PR99 replacement capacitors

[Linnovice]

I’m about to undertake a complete recap of my Revox PR99 MkII. Is there a brand of capacitor that is considered as superior for audio use?

[Ted Kendall]

Panasonic FC series from RS are good and reliable - never had trouble with them.

[Radio Wrangler]

There are several manufacturers of good quality, reliable capacitors with excellent performance characteristics. The main issue is to be confident of getting real ones not faked ones. Stick to buying from a major distributor like RS, Farnell, Mouser etc. who have carefully managed supplies from the manufacturer. Capacitors are cheap, so the savings to be made are small, so it's a matter of the labour of changing them all again if you discover you bought dodgy ones.

There's a lot of hype about audio parts, but that's what it is, hype. Reliability is, however a very real difference, and the large scale manufacturers have excellent research and quality control facilities and it shows in their parts. There are fancy boutique audio capacitors but who made them? are they re-labelled ordinary ones? they could be real. But their small turnover means that the actual manufacturer won't have statistical quality control systems running like the principal manufacturers.

Panasonic, Nippon Chemi-con, Nichicon I've found to be very good and I've tested these things at high temperature, fast temperature changes, high altitude, high humidity and very low (-55C) temperatures all while pushing them hard with very fast pulse currents close to their ratings. Wurth, the German inductor firm have entered the field and they ought to be very carefully done. They have a good reputation in their existing products and I think they want to actively keep it.

We've all assumed you mean electrolytic capacitors. They're more variable than non-polarised types. Polyester/Mylar film capacitors can exhibit dielectric absorption and show some peculiar behaviours in precision low frequency analogue circuitry, so I avoid them myself. Polypropylene and polycarbonate ought to be OK.

Ceramic capacitors are a world in themselves. X7R dielectric are OK for decoupling but I'd avoid them for changing signals. The higher-still dielectric constant materials, I'd avoid. For smaller values, the NP0 and C0G dielectrics are very good. There's a wide choice of reputable makers, Yageo, Vishay, Murata, with not much reason to choose between them.

Be aware that there is now a global shortage of ceramic capacitors. Prices are going silly as delivery dates are shooting out... well over a year for some parts!

Tantalum capacitors are not good choices nowadays, even known name makers are turning out marginal parts to combat the rapid increase in the price of tantalum dust. Ask yourself if an aluminium part will do - they've been improving as tants have got worse.

It's always worth asking yourself why you're changing them all.

Probably the non polarised parts are OK.

I re-capped an entire revox B261 tuner. It had been rack mounted and run 24/7 for decades. Many electrolytics had dried and gone high ESR. Changing the lot was the right move, given the circumstances.

If the PR99 is for yourself, use what you feel like.

If the PR99 is going to be sold, a significant number of purchasers have definite ideas on what capacitors they want and these choices may not be rational, but they affect selling prices.

David

[Linnovice]

Morning all, thanks for the advice. David, especially thank you for the detail and explanation. You ask why I am considering a recap. I’ve got the machine in my workshop for its annual checkup and service (sounds like the car!) and I’ve noticed that when checking the frequency response there’s a difference between the channels of 2-3dB. This is not off the tape but in the machine itself. As it’s around 30+ years old I’m thinking it’s time for the recap. I’ve already changed the Fraco and Rifa caps as well as the motor run caps. In fact, on thinking about it, all of the electrolytics have been changed. Time for the tantalums?

[Ted Kendall]

When you say "a difference" is this one of level or frequency response? E-E the machine should be pretty flat and square. I'd put a 1 kHz square wave in and check with a scope at various points in the signal path to see where things change. This sounds like a specific fault rather than sagging caps.

[Radio Wrangler]

I agree with Ted, you need to track the cause down. 2-3dB is enough to show easily in tests.

David