it were smarter than me.
In some cases I thought they could have used a better product. I understand there are factors in play that aren't apparent today (availability) bean counters saying you don't need so much leeway or quality . We can save a quarter a unit if we use this instead of that.
In fairness, we work on 20-60 year old gear .
I don't think the manufacturer thought these would still be working (or even around) so the reliability point is somewhat moot.:idea:
I can't say I ever thought upping PS caps was going to up output.
 
Last edited:
This is an interesting post and I'm surprised it hasn't gotten many more replies.
I like to hear from people with design experience who are also in the hobby as it brings a quite different perspective to what you usually hear.

I personally restore my gear.

I will change like for like unless something is under rated for it's purpose (like a cap within a few percent of it's working voltage) to help it last longer.
I read many threads where people change electrolytics for films or certain diodes for others and such but I personally don't have enough know how or enough equipment to tell if what I'm doing is an improvement or not. I feel the team that built it in the first place must have known what they were doing and I'll defer to their expertise.

Just recently I watched a video on YouTube of someone building a Class A amp board and with the substitution of what should have been an acceptable pair of outputs the amp had major problems with high frequency stability. If that's just outputs then retransistoring the entire amp with subs, even if they are close, seems like it would almost call for a redesign. For this reason I leave all devices unless I absolutely have to change them in which case I try to make as careful a sub as possible. I also have a decent stash of NLA higher gain devices from back in the day so, at least for my own gear, there will hopefully be no subs if a failure should occur.

With the BOM, most of the stuff I work on doesn't have much info on AK so I go in to the amp myself to make one. I have been burned too many times by changes made after the SM was printed so I verify at least two times before I place my order and then one more time when I place my second order after I realize I missed a few anyway.
 
I have restored and recapped around 15 receivers and amplifiers from 1970s. I think people that designed them would be a bit surprised reading this about how meticulously that gear was designed maybe would have a laugh or two.

I started replacing caps tit for tat same type and specs. I'm maybe the only guy here replacing ta tantalum with tantalum.

Then I measured voltage under some and I winced. 50v rated caps sitting on 51 volts in an early 1970s receiver. But still good after 40 years.

I think those 1970s parts were actually better when new than today's caps. They still measure better esr than today's run of the mill new caps.

Therefore now I usually go one voltage rating up. Same for aluminum and tantalum. With filter caps I go for higher capacitance too especially if original lost some capacitance and are physically much bigger.

After some research and some experience I never replace tantalum with aluminum electrolytic. If not new tants you may easily source film caps for this, up to 4.7uf, nos wima etc. Bigger tants i replace with new branded tants. I replace small aluminum with film too, space allowing.
 
Last edited:
It likely has not gotten more replies because a lot of time good intended threads like this become holy wars, without the jerusalem aspect...

that being said, increasing the capacitance of the big mains only increases the microseconds of power reserve during a large bass hit, at high power levels. if the machine spends its life <1watt in an office or bedroom - waste of time

and of course the change means it shows up as a dead short longer on start up and drains, so the diodes need looked at.

voltage values increase out of necessity - machines like marantz are famous for having power supply/filter caps only a few volts above the subjected voltage. as the caps age, this falls and the machine hums. and pay attention to this everywhere - more volts = bigger cap and in the dark ages, a 50v cap might be too large to fit in the machine, so a 35v cap gets used on a 34 volt supply line - for example. this being the future, a 63v version will likely fit smaller.

