Quality output by-pass caps

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
This one has probably been handled before but I'm new to these forums so be gentle with me.

In the Zen MOSFet Class A from Nelson he speaks of a by-pass output cap to sit accross the large electrolytic which feeds the speaker.

I am after a make/model and specs for such a cap. I have no idea if it should be .47uf or 100uf - Sprague or Solen?

Can some-one make a suggestion on what to use and what the affect is?

I built a Class A SE MOSFet amp (my own development) with a 4700uf output cap and would like to bridge it with a better audio cap.

All suggestions welcome.
 
They say that there are topics that one should not bring up in polite conversation. Religion, politics, sex...but caps should be part of the list.
Even though it was demonstrated conclusively twenty-odd years ago by Jung and Marsh that film caps are superior, you still have a strong contingent who deny that you should use any bypass caps at all. Of those who follow through with bypasses, you get people who fall in love with one sort of cap or another and keep suggesting it over and over in the same thread, telling you that they've had good results with it. The problem here is that you're likely to accumulate a half-dozen of these folks, each touting a different cap, and the result is a little confusing.
Some general guidelines. First off, most any size bypass will help. Second, there was an old rule of thumb--and you could do worse, even today--that you should use a bypass cap at least 10% of the value of the electrolytic. This isn't always practical, but it's a start.
You'll run into questions of availability and pricing. I have no idea what you can get in Australia. Be forewarned, film caps can get dreadfully expensive. Whether they are worth a given sum of money is a question you must consider for yourself.
Given a choice between Sprague and Solen, I'd take Solen.
The old Jung/Marsh paper is available on the web at www.capacitors.com. Read it. It will answer questions you didn't even know you had.

Grey
 
Hi,

I must say that Grey made a brilliant point! You'll just have to experiment a bit and find your ideal combination ...

e.g. - in one of my MC preamps I use ordinary 4u7 bipolar "no name" capacitor in the signal path - totally out of Q considering by most here - BUT - the preamp sings! not plays ... And I think that that cap is one of the reasons!
OK, not much help by this ...

Anyway - as I have Zen v4 and had experimented a lot with him - these are my conclusions:
- first I had 10.000uF/50V Jamicon + 10uF Wima MKS at the output ... so so ...
- then changed MKS for 1uF Wima MKP - a bit better
- the replaced Jamicon for 2200uF/50V Panasonic FC - and this is still the winning combination :) in my opinion (and also by some others that have listened to tha amp) ...

Just my 5 cents ...
 
Well, it seems every time a post sits for a day or 2 and then I respond, I feel walked all over by the details I left out, by the people who feel more inteligent or??? That is usually in other catagories here, not Pass Labs... Grey, you are not one of those people and I don't want to make you feel like I usually do, but you really did not point to any qualitative type of info... Oh yea, You did, The Jung Paper I guess... Anyway, Here's my understanding and rational. Big E Caps have impeadance that rises at higher freq.s. You obviously don't want to loose anything (Output Power) because of these ESR/Dissapation/etc effects. If you go to a film cap data sheet and find the graph that depicts Z or whatever at frequency, you will see they do better than any other type of cap. Ceramics used to be the old standard. You will also see that when you get to the 10uF size range they would be the best for our bandwidth. But, as Grey suggested, $$$, bigger is not only better here, it is $$$. Basically, in the output line, they keep the load Z flat way up in frequency.
Personally, I don't believe my caps in the signal path, or noise bypass in the supply liines, are as big a problem as working out the most linear operating range for the topology, and gain components in the design I'm experimenting with. Consequently, I have not done the subjective testing with them. But, I usually use the X2 type of cap. A metalized Polypropalene film, used usually on the input AC line. The basic reason is that, for me, they are easily available and somewhat plentifull. You can usually find .47-2.2uF at the input of a reasonable supply. I scavenge alot of $$$ stuff anywhere I can. Paralelling them, I believe, is an even better performance enhancer. My theory of the output E-Cap bypass,,, as big as you can afford$$$
You would be wise however to follow Grey's "politically correct" advise and check ot that paper I'm going too!
I guess I have generated enough "fotter" to debunk??? Do I understand correctly??? Anyone???
 
Still would love a solid suggestion

So I have a 4700uf eletrolytic output cap and this I bypass with?

From the combined afforts of the immediate responders (thanks to all <and lets hope there is more> - I really mean it) I will bypass this with at least a 470uf polyestor or polypropolene??

This size puts it way out of the range of Solen, Sprague etc. for size and price. .47uf Solen cost me $AU53 each.

So Let's drop the size a factor of 10.

