How do you know it is time to recap?

  • Thread starter Thread starter sweetbeats
  • Start date Start date
altitude909 said:
Electrolytic caps failing in a power supply (i.e. the really big ones) will manifest itself in a hum. PSU caps ("smoothing" Caps, they filter the ripple out of the DC after being converted from AC) are always the ones to start with since they deal with higher currents and will be the first to go.
Click to expand...

i'm afraid you have a misunderstanding here... smoothing caps is just another name for filter caps... they all smooth the ripple out of the primary supply...

perhaps you meant the decoupling caps that are used around the opamps??? they are there to eliminate parasitic signal from the opamps getting into the supply lines...

and neither of these types of caps have anything to do with current capability... especially since they are dealing somewhat exclusively with DC...
the only caps that deal with AC and subsequently with current "to some degree" are those that are actually in the signal path... like those interstage types in gutar amps... and really not much of an issue...


you are right though in suggesting that the power supply is the place to start... and the only caps that have to be replaced are the electrolytics... the small value caps are generally like tantalum/polyester/polypropylene and dont really go bad that often but can drift a bit...
 
No misunderstanding.

The larger the current draw after the rectifier, the larger the capacitance needs to be to bring down the ripple. Big caps dry out first, result is more ripple on the DC supply line, simple as that. I didn't say anything about current stability

Opamps don't get decoupling caps btw, they have bipolar supplies and are never connected to ground
 
Last edited:
ah... i misread part of your post i thought you were trying to make some kind of distintion between filter and smoothing caps... my bad... though i still dont see the current draw as an issue... they need to be sufficient to pass enough current along to the device sure... but that shouldn't affect the amount of ripple... in fact an underspec cap will exhibit less ripple in some cases as the supply at that point is constantly in a sag state... ymmv...

as to opamp decoupling look at any good schemo and youll find them all over the place... most of the time they are very small values that sit on the supply rails near the op-amp and either go from supply to ground or between the two supplies... take them out and things get noisey quickly...
 
...

electronically speaking, a lot of solid state packages (chips...lol) require tiny capacitors on their ground, or a supply pin, or whatnot...

in a simple block diagram, they are omitted. They overcomplicate the inherently straightforward function you are studying... BUT, they are practical circuit details, either needed for operation, or as in this case, to keep the noise down to a dull roar.

in addition to recapping or whatever else maintenance you are needing/wanting to do, someone mentioned new OP AMPS up there... now THERE'S a place you can spend a few bucks, measure twice and cut once... and really get some bang for your buck, performance wise...

this whole thread SQUARELY reminds me of a ream of similar ones on this one electronics site, whrere they are all EE guys in the audio field...

they have schematics on home brew tube pre-amps, etc etc... if i can find the old site name, I will post it here fro you guys... ANYone liking this forum will go apesh!t about that site, trust me. You into audio? you have a soldering iron? you can read a basic schematic??? *you will get a woody*, LMAO
 
Okay...I'm getting into this...

I'm recapping the PS for my Tascam 58.

I've looked at the 4 retailers listed above and I'm finding that I can't get exact replacements, particularly with case size. I was advised to go bigger on the case size if possible, but for one cap in particular (6800µF, 40mm x 40mm case, 50V) all I can find are smaller cases. How critical is that? Also, the leads on the cap mentioned above are like riveted on terminals rather than tails...can those be rotated? The Tascam PS PCB has the mounting points in an 'L' configuration (i.e. the cap terminals are perpendicular to each other), and I'm concerned about finding a replacement cap with the right spacing/configuration.
 
sweetbeats said:
I'm recapping the PS for my Tascam 58.

I've looked at the 4 retailers listed above and I'm finding that I can't get exact replacements, particularly with case size. I was advised to go bigger on the case size if possible, but for one cap in particular (6800µF, 40mm x 40mm case, 50V) all I can find are smaller cases. How critical is that? Also, the leads on the cap mentioned above are like riveted on terminals rather than tails...can those be rotated? The Tascam PS PCB has the mounting points in an 'L' configuration (i.e. the cap terminals are perpendicular to each other), and I'm concerned about finding a replacement cap with the right spacing/configuration.
Click to expand...

Case size does not matter in general other than if it is too big to fit it is too big. There is a general trend to larger being less dense being more reliable (has to do with internal head and field density) but that is only for a like generation (apples to apples)

If the cap can is smaller for the same value, voltage and temp ratings then most likely you are looking at a newer generation of caps vs an older. With the more modern materials performing better. This assumes apples to apples again.

