Should 35 year old electrolytic capacitors that test OK be replaced?

[JohnCPR]

I am rebuilding an AKAI GX-365 (vintage 1970) reel-to-reel tape recorder because performance has deteriorated. I am testing the electrolytic capacitors to determine if they should be replaced. So far, those I have tested using a Heathkit Capacitor Checker Model IT-28 show the following:
- higher ratings than marked (up to 50%)
- a pass for discharge and leakage
- a power factor of about 5%
The capacitors I would use to replace them are much smaller in size from the Digi-Key catalog of Panasonic radial and axial aluminum series capacitors.
Should I replace them because of age or their lifespan affects their continued reliability?

 

[itsmoked]

I probably would not change them if the tester is not showing any pathology.

What is your basis for "performance deteriorated"?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[JohnCPR]

Thanks Keith for your feedback.

The performance deterioration comes from fluctuating volume from the speakers, sometimes no sound at all, and background hum. It is very likely that the hum comes from deteriorating power supply electrolytics. The fluctuating volume may be due to dirty pots, or perhaps deteriorating electrolytics in the audio circuitry and solder joints.

My experience has been with surface mount electrolytics in analog video cameras. I have found that these have lasted up to 10 years when they start to break down and leak electrolyte. My success with surface mount electrolytics has been in replacing them all rather than waiting for them to break down one by one.

I do not have much experience with the larger through hole electrolytics of the 1970s. I assumed that my experience with the surface mount electrolytics may be the approach I should take with the electrolytics in the AKAI tape recorder. However, the through hole capacitors today are smaller and may be less reliable. I am presently checking the circuit boards one by one by removing and testing the electrolytics and checking solder joints, not knowing if the best approach is to replace the electrolytics because of their age. This is the reason for my question.

 

[BrianG]

Traditional electrolytics tend to dry out with age. Whilst the +50% capacitance measured is quite within normal production tolerances for this type of cap, I suspect your capacitance checker does not measure ESR. This is a key factor in getting a good power supply with low ripple, and unfortunately ESR tends to worsen as electrolyte dries out, so I would change those in the power supply at least.

 

[itsmoked]

Since you say hum I too would change them. I would use the same general style. I wouldn't switch to SMD if that is what you're contemplating.

This isn't the REEM tape recorder that was made by AKAI is it?

You need not worry about the life of the replacement caps as cap life is related to temperature and your recorder will not be pushing those limits at all.

What I see the most is solder joints going bad.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[OperaHouse]

I'm very familiar with those old magic eye capacitive meters and the power factor test is about useless.

Power factor is in % at 60 cycles. The
equivalent series resistance would be ESR = Decimal per cent divided by (2 Pie x Frequency x
Capacitance x .000001) A 2 uF capacitor at 6o Hz with a Power Factor of 5% is

ESR = .05 divided by ( 2 x 3.14 x 2 x 60 x .000001) = 66.3 ohms.

ESR is a better test than capacitance for most electrolytics. If the capacitor is in the power supply it should be replaced. Measurable power factor was ok back in the old tube days when currents were lower.

 

[ScottyUK]

Buy them from a brand you recognise, or at least one one of us recognises. There are some truly awful cheap capacitors out there. I don't do much with little caps, but power applications tend to bring out the worst in any capacitor. Some cheap Chinese ones didn't meet the published spec: can ruptured on a couple (the 'safety relief' must have been purely decorative) and another caught fire. I went back to the expensive ones from Evox Rifa and things behaved properly after that.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[JohnCPR]

Hi fellows,

Thank you for your replies.

From my experience and your feedback, the consensus is to replace the power supply electrolytics. I will not be using SMD. Equivalent size is a problem since electros these days are smaller. I have rebuilt the large can lug type caps by removing the top, scooping out the inside and replacing this with an equivalent value of Japanese origin and smaller in size. This technique has worked with older power supplies for me. These capacitors exist in the main power supply. I presume the power filter capacitors on the remote circuit boards should be replaced as well.