Mods have to be thought out. I know people get out the handy dandy frequency roll off calculator and change things like input cap values - it may be good reasoning or science, or a case of 'I didnt have that so I used this and it sounded good to me so it must be better and lets publish'. dunno. if I see a consensus about a mod, I would adopt it. depends on who says it. If MTF (mark the fixer) suggests a mod then accept it as he corrected a serious design flaw that could have exploded. there are others here as well on each board I accept as gospel for that machine line.

all of that now being said keep in mind that the designers and builders of these machines are likely long dead or sold to russia. there is almost no documentation other than stray manuals we scan in. so it become an exercise of either maintaining the machine, or trying to figure out what they were thinking when you desire to modify. practice makes perfect. I myself have noticed that I now see things, after 15 years or so purposefully repairing these, as an 'oh yeah, btdt' moment.

there are a myriad of reasons why these machines are special and as deserving of attention as a 69 442, but consider them the audio equivalent of a moonshot. with juvenile technology, and sound engineering principles borne out on paper they hit a home-run.
 
I know most of the stuff I get to fix was because someone tried to fix the lamps, touched something and smoke and sparks later it doesn't work. they just want it working, not restored. they put a "<$100" fix. Can't really re-cap a '70s silver face for under a $100.

my own stuff is different. i replace all coupling caps with films, the remaining, old electrolytics with similar value caps (might bump voltage ratings), and replace any remaining tantalums with films. Plus i read, read, read on these forums. a lot of great talent here.
 
As an aspiring nubie I find this thread interesting and have a dumb nubie question. What is the potential downside to recapping to original values BUT with modern caps that are significantly smaller physically? Besides the obvious physical challenges, are there risks involved say for instance in a power circuit that the cap may not be up to the original intended task? I am working on a Sansui amp and trying to stay with original specs the best I can simply because I dont feel qualified to alter specs, but doing that and keeping to original dimensions is a real challenge. I feel like I have done fairly well at locating caps as close to original specs AND dimensions as possible, and I have in some cases even sacrificed the option of higher quality caps for a closer match to the originals physically. I figure even a meh new electrolytic has to be better than a 40 year old one. Am I wrong?
 
Greetings;
You can upgrade the Voltage ratings and maintain the uf cap rating. No problem. might even fit. Depending on your model Sansui may have some modifications to better performance. You should open a new thread. Before that, However on top right of every page is the search box.. enter sansui xx-xxx and you'll get several threads. Nichicon seems very suitable for power supply, preamp and output amp. So e.g. usually a mix. e.g. the preamp audio path you would want low esr.. but for i.e. rail service caps not so picky. Power supply.. seems to be rebuilders choices for mains.
hope it helps.... but don't buy cheap chinese caps.
 
most of the time it is assumed if a 25v cap is used, the voltage is less than 25 volts. Some will bump up the cap voltage to 35 to ensure it is covered (or if the cap is running close to the cap spec).

I typically look at capacitance, voltage and lead spacing and cap series when replacing caps. Some cap series are better for power supplies and some series are better at coupling or tone stacks. dlucy typically describes the difference in his HK restore threads and encourage you to read them.

I think it is good to create a restore thread and post clearly your plans. If you are lucky (and your post is easy to read), someone might have a recommendation. it is the same work so you might as well use caps that fit the application.
 
"I figure even a meh new electrolytic has to be better than a 40 year old one. Am I wrong?"

Maybe, possibly, maybe not. If your time is worth anything to you ,using the "Meh" part is not an option. After measuring scads of caps pulled during restoration work, there are certain caps that have tended to hold to value within tolerance better than others. My experience suggests that some units have whole bunches of caps out of spec (and many that are WAY out) and others tend to have caps that are mostly still within a 20% tolerance, with a few that are slightly out of tolerance.
The Elna's seem to hold the best (for 70's electrolytics) United Chemicon seem a 50/50 and the violet e-caps in Marantz's and little blue Sanyo among the not so close to spec/must go column.

Good quality new e-caps aren't expensive, imo.
 
As an aspiring nubie I find this thread interesting and have a dumb nubie question. What is the potential downside to recapping to original values BUT with modern caps that are significantly smaller physically? Besides the obvious physical challenges, are there risks involved say for instance in a power circuit that the cap may not be up to the original intended task? I am working on a Sansui amp and trying to stay with original specs the best I can simply because I dont feel qualified to alter specs, but doing that and keeping to original dimensions is a real challenge. I feel like I have done fairly well at locating caps as close to original specs AND dimensions as possible, and I have in some cases even sacrificed the option of higher quality caps for a closer match to the originals physically. I figure even a meh new electrolytic has to be better than a 40 year old one. Am I wrong?

