going up with capacitance

[sssboa]

So I will recap another Sony STR-6065.
I looked into capacitors list and it gives the values with -10%+100% or similar
like 470uF -10%/+80%
Sometimes it goes as far as +150%, minimum +50% for electrolytic caps. But it's always -10%.
Now today's caps are kind of -15% never + anything.
So I think of going up with capacitance like this
33uf into 47uf
47uf into 68uf
470uf into 560uf or 680uf
1000uf into 1200uf
4.7uf into 6.8uf
etc.

What do you think?

 

[DRecovery]

This is something you want to do, or something you think you should do?

 

[sssboa]

This is something you want to do, or something you think you should do?

I don't want the amp to be borderline. When I measure the original caps, 33uf measures really as around 47uf. And you measure today's 220uf and it shows you 190uf and I mean expensive series of Nichichon.

I am afraid I may lose something doing it the usual way.

 

[tarior]

You won't lose anything replacing like for like. Guaranteed.

 

[restorer-john]

The trouble is, they aren't like for like. Physically smaller has proven itself to be less capacitance for your money and less capacitance than is printed on the label.

Typical these days unfortunately.

If they are starting out borderline or below their specified value, they are only going to go down as they dry out.

We need to bear this in mind with all our work. There is very little reason to not install PSU caps that are +50% rated over the originals, as many of the originals were actually 50-100% over in the first place.

 

[Lavane]

You won't lose anything replacing like for like. Guaranteed.

Plus you can get better made modern caps for not much $$.

 

[telemike]

I'd be afraid of altering the circuit and messing up the sound

 

[tarior]

The trouble is, they aren't like for like. Physically smaller has proven itself to be less capacitance for your money and less capacitance than is printed on the label.

Typical these days unfortunately.

If they are starting out borderline or below their specified value, they are only going to go down as they dry out.

We need to bear this in mind with all our work. There is very little reason to not install PSU caps that are +50% rated over the originals, as many of the originals were actually 50-100% over in the first place.

I strongly recommend against anyone who isn't a tech just going in and wholesale installing all new caps at +50-100% of original value. Yes, it very likely won't hurt anything in a power supply.
However, in a signal path, bypass, high/low/bandpass filter, Zobel, feedback network, etc., it could make a big difference.

 

[sssboa]

I strongly recommend against anyone who isn't a tech just going in and wholesale installing all new caps at +50-100% of original value. Yes, it very likely won't hurt anything in a power supply.
However, in a signal path, bypass, high/low/bandpass filter, Zobel, feedback network, etc., it could make a big difference.

I did some math, It would be like up to +45% of the nominal value on the can for caps on board. I would like to install 10000uF filter caps in place of 6000uf only because they fit the big 50mm bracket. The 6000uf caps measured 8200uf.

 

[restorer-john]

Read my comment:

"There is very little reason to not install PSU caps that are +50% rated over the originals"

No one in their right mind would advocate wholesale increases in all capacitors.

 

[restorer-john]

They will be fine.

 

[petehall347]

i have tested old faulty capacitors at way over rating for uf .

 

[sssboa]

Read my comment:

"There is very little reason to not install PSU caps that are +50% rated over the originals"

No one in their right mind would advocate wholesale increases in all capacitors.

It's actually what I'm gonna do
all caps up by 10-45% but not more than that.

 

[sssboa]

i have tested old faulty capacitors at way over rating for uf .

Now there are many theories to that.

Some people say that back in the 1970s caps were produced with tolerance of -10%/+100% and similar, just like now they are made with tolerance of -20%/+20%.

Others claim that with time capacitors may in decades actually increase their measured capacitance and dicrease esr that does not mean it's "real capacitance" and "real esr".

 

[petehall347]

its hard to say what the old ones read on the meter when they were new . unless anyone remembers testing new ones back then or if data sheets exist .

 

[tarior]

Read my comment:

"There is very little reason to not install PSU caps that are +50% rated over the originals"

No one in their right mind would advocate wholesale increases in all capacitors.

I read it, and understand you just fine. A lot of people might not. Hence my comment.

I've measured many, many 30-40 year old caps just out of curiosity. Quite surprising, the high majority of them were well within 10% of indicated value.

 

[Bob]

if you're keeping it forever then Ok. otherwise, if you don't document your changes,
the next guy opening the unit up finds all the caps are different from the
parts list and the schematic and yells out...

If it were me, I'd look at 105 degree caps, caps designed with low ESR, bipolar caps,
and if the value is small enough, then mylar or even the better polypropylene
(wimas in both cases). the specific choice will depend on where in the circuit
the cap is. this is going for long term reliability and better sound quality.

the difference between an expensive cap and an "average" cap is only a
few cents and the real costs is in the labor so your DIY allows some leeway.

it would be your choice as to cap vendor. I prefer Nichicons for their color schemes
gold, blue and green. also wonderful for those who like color.

last note, whether you "up" the cap values or you "boutique" the caps, be very careful
regarding the lead spacing and lead type. large caps use a variety of mounting
styles, replacement caps (same value, different value, different cap types) may
have different lead spacing make it a headache. check first. lastly, some caps
are axial (others mostly radial) and these may prevent ordering the next quantity
level for better pricing by grouping all the same values into one line item.

changing values of caps is second guessing the original Sony design engineer
and if he ever listens to it - he might pull out his katana and hack your sony
to pieces saying - that's not the sound as I designed it.

 

[Retrovert]

An important reason exists to not dramatically increase capacitor size in a rectifier: reducing the conduction angle, i.e. the region of the AC cycle in which the filter capacitor recharges, increases the load on the rectifier and transformer. This is a fraction of the entire cycle. The shorter it is, the greater the current draw in that short region.

Here's why.

The capacitor only charges when the applied voltage exceeds that currently stored. Bigger capacitors have lower voltage drop, so the charging time is reduced. That means the charging current must be supplied in a shorter timeframe. Same overall number of electrons, just provided in a shorter time.

Conduction angle is why tube rectifiers have a specific capacitance limit; that is really specifying a minimum conduction angle, but in a way that is far more understandable and far more easily honored.

A silicon rectifier has higher current limits, of course, but the rectifier may not be able to drive that load in a short timeframe without overheating. The transformer is the same way. So conduction angle matters for solid-state, as well.

In addition, the shorter conduction angle begins to look more like a square wave, which modulates the transformer, causing it to ring like a tank at whatever frequency is formed by its parasitics. Snubbers are therefore needed. Which is a good idea in any event. As is switching to lower Qrr rectifiers and adding snubbers to those as well.

Making changes when one does not understand the circuit is risky. When the original design was kaizened to reduce manufacturing cost, it is extra risky.

Will the diodes and transformer supply that increased current in a shorter time? I don't know. Maybe. Maybe not.

I suggest you maintain the original values for the power supply, and, indeed, for the remainder of the circuit, until you fully understand the ramifications of such changes.

 

[restorer-john]

Retro, you need to let this go. Seriously. Regurgitating the same old and tired stuff that is utterly irrelevant to what we are discussing here is tiring. Read the comments, measure a few hundred capacitors, both NOS (new old stock) 30 year old caps, old cap pulls, and brand new stock and come back to the discussion.

We know what we are doing. Many here have studied electrical engineering, many here are technicians with decades of experience. Many are audio designers. We even have avionics technicians here.

We know that modern new stock capacitors on the whole measure consistently low, quite low and very often, under spec. Old capacitors measure the opposite. In PSU filtering, this makes a significant difference to many measured results.

 

[gslikker]

In the past, there were even techs considering any electrolytic cap below value as defective.