CurtPalme.com Home Theater Forum

[mc86]

Guys -

So I read-up a tiny bit on electrolytic caps and testing thereof. I bought the Atlas ESR+ (model ESR70) as per Jask's recommendation.

Please understand I intend on doing no mods or upgrades -- I only want to ID and replace any bad caps on my 2nd ECP 4500 with identically spec'd caps before I start tinkering with my own blend. While I believe both of my PJs are very-low hour machines and I know it is heat, not time that does them in...it easy enough to go through them in two hours or so.

Anyway, my hope was to the table given in the tester manual, which is basically is a lookup for Capacitance Value, the designed voltage rating, and a corresponding typical or "good" ESR value one might expect. IF less than 2x that ESR, probably OK.

After testing my first dozen or so caps, I have three Qs:

1) Most of the tested caps have ESRs WAY under what the table suggests is typical. Should I be using some more appropriate table OR just be thrilled none of the first dozen caps I've looked at so far are above the typical "good values"? Mac didn't go into ESR value high-low on the awesome cap thread, unfortunately.

2) What if a cap has a "good" (low) ESR AND ALSO reads a capacitance higher than rated? I have a 100uF cap reading 178uF...strange. Seems off-spec and worthy of replacement. Is this an oddball failure mode or 3 sigma quality outlier?

3) Presuming I find a few bad ones (like ones near hot parts of the LVPS, transistors, etc), any recommendation for a brand/type to replace them with / specific supplier?

Any input appreciated!

Thanks!
Matt

[Nashou66]

Have these PJ been shut off for a long time?

If so then Id plug them in and run them a bit before testing the caps.

I know that tested caps sitting on a shelf for a few years read differently after I use them for a few hours then re test them.

Athanasios

[CasetheCorvetteman]

That is incorrect about time not being an issue, a lot of Nichicon and Rubycon capacitors fail over time even with little or no use.

Testing a capacitor in a circuit can give totally incorrect readings, such as you see with the one measuring almost double it's rating. Removal and restesting will probably show a different story. Same goes for the ESR reading, if there was a number of them in parallel the ESR reading will be inaccurate.

[macgyver655]

[stefuel]

ESR......What's that? Employer sucks rump

[CasetheCorvetteman]

[macgyver655]

[macgyver655]

[the big E]

[mc86]

Thanks Mac! If I've got it straight, the attached flowchart PDF and text at bottom should be helpful to folks like me. Three related Qs, tho', if I may.

Q1: Is it rare for a resistor to be in parallel with a cap, thereby making the cap's reported in-circuit ESR value read lower than its true ESR value? (I hope I've got kirchkoff's law right on this.) Your implicit acceptance that surrounding circuits will raise measured ESR and not lower the measured ESR makes me wonder if such a layout is just not done/makes no sense.

Q2: Out of curiosity, I'm guessing the lowest ESR available is not always best...there must be tradeoffs with other characteristics. Can any guideline apply when selecting a replacement cap (once capacitance, voltage, and temp are checked) that balances a "low" ESR with the other performance specifications (ripple current, leak voltage, etc)? Seems unlikely, but hard to know sometimes when I am wrong.

Q3: Suggested brand(s) that are of quality?

The procedure should be as follows:

1) Check ESR while in-circuit. Don't trust in-circuit capacitance value reading.
+ If ESR is sufficiently low (lower than the supplied chart, interpreting esr value to be max allowable), cap is probably OK.
+ If ESR is too high, remove cap and proceed to step 2.
2) Re-check ESR value out-of-circuit to confirm high reading (ESR in-circuit value may have been erroneous)
+ If ESR is above nominal chart value, replace cap.
+ If ESR is below nominal chart value, it is generally useable. Proceed to step 3. A lower ESR value may/may not have been originally spec'd.
3) Measure capacitance out-of-circuit
+ If capacitance is within tolerance range, cap is OK.
+ If capacitance is outside of tolerance range, replace cap.

Thanks all,
matt

[macgyver655]

Alright, I guess your going to make me explain this in more detail. You can probably copy and paste this into that cap thread that wasn't continued, LOL.

