Looking for a Supplier

ian field

Joined Oct 27, 2012
6,536
@ RB,
Does it work perfectly?
How accurate does the capacitance value have to be?

I've harvested multilayer ceramic chip capacitors from the old Pentium II cartridges and the driver panels from scrap LCD TVs - I get a lot of caps ranging from about 18uF to about 30uF, which are probably supposed to be nominally 22uF.

To be safe - the capacitor only needs to withstand the amplifier input range - which should be set out in the datasheet.

If that voltage is exceeded, you'll have bigger worries than the capacitor breaking down.
 

THE_RB

Joined Feb 11, 2008
5,438
@ RB,
Does it work perfectly?
I'm extremely happy with its sound quality, and it has never missed a beat in terms of reliability.

After building it I tested with a few different sine frequencies, and pushed the amp close to clipping. I scoped both waveforms (input sine and output sine) and they matched as perfectly as I could overlap them with both 'scope traces.

I'm following this thread with interest, because if there is an issue using a polarised cap for the 22uF cap I will swap my polarised caps for some 22uF NP caps (or two 22uF in series back to back).

That part of the circuit is using a 18k:1k divider from the amp output, so even with the max output swing of +/-35v the max voltage on that cap will be about 1.9v (in either direction).

From my limited understanding of that part of the circuit, at audio frequencies the cap voltage does not really have time to charge/discharge and will just sit around the average (0v) and have very little effect (other than acting like a short circuit in regards to the 18k:1k feedback divider).

Alec_T said earlier it should only get "a few mV swing" which sounds reasonable at audio frequencies. At very low frequencies there could be a couple of volts there?
 
Last edited:

Alec_t

Joined Sep 17, 2013
14,280
Alec_T said earlier it should only get "a few mV swing" which sounds reasonable at audio frequencies. At very low frequencies there could be a couple of volts there?
IIRC from a sim there's possibly ~1V below ~10Hz when the amp is being driven to the clipping point.
 

THE_RB

Joined Feb 11, 2008
5,438
Agreed, that cap only really has an effect at very low frequencies and rolls off the gain at very low frequencies (as the whole amp is DC coupled it needs that).

Do you have the simulator still set up? I'm curious to see if the polarised electro causes distortion on the bottom half of a sine waveform at very low frequencies.

Music has very little content below about 60Hz anyway, neither do normal stereo speakers. And if it's for subsonic woofer use the amp distortion does not matter much as the speaker deflection/flex/cabinet etc are already giving a massive distortion, and subsonic systems are only really good for reproducing the general pulse energy.
 

Alec_t

Joined Sep 17, 2013
14,280
I'm curious to see if the polarised electro causes distortion on the bottom half of a sine waveform at very low frequencies.
Ran a sim at 1Hz. No measured difference in THD for various polarised/non-polarised cap models.
 

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
Interesting..!

But still I will need the 10uf BP for the protector.
So I am going to get them both capacitors for future use.
 

THE_RB

Joined Feb 11, 2008
5,438
Ran a sim at 1Hz. No measured difference in THD for various polarised/non-polarised cap models.
Thank you very much! :)

I'm actually a little surprised though, I would have expected the cap voltage at 1Hz to be distorted, as a bipolar cap should not charge very high in the reverse direction (as they look more like a short to reversed polarity).

Did you check the waveform across the cap when running at 1Hz and large output waveform? That should show if the cap voltage is charging to match the + and - parts of the sine correctly, and what max voltage the cap charges up to.

Of course the THD test you did probably confirms the cap is charging equally to + and -, or there would have been significant 2nd harmonic distortion?
 

t06afre

Joined May 11, 2009
5,934
A PSPICE type circuit simulator will not care about which type of capacitor that is used. The all will used the generic type cap. Then a electrolyte cap is revers biased. You will have a electro-chemical reaction. That the spice simulator will not be able to simulate properly. From this link http://electrochem.cwru.edu/encycl/misc/c04-appguide.pdf
"Application of reverse voltage much beyond 1.5 V causes high leakage current quite like the forward conduction of a diode. Neither of these operating modes can be maintained for long because hydrogen gas is produced by the capacitor, and the pressure build up will cause failure"
For those who will go deaper into this. This free ebook may be of some help http://ebooksgo.org/engineering-technology/ElectrolyticCapacitors.pdf
 

Alec_t

Joined Sep 17, 2013
14,280
Did you check the waveform across the cap when running at 1Hz and large output waveform?
Yes. Clipping seemed symmetrical.
Of course the THD test you did probably confirms the cap is charging equally to + and -, or there would have been significant 2nd harmonic distortion?
Distortion was primarily odd harmonics. Very little 2nd harmonic.
A PSPICE type circuit simulator will not care about which type of capacitor that is used.
I believe you. I'm not convinced that the cap models are completely accurate regarding asymmetric effects with polarised types.
 