Now for the dilemma part. What capacitor is reliable? There are the ones made in China. That includes those exclusively of Chinese origin, those marked by a company from another country but made in China, and those marked as if from a legitimate company but counterfeit and possibly made in any country. I have had my fill of counterfeit like an MJ802 power transistor marked from Mexico with a Motorola insignia containing two small transistor chips mounted in parallel and covered with silicone. I now stick to reliable suppliers like ON semiconductors. I have used Digi-Key in Canada as a supplier in the past with no problems so far for electros. The Digi-Key catalog shows the origin is Panasonic.

With power supply electros and possibly filter caps resolved, what of the coupling electros and the problem of drying out? The AKAI GX-365 tape recorder is large and very heavy with loads of internal space so excessive heat should not be a problem. What is our consensus from experience for replacement of 35 year old coupling electro caps while I am at it?

 

[itsmoked]

Any E. cap from Digikey is fine for your use. Do not worry about 'where built' if you get them from Digikey.

I wouldn't wholesale switch all the caps. I've seen electrolytics work fine for decades.

How about changing these filter caps and supply caps and then re-testing?



Keith Cress
kcress -

http://www.flaminsystems.com

 

[ScottyUK]

Watch the ripple current rating of electrolytics - the big can type almost always have a higher ripple rating than physically smaller ones even though they're the same capcictance and working voltage. That said there are some good PCB mounting types designed for SMPS applications. As a recommendation for can-type electrolytics, the Rifa PEH169 series is my first choice for high power, high current applications where you need a top quality capacitor, the cheaper PEH200 series is also good. A lot of this will come down to your location - I'm in Europe, sounds like you're in the US?


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[JohnCPR]

Thank you all so much for your help to improve and confirm my experience with maintaining electrolytic capacitors. I am amazed at the reliable manufacturters' ability to reduce the size of electros and still maintain the quality of performance. I generally tend to replace all the electros once I have dismantled a device, especially if it takes considerable effort to get to the part in the hope I do not have to do it again for many years.

Thank you Scotty for your advice on power supply ripple reliability. I have used 105 degree centigrade rated electros with ripple current ratings for power supplies. I must admit that I do not fully understand the ripple current rating given for the Pansonic electros given in the Digi Key catalog, namely: 120Hz, 105 deg. 160-450 W.V. 47 - 470 mF, 0.80 at 60 Hz. What does the 0.80 mean?

 

[ScottyUK]

Hi John,

I can't find that sample data on DigiKey's site - any chance you could link to it so we can see it in context?



----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[JohnCPR]

Hi Scotty,

I went to Digikey.com and found some ripple information at the following url:

http://pdfcatalog.digikey.com/T091/SectF.pdf

The values are not exactly as shown in the hard copy catalog. At the url, I found a figure of 0.7 for example listed in one location as a correction factor multiplier or in another as a ripple curreny multiplier in mA.

 

[ScottyUK]

Still not sure if I understand the question - some days I think I should have been born with blond hair - but ripple current ratings are frequency-dependent. Generally the higher the frequency, the higher the ripple current rating. The ripple current rating will be quoted at some stated frequency, often at 100Hz or 120Hz as those are the fundamental ripple frequencies produced by a single phase bridge rectifier running on 50Hz or 60Hz mains respectively. The multipliers are used to calculate the rating at frequencies other than that listed: Digikey (or Panasonic) quote their ripple ratings at 120Hz: if you are using that capacitor at 60Hz then the ripple rating is down to 80% of the value at 120Hz. If you were using it at 1kHz then the ripple rating would be 150% of the value at 120Hz.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[JohnCPR]

Welcome to my lack of understanding and confusion. The frequency correction factor for ripple current in the Digi-Key catalog is shown for several capacitance ranges and for frequencies of 60, 120, 1K, 10K, and 100K Hz. As an example for one row in a table:
6.3-100 W.V.(V DC), 390-100 microF, 0.7@60Hz,0.75@120Hz, 0.9@1KHz, 0.95@10KHz, 1.0@100KHz.
In my case, the power supply capacitors are used in North America (60Hz), usually in a full wave rectifier circuit. Then there is the reference to ripple current multiplier in mA.
A full wave rectifier circuit at 60Hz would give me a ripple of 120Hz so I am trying to understand what the 0.75@120Hz means in terms of the ripple current multiplier in mA (as referred to in one of the tables). I would have expected something that refers to current as the higher the current load, the higher the ripple.