10pF cap to opamp

[snk]

Hello,
In some circuits, I see a small (10pF) capacitor to opamp pins. I understand it is an advice taken from the datasheet, i order to avoid stability issues.
My "general" question is the following : is the 10pF value mandatory, or is it a "ballpark" value? If so, what is the acceptable range?
My "pragmatic" question is the following : I am building a Bajaman Opto Limiter. I have several 15pF (multilayer monolithic ceramic) and 33pF capacitors (ceramic disc), but I don't have any 10pF anymore : Is it safe to use a 15pF or 33pF cap, or should I be patient and wait to receive some 10pF?
Thank you in advance!

[duck_arse]

hmmm. I can't see such a cap on any circuit I could find. and didn't want to trawl 7 pages at the other place.

if you mean a cap across a high-value feedback resistor, it is there to reduce gain as the frequency increases - at some frequency [and above] the cap looks like a resistor of the same value as the feedback R, and so the gain is reduced by half.

10pF is not likely to be worried by 15pF, to the extent that the build fails.

[Elijah]

Hi. Can't access the link. But if you're talking about dual channel op amps (4558, tl072, ne5532, etc.) and the cap is connected to the output pins it's there to limit the high freqs a little bit (51pF in TubeScreamer for instance). Basically it removes that hiss. But if you're talking about compensation caps like in rats which uses single channel opamp and connects pins 1 and 8 it's there to control the slew rate. The lower the cap the higher the slew rate. That's what I can say.

[duck_arse]

Elijah - there is a rule here that we aren't allowed to have links to the other site, so the offending words are replaced with "....". it is left to you to fill in the forbidden words.



if you are following the Holy Roller circuit, the 10pF parallel to the 4k7 sets the upper freq to 3.4MHz, and is basically useless. it should be more like 1nF, which is still 34kHz cutoff.

[amptramp]

I take it you are referring to the capacitor from the output to the inverting input.  If you look at a typical inverting stage, you have an input resistor and a feedback resistor.  There is some capacitance to ground at the inverting input.  This bypasses the feedback to ground at higher frequencies and causes the output to increase as frequency rises.  When this rise intercepts the falling high-frequency open-loop response of the op amp itself, you get a phase shift that can occasionally cause oscillation but more frequently causes an unwelcome transient response.  This can be alleviated by putting a feedback capacitor in parallel with the feedback resistor so you have a capacitive divider for feedback.  This can also be used to set the upper frequency response to limit the noise spectrum.

There are some op amps that require a compensation capacitor which does the same thing for internal stages.  From (bitter) personal experience, I can say do not attempt to skimp on these capacitors.  I wasted an entire Saturday once because a compensation capacitor that should have been 100 pF actually measured 91 pF and I couldn't get the stage under control without oscillation.

[antonis]

if you mean a cap across a high-value feedback resistor, it is there to reduce gain as the frequency increases - at some frequency [and above] the cap looks like a resistor of the same value as the feedback R, and so the gain is reduced by half.

Only at very "some" frequency...
Above it, gain is reduced by more than half 'cause cap's impedance is lower than feedback resistor's value..

P.S.1
Totally agree for particular value/circuit total uselessness..

P.S.2
:cider:  :cider:  :cider:

[snk]

Hello,
Thank you for all your replies, it's all very clear.
Also, apologies if I posted a link to a "forbidden" site, I didn't want to offend anyone or break any forum rules. I had read about "incompatibility" between the two forums a while ago, and had since totally forgotten about it.

So, to go back to the topic :

1- My question wasn't technically accurate enough ("a small capacitor to an op-amp, ok, but to which pin?"), but you gave every answer i need to know (and try to remember from now on).
I already made a TS/SD1 clone with many mods & tweaks, as well as a RAT (with many mods also), so i cam across the two situations you were describing, Elijah, and it's good to know why a component is here, and why it has a given value.

2- In this case, yes, it is the MadBean Holy Roller. In the meantime, I used a 12pF cap, and everything sounds good

About the Holy Roller, I would have 4 questions, but I don't know if I shouldn't open a dedicated thread for it?

