Hi all,

After full recaping my Beogram CD5500 (Oscons, Panasonic FM and FC) I achieved good results- better, more clearer and detailed sound. As output cap i used russian K73-16 22u/63 which are huge, but i found them a place. Now i`m thinking about recaping my Beomaster 5000. For now the amp is working just fine, so my idea is to change at least the signal path electrolyte caps in the preamp -  3x10u  polarized and 3x22u non-polarised Roederstein EKUs per channel with non-polarised (WIMA MKS for example). For my philosophy the electrolyte cap are not sutable for signal coupling - for my ears they "color" the sound.

I read the thread about the polyester caps and I have a question, especially for yachadm (Menahem) - I read that those caps could be changed with Wimas with much smaller values, I have never thought about that. Because the 10u Wimas are huge (the space in Beomaster is small) and expensive, what value and type would you recommend? And how about the CD, the huge russians?

Thanks in advance,

Tonislav

You can never replace a capacitor with a much smaller value. That would most certainly cut the bass. You can replace electrolytics with polyester types if you want. Or you could replace polyester with polypropylene if you're a real purist.

By the way, it seems some of you routinely replace the tantalum electrolytics. Why exactly? They don't have a liquid electrolyte that evaporates, do they?

Yup,about the capacity that`s exactly the  same as i thought. This is from Menahem`s post:

"

Today, with the advanced technology of film caps (Wima MKS2), you can quite easily use a 2.2uF Wima cap instead of the original 50uF (or even higher value).

I've routinely used Wima 3.3uF as a substitute for an electrolytic 220uF. And the Wima does a better job of blocking the DC.

The much smaller value means even less signal coloration than the original high uF electrolytic.

And that, after all, is the end goal of audio - to ensure that the original sound reaches the ear of the listeners.

Bear in mind, I am talking about only the coupling caps on the signal path, nowhere else.

"

Guess it`s OK to copy it without his permission, but it really interested me and made me start the thread.

About the bipolar caps, as I knew they are electrolyte, not tantalum. They contain two capacitors connected in series opposition. This conection is even worse for the signal than a single electrolyte and I`m convinced, that a polyester or polypropylene would do it better. In the other hand, if they were tanlalum I would replace them either, because i`ve heard tantalum as coupling cap and I would not recommend it.

Agent00soul:

By the way, it seems some of you routinely replace the tantalum electrolytics. Why exactly? They don't have a liquid electrolyte that evaporates, do they?

No liquid as far as I remember, but these old drop tantalums still can go bad. And often in an unpleasant way, like transforming into a smallish resistor that can actually burn out if it is across a supply line. Otherwise they are more reliable than aluminum electrolytics.

Everybody has his own ways, but as a first rule, I tend to leave alone components that have nothing wrong with them. Especially if the kit under scrutiny isn't working at all, it might not be wise to alter too many things at the same time.

The main reason B&O used so many tantalums in these old designs was most probably just the physical size. If you don't care about the price, the smaller tantalums in the signal path can well be replaced with MKS - conveniently smaller now than 30 years ago - or whatever. The end result might be better than new, or at least not worse. But in many other cases the circuits may well have been designed with the deficiencies and characteristics of certain component types in mind, and it won't be sensible to go "upgrading" everything.

After I first got my BM5000 working, I thought it sounded mostly like crap. Along my first rule, I replaced the electrolytics that a) were bad according to my ESR meter and b) actually had a chance of affecting something with their badness. As an exception, if I find a cap bad in one channel, I replace the other one too to keep everything symmetric, and also if the cap type or location is known to be troublesome. No use leaving a brown Roedenstein stay when the other 29 of its kind have already gone bad.

I then replaced the 4.7u and smaller tantalums in the signal path with MKS's. I thought about replacing the bipolar caps as well, but they didn't measure too bad and it is hard to find non-electrolytic replacements for them. And more importantly, the receiver sounded absolutely fine as it was at this point, so it would'be been just wasted time & money to me.

I do agree that the signal path caps may deserve some attention. B&O was an early adopter of technical solutions and circuits, like digital volume & tone controls and signal switching, that required plenty of decoupling caps. I once went over the schematic of some upper-class integrated amplifier (far from high end silliness though, I think it was a Sansui from their better days) and it managed with four or five capacitors on the signal path from a line level input to the speakers. I think a Beomaster 5000 has more than double that, and it isn't the worst!

fagaldi:

Today, with the advanced technology of film caps (Wima MKS2), you can quite easily use a 2.2uF Wima cap instead of the original 50uF (or even higher value).

