Capacitor advice please

[Reference1042]

OS-CONs are extremely good in terms of low dissipation factor. Don't use them where they can resonate with wire inductance or cause other circuit problems because of that low dissipation factor (= high Q). Those cases are rare. They filter high frequency junk like crazy, better than almost all standard electrolytics. There are a few specialized electrolytics designed to compete with them that cost a bit less. Illinois Capacitor makes some. The big problem with OS-CONs is that they don't come in higher voltage ranges.

Heat is a big cap killer. You'd be amazed at the reduction in cap life that's caused by heat flow from a hot power resistor, to a cap lead, through a copper trace. It gets coupled right to the center of the cap, which then lives about half as long as other identical caps on the board.

Low voltage electrolytics are also prone to problems. I've seen all manner of failures in 6V and 10V caps. I always replace these with 16V minimum, 20V preferred, if it will fit.

Once you get away from the heat problems, and away from the cheapest of caps, cap life is way longer than most realize. Those published life specs are for maximum temperature and worst case conditions. If we ran our caps that way, we'd have to rebuild our equipment every year! Or even every few months. OTOH, under cool low stress conditions, most decent caps in audio equipment can last many decades. That said, there have also been some pretty lousy caps used in low cost consumer audio equipment.

My thoughts on the matter.

Apologies, but im still a bit of an amateur in electronics. And didnt understand some of that.

Which areas would i be best to keep oscons away from? Could you clarify in laymans for the noob

And yeah, i always wondered about the quoted lifespans. Ive been running my audiolab 8000s for around. 5 years now. It has got elnas in it, which are quoted with a reasonably short lifespan. They are still going strong!

 

[Conrad Hoffman]

All caps have a bit of resistance. The reality is more complicated than a simple combination of a resistor and a perfect capacitor, but you can think of it that way for the moment. Now, if you connect a capacitor and inductor together, in series or parallel, you create a tuned circuit that can ring like a bell at a certain frequency. That's how tuned circuits in radios are made.

Adding some resistance damps out the ringing. Since wiring has inductance, there is the possibility of creating a tuned circuit whenever you add a capacitance. Film caps, mica caps, some ceramic caps and OS-CONs have very little internal resistance, so you have to be careful just popping those types of cap into a circuit designed for electrolytics. Electrolytics inherently have more internal resistance. Usually it's no problem. One place where it can be is the output of a voltage regulator (they have a sort of psudo inductive character). It can cause oscillation or a noise peak. Another place is in parallel with a larger electrolytic. When in doubt, use the same cap type (not brand) that the circuit was designed with.

IMO, Elnas are pretty decent.

 

[Reference1042]

Well i hope i have got it correct. Order is done now. But this is just a learning project anyway. So if it goes Pear shaped, ill fix it, or build a new one

But thank you for the extra food for thought. Every new piece of information is very much appreciated.

I was following advice in a dac mods primer article posted earlier that spoke about oscons, and advice given to me by someone who mods dacs.

The values of all capacitors were kept the same as the original spec. Would this make the issue you mentioned less likely?

———————————————

 

[indianajo]

This thread is fine. See you back if you have interesting results.
My theory is if you buy a premium product, you get a premium product. Some people watch for low ESR in a cap, some watch for initial tested ripple current spec. I worry about life spec since I have had to re-e-cap an amp and preamp 3 time in 40 years, and the resulting long life caps I can now find haven't caused problems yet. I'm got about 10 year life at 24 hours a week on 1000 hour rated caps, which is good, but not good enough for me. I hope the 10000 hour ones go the rest of my life, my vision is not going to get any better. There are also expensive "audio" caps some vendors sell that don't have any test numbers to back up the premium price. I like to read the numbers. Interesting about the ringing on Oscom, similar to what people warned about on the thread about paralleling film and electrolytic caps where the design was e-cap alone.
I'm just dipping into the realm of low voltage caps, every thing until the last two years has been tube designs or the 100v rated ones in the ST120. I've re-e-capped a 16 year old Peavey amp where no 10v e-caps failed that weren't blown by the 95V runaway through the blown output transistors, so those small caps were okay but that is a premium priced product. Cost Cutter PCAT power supplies, something blows up every year, so that is the other end of the price spectrum. I've just bought a 1968 organ where the factory bankrupcy receiver warns everybody that all the small e-caps need to go even before the big ones in the power supply, so I've got my work cut out on that one. I doesn't even whimper right now. Plenty of room for 10000 hour 25 v caps, though in that beast.
Good luck.

 

[Reference1042]

Got the DacMagic 1 modifications finished today! Was up till 5am in a haze of solder fumes..... Cough Cough

Was a bit of a faff squeezing the larger oscons into some areas. And a challenge desoldering the opamps, but apart from that it went well.

From brand new the mods seem to have made a difference. But obviously the caps need time to burn in.

Im a little reluctant to say how much of an improvement until i have done some more listening and the caps are burnt in. But overall very pleased with the results so far.

Ill post some more info and some pics shortly.

@ conrad hoffman

The oscons dont seem to have caused any issues due to lower resistance. I gather this would be evident if there was a problem of this type? Or would this only be evident with an oscilloscope?

Thanks again to all for the advice given.

 

[indianajo]

I found out I had 1 mhz oscillation in my disco mixer output after new fast op amp upgrade due to the fan running full speed on the PA amp @ 1 Vpp output (audio). No tell tale audio cue. Had to buy a scope to see it. Had to put .1 uf power supply bypass and 22 pf feedback resistor bypass caps on the fast slew rate op amps to get rid of it.

 

[Reference1042]

Short update......

I completed the mods to the dacmagic1. Replacing all the caps listed above. And installing the 8pin sockets for the opamps.

I have been trying different opamps, as was slightly disappointed with the results. I tried the ne5532p and lm4562na. All sounded ok. But...

Thought the lm was a bit wooly in the bottom end. And a rather harsh top end.

Then tried the ne and was slightly better. But seemed to lack mid range.

Neither had the WOW factor that IMO the opa2134pa had.....So went and bought 2 today.

I need to listen further.

Let you all know how it turns out.

Regards

 

[Reference1042]

Well, looks like the dacmagic doesnt like the lm4562's....after putting the opa2134's back in it sounds absolutely superb!!!

Now cant stop listening to the hifi again

Im a bit confused as to what has caused this, i looked at the datasheets for the lm and opa and there is a slight difference in how they are wired internally which i didn't notice before. Im not sure if this would cause the poor sound quality i was experiencing. But one thing is for sure, the dacmagic loves the burr browns!!

Has anyone else experienced similar results when upgrading to the lm4562's, and if so how did you get round it?

 

[starkeyg]

Did you put a power supply bypass caps accross the power supply pins or make sure the bypass caps on the pcb are quite close and a type that can handle high frequencies?

This chip is known to get into osciallation at very high frequencies if not bypassed well. Can sound like crap if this happens.

I've used these chips for a while and have really liked the sound I'm getting out. IMHO - YMMV.