As that we are talking about power supply caps the exact values are not as important as in (say) a timing circuit or an EQ filter.

In recapping a PS I would select a cap that has the same value or the next cap up. (100 uf in place of 50uF for example), has a 105 degree temp rating and fits the position. You can bend the leads around to fit but keeping leads short is good practice.

As for op-amp bypass caps...As was noted a bypass cap from rail to rail or from each rail to ground is common and good practice. typical values are 10 nF. They should be mounted as close to the device as possible. Their function is to keep the part from oscillating. Not so important with the typical 4558 used in the Tascam 58 channel cards but very important id you upgrade to (oh say) an LT1358 or some other high speed op-amp.

Regards, Ethan
 
Thanks, Ethan. That makes sense and will help tremendously in locating replacements. ;)

As for op-amp bypass caps...As was noted a bypass cap from rail to rail or from each rail to ground is common and good practice. typical values are 10 nF. They should be mounted as close to the device as possible. Their function is to keep the part from oscillating. Not so important with the typical 4558 used in the Tascam 58 channel cards but very important id you upgrade to (oh say) an LT1358 or some other high speed op-amp.
Click to expand...

Huh....what I am about to ask has likely been addressed:

  1. elsewhere
  2. here
  3. all of the above

I don't understand how a small cap between the +VDC rail and ground doesn't just short to ground....
 
Caps block DC

When you first apply a voltage across a cap it appears like a short. That is to say that is passes current as if it was a resistor (which limits the number of electrons per second, er, current flow). The cap charges. As the number of electrons in the cap rises the voltage of the cap rises. THis decreases the current (current only flows when there is a voltage difference) until the cap is at the same voltage as the source. Then no current flows and the cap is charged and it appears as if there was not connection. Thus the cap blocks DC.

This is what prevents bypass caps from shorting the PS rails to ground. They charge and then stop passing current.

Is a PS the filter caps do exactly this. As the pulsating DC from the diodes rises to its maximum value the filter caps voltage increases. The source pulsating DC heads to 0 and the cap is able to supply the current needed by the circuits the PS supplies. THe circuits use less than is stored inthe filter caps and the source voltage ramps up again to refill the cap. They smooth out the flow of current....

PS caps are quite large in value, bybass caps are much smaller.

The PS cap works because it is never discharged fully. It takes time to fill and time to discharge. Larger caps take more time and thus work better in PS.

But what about those small value bypass caps from the PS rails to ground? They quickly fill up and then do not conduct DC again. Get a change in the supply voltage and the cap starts to conduct again. Increase the source voltage and the cap voltage goes up. Decrease the supply voltage and the caps voltage decreases.

One way that high speed opamps oscillate is through interactions in the PS rails. Minor changes in the signal change the PS voltage which changes the opamp which changes the PS voltage which....and so on till the device is oscillating at several MHz or tens of MHz (or 100).

The bypass caps counter this by making the PS rails appear shorted at "high" frequencies.

Man you are making me dig up things I've not studied in ages. And our understanding is better now than then.

Regards, Ethan

http://en.wikipedia.org/wiki/RC_circuit
http://en.wikipedia.org/wiki/Capacitor
 
Oh, Ethan, thanks so much...I am literally laughing for joy at the new understanding...kinda wierd maybe, but it explains so much. Thanks again.

Caps work the opposite of how I thought they worked...I was thinking it was like an uninterruptable power supply, like a tiny battery backup (cap charges and then spills over...supply sags and stored energy in cap compensates...) Very cool. You can see now how confusing that would be to try and understand their place/purpose in the schematics with which I've been working.

I realize I could have gone Googling to get that info, and I really appreciate you acquiescing to my laziness :p, but what you have done is given me the big picture (along with some very valuable "real world" application information) that will help me as I read the links you included as well as other information I come across; the big picture that would likely have been more difficult for me to get from Googling. It will help me tremendously as I take your post and apply it to my schematical studies. My brain and comprehension skills work better this way (big picture first, apply big picture to real-world situations and get lots of little pictures).

So when a cap goes bad or gets cooked, current either flows when it is not supposed to (becuase the cap can no longer limit current flow...a short if it is a smoothing cap between the PS rail and ground), or it won't pass current at all so it no longer filters or smooths the current. Is that right? I'm speaking of an extreme "goes bad" but there are all points in between where the cap loses capacity to store electrons per spec and it is less effective at smoothing, righrightright??