So which big brand music cap maker has a 47uf offering? The amp (all two watts of it) has cost me $AU350 in parts to build but the result has been astonishing so I'm happy to sink more money into it.

I used Sprague .47uf input caps (orange drops) and there is only a total of two caps in the sig. path so a good bypass may even lift it more. The effects of a better output path should be noticed when there are only a few other components clouding the result.

Am I on the right path here? Would some-one like to suggest a better (and possibly cheaper) solution. Should I just use a polyester or am I better to go for one of the big names in audio caps?

Thanks for the help so far.

happy listening.
 
I am reluctant to tell people how much money to spend. That's your decision.
You will discover that 470uF film caps are rare, not to mention expensive. This is what I meant above when I said that it's not always practical to adhere to the 10% rule of thumb. You can get around the 470uF problem by paralleling two 220uF caps or perhaps five 100uF caps, but then there's still the money aspect to deal with.
I confess that I'm having difficulty reconciling your statement that you don't mind spending money yet want something cheaper.
Bear in mind that you don't need a really high voltage rating for this and that lower voltage caps are cheaper, all things being equal.
As far as largish caps go, the best candidate is polypropylene in the audio sense. Polyester is a lesser choice.
Read the Marsh/Jung paper before going any further. It won't help you with the pricing, but it will help you make somewhat better decisions about what sort of dialectric.
Incidentally, film & foil is generally better than metallized film, but in large caps you're not likely to have much choice. Most anything is better than an electrolytic. The only reason they're tolerated is that they are the only practical way to get thousands of uF in one reasonably-sized, affordable package. If it weren't for that, no one in audio would buy the silly things.

Grey
 
Bridging will work to the degree that DC offset is the same on both sides. If it differs by more than, say, a tenth of a volt, you're going to have serious trouble. Obviously, you'd prefer less.
Assuming that you have the DC under control, you'll need to bias more heavily for the same load. You could also parallel amps to provide the current. The hardware starts to add up, but it can be done.
Or you could just build a SOZ and get it over with. Current sources will work there, too.

Grey
 
Bypass Caps - choices

Grey - what I meant by spending money/something cheaper is "am I waisting money".

For example if a polyester will do and cannot be diffrentiated from a Teflon and oil (under critical listening) then the poly is the choice. But if it is the other way, that Teflon & Oil will eat the poly (but at a price) I'm happy to spend.

The suggestion was made to use a .47uf Sprague as a bypass on the 4700uf. Is this too small or should it do the job? It would be an easy solution because I bought 4 so I have two spare. Once more - it sounds a little small??

While I have everyone's attention - input caps direct into a power MOSFet -> I'm using a .47uf Sprague. I see amps with 10uf input caps and some with .1uf caps. I realise that with relation to input impeadance etc. bass can be effected - what is a good choice in size. I have seen Pass use all sizes into simular style amps.

The quality in this case I'm aware of - Teflon and Oil anyone??

Thanks for all your help so far.

happy listening.
 
hey - look - it's quite simple - you can calculate the frequence via:
f = 1/2*pi*R*C

For input - e.g. input impendance is 47kR and with 0u47 capacitor:
f = 1*1000000/2*3,14*47000*0,47 = 7,2Hz. Everything bellow 10-15Hz is allready more than OK!
For 10kR input impendance - 0u47 cap ain't good - you have 33Hz - that's too high ...

As with the output - the formula stays the same - just R is now few ohms (speaker impendance) and C is much higher (in your case 4700uF) - so from this point of view it doesn't matter if the bypass cap is 0u47 or 4u7 - what you get is total capacitance of 4700uF47 or 4704uF7 ... not a big difference ;)

As said - try it out with those 2 spragues! I know I would :cool:
 
I bit the bullet and along came poly

Thanks for all the comments and the formulas though have used them before – it was more where to use them that I needed to know. As an interim I dropped a 10uf poly across the 4700uf. The way I see it (and I remember from my speaker mass-production days) electrolytics act like diodes at high frequencies. They pass (not Pass) better in one direction than the other. This ‘tends to push the cones forward a little and hold them up there for short periods.

The higher frequencies only will pass through the 10uf poly bypassing the diode-imitating-fat-electrolytic. This should give me a cleaner top end – well it does.

I did the deed last night but finished it far too late to play. I had to hang the polys off the bottom of the cct. board (component to component wiring used throughout) because there was just no space on top. So while upgrading some in-expense book shelves, this morning, I have had about two hours of listening.

Rather than brighter top-end I would say cleaner. Maybe a hint more of intimacy with voice. Not a huge difference but enough to say the few hours it took to do the job was worth it. I decided to hang onto the Spragues for another amp to use as input caps – I think the Sprague are a good balance between cost and quality.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.