I've routinely used Wima 3.3uF as a substitute for an electrolytic 220uF. And the Wima does a better job of blocking the DC.

The much smaller value means even less signal coloration than the original high uF electrolytic.

And that, after all, is the end goal of audio - to ensure that the original sound reaches the ear of the listeners.

Bear in mind, I am talking about only the coupling caps on the signal path, nowhere else.

This was the worst piece of crap I have read in a very long time! Yes, of course it will be good at blocking DC. But it will block your bass range too!

You can't beat the laws of physics. Your cutoff frequency will always be F = 1/(2*pi*R*C). If you make C smaller, the cutoff frequency of the high pass filter will be higher and you will lose your bass range.

I usually will just replace the whole lot when we are speaking of electrolytics. Often I will spot leaking ROE's throughout the whole system on some of the older stuff say pre 1980.

Some of the earlier Beomaster 5000's used them as well. My thinking is I am already in there so why not.  I buy in batches a couple times a year. What I dont do is mess with the MFD value. If I am replacing a 50uf thats is what I replace it with.

My caps of choice lately have been Fujicon RK seriies and Xicon. I have liked the results and I can can buy cheap in say lots of 25.

Hello,

To retain the character of the (vintage) equipment, I now replace with the exact same type of capacitor. Not quite audiophile, but I am not one anymore!

Thanks for your replies.

I`m also an audiophile - I prefer listening high sensitive fullrange driver in horn, shortest signal path and stuff. Obviousely the Beomaster is not exactly an audiophile product - too many chips, signal caps, low power amp bias current, but I think that the CD5500 and Beogram5500 are. I bought the Beomaster mostly because the remote function, but for now it`s my main amp for my stereo system. I like the whole system very much - stylish, small, looks like "out of the box".I even bought the deck because of that-i kind of falled in love in my system :) . That`s why I`ll try to squeeze the best from my Beomaster and than decide whether to use it for amp or use only its comunication features with an external tube amp.

Tonislav

Agent,

"This was the worst piece of crap I have read in a very long time! Yes, of course it will be good at blocking DC. But it will block your bass range too! "

With all due respect, your comments overstepped your mark. I specified that these are the signal-path COUPLING capacitors, not decoupling. You are right for decoupling, but you are off the mark for the COUPLING capacitors.

One has to look very carefully at the circuit, before making a decison. Neither I nor anyone else will take responsibility for random changes made by a member with no understanding of what goes on in the circuit.

Many times, if you look at the circuit, you will see that B&O put, for example, a 10uF coupling cap at the source, and then further down the line a 22uF cap. Well, there's no sense to that at all, and I cannot understand why it was done. You cannot increase signal that was never there in the first place. If you know what you're doing, and see this kind of thing, you can change the 22uF to a 10uF, and not affect the sound.

Furthermore, using a 4.7uF Wima instead of a 22uF old Philips electrolytic  as a COUPLING capacitor does make sound-sense. You can do a hearing experiment, swap them out, and listen for yourself.

These are my conclusions from Practical Work, not numbers and theory.

Again, I refer to COUPLING capacitors.

And before you make statements like that again, make sure that you read ALL the facts first, and fully absorb and understand what is written. Clearly here, you chose to ignore the fact that I referenced COUPLING capacitors only.

Menahem

Yes, this wasn't very nice at all. People here are sharing their knowledge and it's free. We are all glad to do it!

Menahem, your reaction is that of a gentleman, which we appreciate. And please keep informing us about what you do!

yachadm:

Agent,

"This was the worst piece of crap I have read in a very long time! Yes, of course it will be good at blocking DC. But it will block your bass range too! "

With all due respect, your comments overstepped your mark. I specified that these are the signal-path COUPLING capacitors, not decoupling. You are right for decoupling, but you are off the mark for the COUPLING capacitors.

One has to look very carefully at the circuit, before making a decison. Neither I nor anyone else will take responsibility for random changes made by a member with no understanding of what goes on in the circuit.

Many times, if you look at the circuit, you will see that B&O put, for example, a 10uF coupling cap at the source, and then further down the line a 22uF cap. Well, there's no sense to that at all, and I cannot understand why it was done. You cannot increase signal that was never there in the first place. If you know what you're doing, and see this kind of thing, you can change the 22uF to a 10uF, and not affect the sound.

Furthermore, using a 4.7uF Wima instead of a 22uF old Philips electrolytic  as a COUPLING capacitor does make sound-sense. You can do a hearing experiment, swap them out, and listen for yourself.