So, okay...anybody...and maybe this is opening a can of worms/subject of long-term debate, but what is it about a faster opamp that makes it better for audio applications? Transient response?
 
So how does this look?

Ethan, I wasn't sure what you meant specifically when you said

In recapping a PS I would select a cap that has the same value or the next cap up
Click to expand...

You gave the example of of 50µF to 100µF, but of course there are lots of "steps" in between that...so what is the equivalent "step up" for a 6800µF cap? :confused:

Any brands to avoid?

I'm looking at something like this to replace the exisiting 6800µF 20% 85°C 40mm x 40mm cap.
 
sweetbeats said:
Caps work the opposite of how I thought they worked...I was thinking it was like an uninterruptable power supply, like a tiny battery backup (cap charges and then spills over...supply sags and stored energy in cap compensates...)


So when a cap goes bad or gets cooked, current either flows when it is not supposed to (becuase the cap can no longer limit current flow...a short if it is a smoothing cap between the PS rail and ground), or it won't pass current at all so it no longer filters or smooths the current.

but what is it about a faster opamp that makes it better for audio applications? Transient response?
Click to expand...


the battery analogy still holds true ... except what it does is releases it's charge as the transformer goes the other direction... thus holding the over all charge up... make sense???

when caps go they short... sometimes they act kinda leaky but that's really just on the way to short...


transient response... bingo... also newer opamps can be capable of operation from rail to rail... in other words assuming a +/-15v supply you can push it to within a hair of the 15 volts... older op-amps were lucky in some cases of doing 10 volts out of 15... and that's assuming they could stand a 15V supply at all...
 
Huh!

the battery analogy still holds true ... except what it does is releases it's charge as the transformer goes the other direction... thus holding the over all charge up... make sense???
Click to expand...

Yes, I think it does! I re-read Ethan's post in concert with yours...current exists..only when..there's a voltage differential...so if a charged cap sees a voltage sag on the supply side current will flow essentially "upstream" out of the cap, compensating for the sag like you said...so then if the voltage supply increases beyond the voltage stored in the cap, that is just stored in the cap (up to its capacity) instead of heading on downstream and hitting components? And that explains why, for instance, C29 (a 1µF 50V cap) on the amp card in my Tascam 58 is rated at 50V and resides on a much lower 24V circuit. It'll take whatever the supply will dish out (unless something really horrid happens in the main power supply that propogates all the way to that cap), but it is only 1µF because that is all that the engineers calculated would be demanded of that component to elicit the desired smoothing result...am I getting it?

See, my old "battery analogy" lacked the very important piece that the cap stops passing current to ground when the differential reaches 0, and I wasn't thinking of current flowing out of the cap back into the circuit...I was only considering the current flowing from + to ground...

transient response... bingo... also newer opamps can be capable of operation from rail to rail... in other words assuming a +/-15v supply you can push it to within a hair of the 15 volts... older op-amps were lucky in some cases of doing 10 volts out of 15... and that's assuming they could stand a 15V supply at all...
Click to expand...

So transient response and dynamic range as well right? That's really exciting...so I'm reading into this that there is some real potential for improvements to my M-520's performance with some opamp upgrades, yes (assuming the rest of the circuitry is up to the task...)? Not that what's in the M-520 stinks, but what was industry standard and/or leading edge 20 years ago is different now...
 
there's something in your understanding of the caps that's still not quite there and i' need to think it throughsome to clarifiy... though i susect it's in the different uses of caps... the filter supply caps (some call smoothing) are different than the decoupling(bypass)and the coupling(interstage) then there's what i(was taught)call bypass... i'll see if i can work up something to differentiate them for you...

as to the opamps... there's some other things that newer can do as well... some use far less current than predecessors so efficiancy is improved... and some will pass current through better which makes them better for driving output stages for instance... that's why you need to know what that particular device is doing... make sense???
 
there's something in your understanding of the caps that's still not quite there
Click to expand...

For some reason I'm not surprised...:D

Thank you in advance for taking the time. I do realize that a cap in a PS application the current does flow unidirectionally since there is always a negative voltage differential downstream of the cap...but any help you can offer would be much appreciated.

as to the opamps... there's some other things that newer can do as well... some use far less current than predecessors so efficiancy is improved... and some will pass current through better which makes them better for driving output stages for instance... that's why you need to know what that particular device is doing... make sense???
Click to expand...