THE_RB

Joined Feb 11, 2008
5,438
Very interesting!

Maybe when Rif@@ gets his amp to the test stage he can scope it with a real-world polarised cap?

Otherwise I have everything needed here to 'scope my amp (that has polarised caps), but it is assembled into a rack full of music equipment at the moment.

I think the key might be what t06afre quoted;
"Application of reverse voltage much beyond 1.5 V causes high leakage current quite like the forward conduction of a diode. ..."

The waveform on that cap in low freq use is about 18 or 22 times smaller than the amp output waveform, so on my amp that's about +/-30v max which makes +/- 1.67v on the cap. It's possible the cap is able to charge and discharge fairly normally provided the reverse voltage is only 1.5v or so?

I can easily connect a sine generator through a resistor to a polarised cap and see what voltages it will follow. I'll try that if I get some time in the workshop later today.
 

t06afre

Joined May 11, 2009
5,934
The amp Rif@@ use behave like an opamp. So the voltage drop across the resistor and the series cap. Will be equal to to the voltage on the amplifiers + input pin. Then operating as normal. At reverse bias. An electrolyte cap behave kind of like a non polarized cap. At voltages around 1 volt. At least if it is not subjected to long term reverse biased voltages
 

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
I do have a scope but no function generator. I intend to use cool edit to generate sine wave from my PC.
My amp won't be ready soon cause the LM3886 has not yet arrived. Nor the PCB. The caps are yet to be ordered. My budget ran out.
It will take a month or so to get it running I presume.
 

Alec_t

Joined Sep 17, 2013
14,280
@RB
I can easily connect a sine generator through a resistor to a polarised cap and see what voltages it will follow.
I dare say it will follow ok; the question is will it damage (de-polarise) the cap in the long term?
 

ian field

Joined Oct 27, 2012
6,536
@RB
I dare say it will follow ok; the question is will it damage (de-polarise) the cap in the long term?
Tantalum capacitors are very vulnerable to going leaky if reversed - even small signal!!!

Aluminium electrolytics are a little more resilient, but may need re-forming after mild reversing.
 

ian field

Joined Oct 27, 2012
6,536
I have to ask, what is meant by re-forming a cap?
Its a procedure that's sometimes required on aluminium caps that have been in storage a long time.

The dielectric is actually a chemical layer on one of the porous aluminium foil plates - the electrolyte is caustic and eats away (very slowly) the dielectric in the absence of polarising voltage.

You "re-form" a cap by applying a current limited polarising voltage to gently and carefully get up to the rated voltage, you generally hold them charged for 24h then discharge them ready for use.
 

THE_RB

Joined Feb 11, 2008
5,438
@RB
I dare say it will follow ok; the question is will it damage (de-polarise) the cap in the long term?
That's an excellent question! :) I always assumed the damage occurs when the reverse current rises, making heat/electrolysis etc.

Assuming it can follow the 1v or 1.5v AC waveform without distortion, I would not expect much damage as the currents would be very low, and the AC voltage being an average of zero (reforming?).

In the case of a bipolar electro (or one made from two caps in reverse series) the mode of operation is that the reversed electro gets about 1.5v reverse max, where it starts to behave like a short, and then the other forward cap charges normally.

So in a bipolar cap the reversed cap half is still subjected to the reverse 1.5v, and the overall charge current. I don't think they get "damaged" from that?
 

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
The Gainclone amp I built from two kit amps just used a 22uF polarised cap in that part of the circuit;


http://www.romanblack.com/gainamp.htm
I was looking at ur Regulator circuit tht u build. I like to use it and need ur permission. so Can I ? pretty please. :D

PS.. I know u will say yes. So I was wondering what the is tht cap and input pin of LM317. The 0.0068 HV and the rest is all greek. What is tht u have written below HV.
U said u use 47μf as input bypass but in the diagram I see 33μf ( before the 2.2Ω 2W R between BE of pass Tr). So which is it?.
 
Top