• Should I replace the 10 (or 12, here) pF cap by a 1nF, or leave the 12nF cap as is?
• I want more bass to come in, I guess I should increase the 100nF cap after the FET, or is there anything else to do?
• The circuit calls for a LF353N for the upper opamp. But if I want a more "hifi" sound, should I use a 5534 or a 2134 instead?
• I don't have a  MPF102 here, but I used a 2n5457 and it works fine. But i have other transistors, and I am wondering which transistor would be the best in this situation? Is the 5457 just fine, or should I try some other Jfet here?

[antonis]

• Should I replace the 10 (or 12, here) pF cap by a 1nF, or leave the 12nF cap as is?
• I want more bass to come in, I guess I should increase the 100nF cap after the FET, or is there anything else to do?
• The circuit calls for a LF353N for the upper opamp. But if I want a more "hifi" sound, should I use a 5534 or a 2134 instead?
• I don't have a  MPF102 here, but I used a 2n5457 and it works fine. But i have other transistors, and I am wondering which transistor would be the best in this situation? Is the 5457 just fine, or should I try some other Jfet here?

• Replace it with a cap of value set by 0.159/(4k7*f), where f= LPF cut-off frequecny of your taste
• C2/R3+R4+R9 HPF cut-off frequency is about 11.6Hz.. C1/R2 one is about 16Hz.. C8/R7 one is about 28.4 Hz.. Make your choice..
• a 2134, although it shound't make any difference at all..
• the 5457 is just fine.. after all, it's just a humble voltage follower..

[snk]

Thank you Antonis for your comprehensive answer.

C2/R3+R4+R9 HPF cut-off frequency is about 11.6Hz.. C1/R2 one is about 16Hz.. C8/R7 one is about 28.4 Hz.. Make your choice..

Yes, in the meantime, i made some more hearing tests, and noticed that in fact, the effect has a wide spectrum range : it is one of the compression artifact to lose some of the bass content while compression more.
From the figures you gave above, the HPF lets all teh audible bass pass through.

a 2134, although it shound't make any difference at all..

Oh, really? I'm getting some distortion on long, sustained bass notes (even when there is no compression), and was hoping that swapping the op-amp could improve that (because else, I don't know how to fix this issue).

[antonis]

I'm getting some distortion on long, sustained bass notes (even when there is no compression), and was hoping that swapping the op-amp could improve that (because else, I don't know how to fix this issue).

You have to define the origin of distortion..
e.g. is THD percentage high due to "noisy" op-amp (resistor values are more than acceptably low) or is just IC1A Gain in combination with input signal amplitude..??

[snk]

The distortion was coming from the opamp : even with compression off (well, by that i mean : with the "compression" pot set to minimum, and indicator led kept off), i could hear some distortion on percussive material with sharp transients and bass heavy content. Increasing the compression amount was not increasing the distortion, so I thought it was not an artifact from the vactrol.
I tried with a 5534 (instead of the 353) : the sound is clear again
I found a 5534 in my opamp drawer before finding a 2134, so this is what i used, but maybe that tomorrow I will try with a 2134.

[anotherjim]

We have seen trouble with something like this one before.
I think that C8 on the Vref buffer should be gone!

[Rob Strand]

Regarding the 10pF caps.

Amptramp's is covering the precautionary cases of not having the thing go nuts an oscillate.   Whereas Antonis's post is dealing with deliberate high frequency roll-off.   

For the precautionary cases there is usually a minimum safe cap value.   You can usually tell if it's the precautionary case or deliberate from the cut-off frequency.

For the special compensation  pins Amptramp mentioned you should follow the datasheet.  In general there is a minimum value for a given configuration.  If you have both types of caps present you usually need to choose the compensation cap for unity gain.  That means the cap is larger than some of the configurations you might seen on the web or in the datasheet for the same chip.

[antonis]

Additionally to what said above, ALL gain stages (IC1A, IC2A & IC2B) should have caps in series with respsctive resistors going to VB for rolloff DC gain to unity..

Especially for IC2A, without that cap, LIMIT pot will cracle due to Lug3 - Lug1 DC voltage difference..

[snk]

Thank you all : thread bookmarked for when needed