Yes, absolutely. Thanks! Not too far down the road I'll be working on mods for my M-520...I started a thread awhile back with no responses yet, but hopefully that will change as I get more specific and actively involved in the project.
 
first... dont get hung up on the idea of caps passing current... for the most part they dont! they block voltage and pass ac signals which is really more of an inducement of that signal into the next stage...

as i see it there are basicly 4 types of uses for caps...

1. filter supply caps.... what we have here is a large value cap that is hooked between the supply and ground... so the line in just passes across the "hot" side... it doesn't realy do anything relative to the ground... the groung is just a reference... if you take the cap out of the circuit what you would see is a hump looking signal for the positive line and a cup for the neg... so you put the cap back in and the hump becomes smaller humps resting on a dc charge...as you increase the capacitence the humps get smaller and the dc goes up till you reach the effective limit... so then you dont see any of the hump at all... if we assume then that the device is drawing on the dc if the capacitence is marginal the you could see the signal reflected in the dc line... because as the draw increases the supply can sag... thats part of what gives tube amps their sound... increase the cap some until the sag goes away....

2. coupling caps... these aren't used in solid state stuff much nowadays... but very much a consideration in tube gear... they are mostly (lets call them) mid value types that block the dc of one stage from entering the next while passing the signal which is effectively an AC signal... look at any guitar amp schemo and you'll see lotsa them...

3 decoupling caps... these are small value caps that sit either from supply to ground or between +/- supplies typicly right at the op-amp... as has been mentioned earlier they can keep op-amps from oscillating... they also can reduce whats known as parasitic feedback/forward... remeber how i said an under capped supply can reflect a signal... well this is the same on a microcosm level... the op-amps draw is reflected here and the cap passes it to ground... imagine if every opamp in this board were to put miniscule amounts of it's own signal on the line... and add them up... this looks like high freq noise on the lines if not dealt with...

4. bypass caps... these are used in the feedback loops around op-amps... they can affect the freq response by either passing the higher freq along... or increasing the amount of them fed back for gain....


so when you look at a schemo you can assume two things... if it's between someline and ground it probably there to reduce noise (the exception being some tone controll circuits)... if not from line to ground then its there to block dcand/or to affect freq response...


remember there will be a quiz on thursday....
 
Don't take the water anology too far. There is no spilling over after full in caps.

In DC the current starts at one place and follows the circuit back to the source in one direction only. Voltage is anologous to the pressure in a host. Current is the actual flow of electrons (or water....) You can have high pressure with low flow or low pressure with high flow and so on.

It is a circle with electrons doing the moving.

now break that circle and put a cap in the break. Turn on the DC and electrons will rush through the conductors. On the negative side the electrons will pile up in the negative plate of the cap (remember that the cap is really 2 plates that do not touch with an insulator between them) and on the positive side a depletion of electrons will exist. There is a current or flow of electrons up to the point that the potential in the cap (its voltage) is equal to the supply voltage (minus some resistances but why cloud the idea).

Think of the cap as a bladder or balloon that is hooked in a tee to your hose. Turn on the water and it starts to fill, but water still comes out the end of the hose. Turn of the water and the bladder empties causing its water to come out eh end of the hose. Turn the water on and off and the bladder will smooth out the flow out the end of the hose.

thats the DC filter cap example.

For a coupling cap think of what AC is. It is a rush to one voltage followed to a rush to the oposit voltage. If we take that loop of wire and apply an AC voltage across it the electrons will first run clockwise then counter clockwise.

Now put a break in the circuit and install a cap. Electrons rush to one plate then to the other. The cap never really fills up because the voltage is always changing. the voltage in the cap lags behind the source voltage.

To use the water example. We cut the hose and put 2 bladders in. One on each end of the hose. The bladders are touuching such that if one is full the other is empty. Run water clockwise and the left baldder fills and the right empties. Reverse the direction and the left empties and the right fills. We are passing current now without end. Caps pass AC.

This shouuld add to your confusion :eek:

Oh back to smoothing. Caps turn big (max to zero) ripples into little (a few mV) ripples. Regulators get rid of that for the most part.
 
sweetbeats said:
Hi folks.

I apologize in advance due to the fact that, at my skill level, I am much more likely to be a taker here than a giver. :o

I started a thread about recapping over in the Analog Only forum thinking it might be more appropriate over here, and another member reinforced that idea, so here I am. ;)

The original thread is here.

As you'll get from the original thread, I am comfortable with a soldering iron, have a pile of mid-80's Tascam gear that I intend on treating right for many, many years. I want to keep it going and going well. I have heard about the degredation in performance that bad caps can cause, not to mention the potential damage. I especially want to avoid the latter. So in the spirit of good stewardship to my gear and my craft I'm trying to evaluate if I get into recapping now, and where to start.