These are my conclusions from Practical Work, not numbers and theory.

Again, I refer to COUPLING capacitors.

And before you make statements like that again, make sure that you read ALL the facts first, and fully absorb and understand what is written. Clearly here, you chose to ignore thae fact that I referenced COUPLING capacitors only.

Menahem

Menahem and I have not always agreed on everything, we both know that.
I would have a hard time believing if Menahem knows this little about
capacitors and electronic basics so I suppose he just got something mixed up.
To prevent this from ending in a war on words, let's straighten this out so everybody
knows what we are talking about:

Coupling capacitors are doing exactly that.
They couple from one stage to the next. As that, they can often be seen
as a series component in schematic diagrams.
In other words; They conduct wanted AC signals.
They isolate the DC working levels from the previous stage to the next.
At the same time, they allow the programme signal through.
The capacitance value of the coupling capacitor is calculated based on
several factors; The desired frequency range is important but the
output/input impedances of the actual stages are even more so.
Having a 10uF coupling capacitor one place and a 22uF coupling capacitor
in a later stage can make perfect sense.
It's all down to the impedances of the individual stages.
Think of it as RC-filters, where the R is the input impedance and the C is
the capacitor. Then calculate the filter to avoid influence within the desired
frequency range, while still adapting impedances.

Changing a coupling capacitor from f.e. 22uF to 4,7uF will spoil the frequency
response, the impedance matching will be ruined and the sound level will
be lowered at certain frequencies - or the whole range, taking the linearity off.
Fact.

You can mount different types of coupling capacitors to achieve a slightly
different sound "color" but you shouldn't change the values unless changing
the circuits accordingly.
If you like the sound better with other cap values, then fine
but it does not calculate nicely and it makes no technical sense. Quite contrary.

Decoupling caps usually goes to ground to remove AC from
specific points in a circuit to provide a stabile DC working level
for the specific circuit and/or to prevent self-oscillation in amplifiers
with a certain amount of positive feedback.
In other words; They conduct unwanted AC signals.
In amplifiers they can also adjust the amplification by regulating
the feedback themselves, but that's a longer story and they still connect
in some low impedance way to ground.

Changing the value of decoupling capacitors can have little, some or a lot
of influence depending on the actual location in the circuit but generally
these should not have their value changed either unless radically modifying
the circuit in which they work (as Die_Bogener suggest when changing OpAmps, I
think it was in the Beolab 2500 where the decoupling caps are no longer
needed due to a different construction of the new OpAmp) or if the circuit proves
unstabile.

Your statement:
"The much smaller value means even less signal coloration than the original high uF electrolytic".
is hopefully not what you mean;
Capacitors with large capacitance values transfer AC signals with less loss than caps with smaller values.
Fact !
The rest depends entirely on the position of the capacitor in the circuit and, as stated above, a
change of capacitor value, be that to a larger or smaller value, can cause havoc.

Martin

Thanks Martin for taking the time to explain the facts once and for all!

chartz:

Yes, this wasn't very nice at all. People here are sharing their knowledge and it's free. We are all glad to do it!

Unfortunately, in this case he was not spreading knowledge, but total nonsense. It is totally against signal theory and his "advice" will at best cause disappointment, when followed.
By the way, his reply in this thread is just as much nonsense. Read Dillen's post above if you want to learn the facts.

Thank you Martin.

I am always willing to be corrected when wrong, however, being an experimenter, and tenacious to find the best solution - I rarely give up - I have found IN PRACTICE, that swapping and experimnenting with COUPLING capacitors sometimes makes a big difference in sound.

Yes, I have found that in a certain circuit, putting a Wima 4.7uF in place of a 22uF electrolytic, sounds clearer and more musical. Why - it doesn't match the numbers, as Martin correctly writes. But it just does.

Same idea as why an OPA2134 opamp sounds better than than some other opamp, even though the numbers say differently.

Or an MJ15003 sounds better than a 2N3055 - just does.

I have come to the conclusion that achieving the "best sound" is a black art at best, and in the SIGNAL path, unexplained variables come together to please the ears where the numbers theory may postulate differently.

So, maybe I am not putting it in English clearly enough, but I prefer to keep an open mind about this signal path magic, and if Agent wants to call it nonsense, then I think he's missing out on a lot of creativity in this very enjoyable game.