First to your Q1. It's is by no means rare to have a parallel resistor with a cap, however this generally will not affect ESR. ESR means Equivalent series resistance. Some call it "effective series resistance" which would possibly be more accurate. Anyways, if you break down these words you have "Equivalent" or equal to, then "series" which means wired as in series, and resistance which means resistance. Now a cap cannot be tested for resistance with a regular ohm meter because there is no actual internal through connection. Yet the fluid will allow flow. So with your ESR meter your not testing true resistance. Your ESR meter actually sends a frequency through the cap and monitors it as it comes out the other side. It then calculates the variation in the signal and determines how much effort was introduced in the process. This induced effort is your ESR resistance reading. This induced resistance would be like: If you have a perfect cap with zero induced resistance and then wired in series a 5 ohm resistor and then tested across both the cap and resistor at the same time with your ESR meter you would get an ESR of 5 ohm. So a cap, by itself that has an ESR of 5 ohm is "like or equal to or equivalent to" a good cap with a 5 ohm "resistor" wired in "series". GOT IT !!!! LOL.

So what causes a higher then acceptable ESR. Well you have the contact points in the cap where the legs attach to the plates and you have the fluid which allows good flow. If a contact point becomes weak it can create some resistance and/or a "LOSS" of fluid can cause a lesser ability for the signal to transfer. So most of the time, when your getting a higher then acceptable ESR it's probably that the cap has lost some fluid either by leakage or evaporation. Plate issues are also a possibility.

Now As to your resistor in parallel effecting ESR, not likely since your testing with a frequency through the cap. It's not the same as testing with an ohm meter where a parallel resistor will affect a reading. I suggested rechecking after removal because there are external things that could affect your reading. If you test at the caps solder points and there is a poor solder joint it can give you a false ESR reading, where your reading is actually showing the poor solder connection resistance and not the caps actual ESR. Also, your test leeds if they don't have good contact with the caps legs could show a poor ESR reading and after removal you would usually end up testing on a clean area of the leg. The surrounding circuit won't "generally" affect ESR reading but there can always be a exception to the rule.

As to Q2: although low ESR is strived for, there are some circuits where this could have a negative affect. But that's a whole other conversation. Some are only happy using a low ESR cap but I believe that's more of an opinion thing. I never insist on a super low ESR cap but mostly because I don't think it's needed at the extra expense.

As to Q3: Many people have their cap brands of choice but 1 thing I always insist on is to buy them from a reputable supplier, 1 that you would hope are very particular as to who they get their stock from so as to avoid getting knock offs. I would never buy caps off Ebay. And don't forget to test that new cap before install, lol.

Nice chart by the way.


Are you guys happy now?

[gjaky]

One thing I want to add: the need of the formation of the capacitor. Either the cap is old but wasn't used for a long time or it is new, newer used, the capacitor should be formatted before any measurement. This is important because the electrolyte builds up an oxidation layer on the aluminium foil if voltage applied on it, this oxidation layer is crucial for proper working, so an unformatted capacitor may seem to have low capacitance and high ESR value, though after formation it is ok.
The formation consist of a few charge/discharge cycles on about the 1/10 of the maximal ripple current rating of the actual capacitor.

[CasetheCorvetteman]

Macgyver, what effect on the reading will multiple capacitors in parallel have?

[gjaky]

[CasetheCorvetteman]

[stefuel]

[CasetheCorvetteman]

Ive done it all before

[mc86]

Thanks for all the great replies. A few more Qs while I'm gearing up for this:

1) Nashou also mentioned the Gjaky's "formation" process (he suggested powering the unit if it'd been off a long time). Anyway, is there a "good" (ie, easy-enough) way to do this? That is, I have a gigantic variable DC power supply that can deliver up to 120VDC at several amps (for a huge magnet). Can I apply +/- leads to the cap and just slowly bring the cap up to the rated voltage, then turn voltage down, and then discharge the cap? The discharge would be via resistor to ground or with the ESR meter if the cap is "small".

2) Can you folks expand on the "pulling out the caps" as recommended in this thread? I gather the technique should be like in this youtube video, right?

3) Do you guys recommend a desoldering iron (presumably with built-in manual vacuum) OR should I just not worry about sucking up any melting solder/use wick/etc? I am a bit persnickety about things being done nicely if a little extra effort is involved. I have access to a hakko fx-888 solder station if I need it, but I planned to use a butane powered solder pencil at work. I can get really fine tips for that.

Thanks again,
Matt

[Nashou66]

For desoldering I tried everything and the best is to just use soldering wick. I dip it in flux then slop it on the solder I want to remove. then just clean the board with de natured alcohol when done.

Drags has a brand of wick he likes but for convenience I just use radio shack stuff.

Athanasios

[stefuel]