Thanks in advance for the time you expend in this newbie! :p
Click to expand...

Electrolytic caps (of the old) have a lifespan of about 15yrs. Anything beyond this age needs to be re-capped. The issues is that the caps start to dry out and despite the ones who will try to tell me different, caps dry out and short eventually. In the meantime, they become noisy and do funky things to the signals that usually go unnoticed until it becomes bad.

Just because your 20 year old whatever seems to work fine, any electro this old is on shakey ground.
 
MCI2424,

Thanks for the focused post.

I've been wondering overall about this as my 20 year-old Tascam M-520 console seems noisier than I'd expect.

Now that I've started getting my hands dirty and have pulled the caps out of the power supply on my Tascam 58 I can see that recapping can be a fun and satisfying project. I'm looking forward to doing the M-520, and hoping that it provides for a longer service life and helps with the noise issue.

Some questions that remain at-large that are hindering my next steps (i.e. selecting replacement caps):

  1. evm1024 stated "In recapping a PS I would select a cap that has the same value or the next cap up" to which I had the question "[evm1024] gave the example of of 50µF to 100µF, but of course there are lots of 'steps' in between that...so what is the equivalent "step up" for a 6800µF cap?" Anybody?
  2. Any brands to avoid or can I be reasonably sure that anything I get from Digikey/Mouser/Allied/Newark will be a reliable product?
  3. Does the "you get what you pay for" adage apply? For example if one cap by one manufacturer is $2.00 and the exact same spec cap from another is $3.00, do I assume that it would be worth spending a little extra?

Thanks to everybody who has put up links and type-written info regarding theory, function and application of capacitors. It is very helpful.
 
sweetbeats said:
I noticed that there are many caps that have the same capacitance and voltage rating, but there are additional bits that are different. Can somebody explain those differences to me?

Like look at these 3 caps...All the same capacitance and voltage rating, but the percentages and that pesky "(B. P.)"...what is that?

10uF 16V
10uF 16V (B.P.)
10uF 16V (B.P.) 20%
Click to expand...


B.P = Bipolar
B.P 20% = Bipolar with a 20% value (+/- 20% of the value)
 
sweetbeats said:
MCI2424,

Thanks for the focused post.

I've been wondering overall about this as my 20 year-old Tascam M-520 console seems noisier than I'd expect.

Now that I've started getting my hands dirty and have pulled the caps out of the power supply on my Tascam 58 I can see that recapping can be a fun and satisfying project. I'm looking forward to doing the M-520, and hoping that it provides for a longer service life and helps with the noise issue.

Some questions that remain at-large that are hindering my next steps (i.e. selecting replacement caps):

  1. evm1024 stated "In recapping a PS I would select a cap that has the same value or the next cap up" to which I had the question "[evm1024] gave the example of of 50µF to 100µF, but of course there are lots of 'steps' in between that...so what is the equivalent "step up" for a 6800µF cap?" Anybody?
  2. Any brands to avoid or can I be reasonably sure that anything I get from Digikey/Mouser/Allied/Newark will be a reliable product?
  3. Does the "you get what you pay for" adage apply? For example if one cap by one manufacturer is $2.00 and the exact same spec cap from another is $3.00, do I assume that it would be worth spending a little extra?

Thanks to everybody who has put up links and type-written info regarding theory, function and application of capacitors. It is very helpful.
Click to expand...

Stick with the rated cap value if you are not sure of it's function. Some caps are for timed circuits and if you change the value, the timing goes to hell (Fostex uses this in it's transport control). If you know for sure that the function of the cap is just for filtering (power supply caps) then you can safely up the value and voltage rating. The Panasonic 105 degree line of caps are very good without going overboard on the price. You can order them from Digikey and they have an online catalog.

First, take an inventory of all caps needed and make a list. this makes ordering quantities much easier and you will not have to re-order forgotten caps (costing more because of the low quantities).
Second, you should get a good soldering station (Weller for instance) with a new tip. The solder sucker should be ESD safe. The most important thing is to practice removing caps (and not etches/solder pads!) on a throwaway PC board. Use any through hole PC board (from the dump?) that you can find. Through hole boards are the toughest to remove components and if you can remove them without damaging the pad/etches, you can remove surface mount anything (except BGAs).

Good Luck and take it slow, you should be alright (buy a static strap and use it!)
 
You must log in or register to reply here.
Back
Top