New electrolytics are, in my experience, significantly smaller...making for easier placement. Not sure why you are seeing physical challenges in fitting them. Assuming correct specs for the application, there is no downside of which I am aware as regards using modern electrolytics which are physically smaller than the electrolytics I am replacing.

For small value caps in the signal path, I use film caps. In this case, they are often larger than the electrolytic they are replacing...but I've found that you you can usually use films for values up to 3.3uf. As for replacing tantalums, I use films there as well...if they fit. I see absolutely no benefit to replacing such caps with new tantalums.

Bottom line, and again assuming correct specs (capacitance, ESR, leakage, etc) I don't see any reason to worry about physical size consistency. The one exception might be filter caps, which are usually mounted in brackets. Ideally, new filter caps will have the same diameter so that they can be mounted without modification....but this is an issue of aesthetics and convenience...not a technical issue.
 
Last edited:
I have had personal experience of what a bad electrolytic can do in several Revox machines. When electrolytics have a rated life of 1-5000 hours, and we know that being unused for a long time severely degrades them, it's a simple conclusion that those in 30+ years machines cannot be good.

The fact that most of them function is - I think - not the result of them being OK but the result of the designs having a decent margin. In many cases function and operate within nominal specifications is completely different.

Regarding upgrading them:
Upgrade is a good word and cannot be bad. But in this case the issue refers to "uprating" them. This is more complex.

In several old designs, capacitors were of inadequate size because of cost. So a larger one should have been used and could be used, but this is a decision that needs knowledge, experience and a bit of reverse engineering.
Uprating the voltage rating is always beneficially if done with measure. Going from 10uF/25V to 50uF./250V is both expensive and can cause issues. Since modern capacitors are smaller than vintage ones, one can easily uprate the operating voltage to at least the next step without size or functional concerns.
 
New electrolytics are, in my experience, significantly smaller...making for easier placement. Not sure why you are seeing physical challenges in fitting them.

Bottom line, and again assuming correct specs (capacitance, ESR, leakage, etc) I don't see any reason to worry about physical size consistency. The one exception might be filter caps, which are usually mounted in brackets. Ideally, new filter caps will have the same diameter so that they can be mounted without modification....but this is an issue of aesthetics and convenience...not a technical issue.

Sorry, by "physical challenges" I just meant in attempting to match with original components. Ive been trying to match values and dimensions with my caps and it is indeed a challenge. No downside to smaller dimensions is good though! Thanks, that's all I was wondering about. Makes things a whole lot easier.
 
Yep...much easier to source, and much more breathing room on your circuit boards...even when bumping up the voltage rating (which I usually do, especially on the very low voltage caps).
 
"I figure even a meh new electrolytic has to be better than a 40 year old one. Am I wrong?"

Maybe, possibly, maybe not. If your time is worth anything to you ,using the "Meh" part is not an option. After measuring scads of caps pulled during restoration work, there are certain caps that have tended to hold to value within tolerance better than others. My experience suggests that some units have whole bunches of caps out of spec (and many that are WAY out) and others tend to have caps that are mostly still within a 20% tolerance, with a few that are slightly out of tolerance.
The Elna's seem to hold the best (for 70's electrolytics) United Chemicon seem a 50/50 and the violet e-caps in Marantz's and little blue Sanyo among the not so close to spec/must go column.

Good quality new e-caps aren't expensive, imo.
Point taken. As mentioned, I was trying to keep them all the same dimensions. Now Im not so concerned about that though. One interesting thing I noticed in doing so though was that sometimes a cap described as "general purpose" is often given a far longer lifespan rating than a low esr cap. I found that rather odd.
 
Back
Top Bottom