Another capacitor question

[Nebogipfel]

Hello chaps.

I know I have read this but can't seem to find the answer in searches .....

When substituting modern caps I know you choose the nearest modern value eg 47uF for 50uF

Not all values are that close and my question is, as a general rule do you go down to the nearest modern value or up?

On the A205 amp board for example the originals are 40, 80, 125 and 400 .....

Sorry if this is boring and thanks

[Eddystonepete]

Hi N, it used to be said that the tolerance on electrolytics was -50 to +100 per cent. In those days we didnt have any way of measuring them. Today cap measuring meters are cheap and accurate, Ive found that most small el caps are within 5 percent of marked value. As to what to replace older values with, Its not that critical. If a designer wanted an exact value of capacitor he would never specify an electrolytic. either go up or down one value and note the result. Why not experiment a bit and report back to the forum!. Many years ago someone told me not to use a much higher voltage rating than the original electrolytic as under voltage will prevent them from polarising properly. Seemed sensible at the time.

Peter

[tiv48]

Using too high a voltage will be ok but after a time the cap may not work at its rated voltage
Rob T

[boater sam]

Not boring at all.

As has been said, the tolerance was massive in years past so don't worry too much about the exact values.
For smoothing capacitors, bigger may be the way to go.
For reservoir capacitors - beware - with valve rectifiers there is a manufacturer's maximum value to avoid valve damage so near or slightly lower value is best.
FM discriminator capacitors are usually small values, 5uf equates to 4.7uf etc.
Cathode bypass capacitors, often 25uf so anything from 20uf to 33uf, larger will increase the negative feedback in output stages, a bit less gain.

Hope that helps, 
Sam.

[ppppenguin]

Sam, I think you mean that SMALLER will increase the negtive feedback. Unless you go much smaller it will only affect the bass

[boater sam]

Thanks Jeffrey, you are correct of course, non is negative feed back max. big is base cut.
Don't be surprised at the odd mistake, I've been in the sun too long, I'm aging by the minute, waiting for a beer.
 
Reading some of the questions on another forum makes me wonder why  I bother replying but I try to help.

Sam.

[Nebogipfel]

OK thanks for that chaps. I was specifically thinking of Hacker transistor sets but assume the tolerance is pretty wide on those too.

[boater sam]

If you read through a lot of the forum threads you will find that folks with less experience worry about tolerances far too much. Old components were  never made to close tolerances, it cost the set makers too much brass. Rarely did they use any resistors better then 20%, paper/wax caps were possibly worse, hard to tell as they are mostly duff now, and what few electrolytics they used were as bad as +100%/-50% and they were expensive then.

Things were a bit better when transistors came in but the cost cutting got worse if anything and component quality suffered, look at the Hunts caps problems in the late 50s sets and the poor electrolytics in in many tranny portables.

Sets were designed to be less critical of deviations, that's why so many still work "off the bat" today. I always advise folk just to get it working first rather than chase around swapping parts that are probably still acceptable and get into all sorts of mess.

Sam.

[Terry]

If you read through a lot of the forum threads you will find that folks with less experience worry about tolerances far too much ... 

I tried to dispel this worry some time ago and explode the myths surrounding tolerances in this VRAT thread

One thing about resistor values, in particular, which confuse the novice is the seemingly random sequences of values whereas if to look at the chart I drew (reproduced below) you an clearly see how cleverly the values where chosen to neatly produce a continuous sequence, irrespective of tolerance.

... Old components were  never made to close tolerances, it cost the set makers too much brass. Rarely did they use any resistors better then 20%, paper/wax caps were possibly worse, hard to tell as they are mostly duff now, and what few electrolytics they used were as bad as +100%/-50% and they were expensive then ...

I would guess that early E6 batches were made by guesswork and the value allocated by measurement afterwards.

I you look at the early colour code - body - tip - spot instead of the later rings - this might give us a clue.

Say you intended to produce a batch with values of around 500Ω, say.

Now measure them. Analogue meters would be used, of course, so it would make sense to have scale with 6 divisions on it, one for each value in a decade, as the absolute value didn't matter anyway! You don't need values, either - colours will do!

So resistor one falls in the orange sector - dip it in the orange paint pot and hang out to dry. The next one scores blue, so dip it in the blue paint pot and leave that out to dry, too. Now one scores in the yellow sector - repeat using the yellow paint pot this time.

When this initial sorting is finished, each resistor will have body colour corresponding to the first figure of its value. Resistors painted red or orange can skip the next stage, because the second digit of their value is the same as the first but the rest will have one end dipped in the appropriate colour for this value, so the blue ones will be dipped in grey paint, the yellow ones in violet paint and so on.

Finally, all of the resistors will have a brown spot added to complete the code.

Easy, Innit?

Care would need to be taken with the first two segments, as they would both be black, and the resistors would have to be segregated during the drying phase so that they received the correct second colour and further care would be needed where selections crossed the decade barrier to ensure the correct spot was added but each procedure could be carried out  very rapidly, which would be vitally important, of course.

   

[Terry]

Care would need to be taken with the first two segments, as they would both be black ...

That should, of course read brown!