And even though this is my work, having been building radios for more than 40 years now since I was a kid, I still enjoy the experimenting  very much. I cannot limit myself to laboratory-style sterility of strict numbers - I think I (and my customers) would miss out on a lot of satisfaction.

Experimenting with signal path COUPLING capacitors creates no electrical danger, or reliability issues, so why keep such a closed mind?

I've forgotten the vast majority of the formulas I learnt in university over 30 years ago, and that's fine with me. It seems that Agent knows his theory well, and would probably beat me in a test - he's probably younger than me, and so he should know his theory better. What's replaced the theory, is valuable experience. And that experience is what guides me today. But I'm not too proud to learn new things from others who may have accumulated different experience from me.

B&O invests a lot of effort conducting listening tests on new products before releasing to production. A smart idea. And justifiably so. They also understand that no matter what the theory and the numbers say, the "ear" factor is what sells. I'm sure they have changed certain components, when the theory says otherwise, just because the listening tests required it.

Remember, we are dealing with mostly vintage equipment on this site. 30 years ago, the type of capacitors available were completely different to those of today. We have a lot more choice available today, as well as real audio-capacitors by Elna, Nichicon and others. It's worth it to experiment in the signal path!

My two cents!

Menahem

Menahem, it seems you have still not understood the meaning of the words coupling and decoupling capacitors.
Coupling capacitors couple AC signals from one circuit to another, but block DC (read Martin's post again). Decoupling capacitors are for GETTING RID of UNWANTED signals from a circuit. This is usually the case in power supply lines. In audio lines, you might want to decouple HF to reduce noise. Decoupling capacitors are basically connected from the circuit to ground, or sometimes from the output of an inverting amplifier to the input of the same amplifier (for decoupling HF). The latter is called a Miller capacitance.

Agent

It appears we are talking at cross-purposes here.

Believe me, I know the difference between

1. COUPLING - which transfer the AC signal down the signal path from source to loudspeakers. I never forget the old saying - No capacitor is the best capacitor - applies especially to the signal path COUPLING capacitors. To approach that true condition in which the capacitor is totally neutral, and does not color the signal at all, so we must keep an open mind and be creative in selecting the COUPLING capacitors.

I have found very good results in investing in audio-quality caps for coupling, and IMO, are worth the investment.

and

2. DECOUPLING, which are usually connected to ground, and which I almost always increase (and never decrease) the capacitance to provide better signal timimg (for example, cleaner bass, among other improvements).

I am still very skeptical of the value of audio-quality caps for use in decoupling - my ears have not convinced me of their value.

It is only the COUPLING capacitors which I would consider decreasing the capacitance, and then only according my ears!

Regarding Decoupling - in the vintage circuits here, the designer's capacitor choice was limited in their capacitance by their physical size. Today that same space can accommodate a unit the same size, but of a much larger capacitance, with substantial improvement bin signal quality. And being that the older capacitors were often specc'd at -20 +80%, there is certainly a lot of flexibility to increase their capacitance today.

Menahem

yachadm:

It is only the COUPLING capacitors which I would consider decreasing the capacitance, and then only according my ears!

Well, the coupling capacitor (C) forms a high pass filter together with the load impedance of the following circuit (R). The cutoff frequency of this high pass filter (F) is 1/(2*pi*R*C). If you reduce the coupling capacitance, you will increase the cutoff frequency of this filter. This means you will filter out some of your bass range unless the cutoff frequency was chosen extremely low by B&O (not very likely).
You might like less bass but I doubt most other readers on this forum would like to ruin their bass response.

OK, I got your theory, and I understand it!

Just my ears tell me differently sometimes!

And I do like bass - lot's of it!

I couldn't find any commercial subwoofer which I liked, so I built my own - 7cu ft, 2 ports, 900W RMS with TS2000 driver, Fs 17Hz.

Perfect!

Menahem

Ok.

I just like to add that any filter has a phase shift around the cutoff frequency. So for this reason, you might want to select the coupling capacitor value for a considerably lower cutoff frequency than the limit of hearing – say 5 Hz. If B&O actually did this, you might not lose much bass response if you reduce the capacitor value somewhat. But you will affect the phase response in the bass range. Bearing in mind that B&O developed their speakers for the best phase response, this seems a bit unfortunate.
Also remember that the level is 3 dB down at the cutoff frequency. So if you have say three stages with the cutoff at 20 Hz, the entire amp will be 9 dB down at 20 Hz and the 3 dB point will be at a higher frequency. And the phase shift adds up too. So with three stages you'll have a 270 degree phase shift at the cutoff frequency.