Why aren't my opamps oscillating?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
So, I just soldered my pairs of opa627's into brown dog adaptors and plugged them into sockets I fitted in my CD63 and wowww! They sound great.
Thing is though, I was expecting these to be buzzing their tits off because I've not added any bypass caps yet. A lot of stuff I've read here suggests the 627s NEED bypassing, particularly when sockets are used.
I can't hear any distortion, and the opamps are not getting hot, so am I right to assume they are free from oscillation?

Would bypassing improve the sound even if not needed for stability?:confused:

Cheers,

Float
 
Ok here goes

AFAIK oscillation is caused by the ripple on the power supply rails.
This ripple would then be superimposed on to the input signal, thus resulting in an amplified oscilating output.

Sorry if it doesnt make sence, but it does to me :p

Ok back to the power supply.
If the powersupply is well designed it may have very little oscilation on the rails to begin with, therefore the p-p ripple maybe within spec for the opamps, negating the need for bypass caps.

As we are very lucky and live in the UK, we also have a very clean power supply, supplied from the national grid, another advantage of 240v over 110v.

In other countries (the rest of the world :p) mains power can be found to be very noisey, the noise is then added to the components psu, more noise = possibley more ripple, etc etc etc.

Im probably going to get flammed a little for these statements, but the above is very true in respects to instument amplfiers I.e OPA627.
My background may not be audio (but it is signal conditioning), so i have enough knowldge to make the above statements, with confidence.

However ourpower supply is not always that clean, in peak times, i.e. cold winter nights, the demand is obviously greater, and this is where things start to get difficult.
Well thats another thread for another day.


Adding bypass caps may furhter increase the quality of sound by further reducing the ripple from the supply , you will never know unless you try :)
As i said my backgroud is not audio, i just go with the flow around here.

Do a quick search for OPA627 and you will see thread upon thread of discussion, and some heated debates about bypassing in everyone of them.

Hope this helped

Please be gentle with the flaming :cannotbe:

n00beR
 
Ex-Moderator
Joined 2002
There are decoupling caps in CD63.

Please check C611, C612, C613 and C614. They are 220 uF/16V and can be replaced by low ESR types like Panasonic FC. If there would not be any caps at all things would oscillate for sure.

I recommend to solder a 100 nF MKT cap directly from the + supply pin to the - supply pin with each OPA627.
 
Ex-Moderator
Joined 2002
Oscillation is caused by too much gain at a particular frequency causing problems in a feedback loop. It may not cause heating, and with fast opamps like the 627 may be way beyond audible, in the MHz range. The only way to ensure that your circuit is not oscillating is to put a scope on it.

The initial signal that causes feedback could be from the mains, or RF, or ground problems, so eliminating these issues helps, but the only certain way to solve the problem is to reduce the gain at the appropriate frequency(s)
 
Thanks all.

nOObeR, I never thought of the UK as having good mains, what with all the whinging and the plethora of treatments availible, but I guess other countries have it much worse.

My player does have the standared 220uF (stargets?-its a KI-S) in positions C611-614, but these are variable ( and fairly large) distances from the opamps as you know.
I have some MMK polyester 100nanos which should just bridge across the pins, so I'll try these later.

I have read some of the gentle discussions:D about decoupling, and don't want to start these up again, but is there a point when enough capacitance on the pins becomes too much?
I have some 100uF panasonic FCs that I could try and fit to the opamp pins- but this would mean having to take ground jumpers from nearby components, and would be quite messy from a wiring perspective.
Would there be any real gains - Burr Brown seem to think 1uf is adequate here.
:xeye:

All the best
 
OK, heres an update

I soldered 100nF polyester caps directly acroos the +Vs and-Vs pins of the opamps and listened. I think it may have changed the sound for the worse.
Is this possible?
It could just be a bad mains time or even just imagination.
Has anyone had bad experiences with decoupling or do you think it always works for the better?

I also tried resistor biasing the 627s into class A with 1.5kohms and this did seem to help, but I still think this combo sounds worse than the opamps left bare.

These additions were quite fiddly, and I'm trying to minimise the heating/unsoldering/resoldering on the opamps so I'd welcome opinion before I strip it down again.

Thanks in advance
 
Ex-Moderator
Joined 2002
What brand and type polyester ( MKT ) caps ? Short leads etc ? A picture says more than a thousand words.

Maybe overheating had its impact on the devices because I never experienced worse sound after adding the caps.

The opamps won't like soldering and desoldering on their pins too often. This in combination with the price of OPA627 make me advise you to leave it just like it is. I would change the Stargets.

The "biasing in class A" which is a wrong expression for drawing some current is personal taste. Till now I did not like it enough for implementing it myself.
 
Ex-Moderator
Joined 2002
float said:
I think it may have changed the sound for the worse.
Is this possible

If the opamp was on the edge of oscillation, it might have been ringing on transients. This would produce more high frequency distortion, and could have given an impression of brighter, clearer sound.

When you addded the cap this may have been reduced, leading to a more accurate, but subjectively duller sound.
 
I used EVOX MMK poly caps, these are small box caps; leads are no more than 5mm long and they sit right on top of the opamp.
I was careful to apply as little heat as possible to the pins so I'm assuming no damage has occurred.
The onboard decoupling caps are actually 220uF cerafine, not starget- are these any better?

Lucpes- are these silver-micas not too small in value to be effective? I do have some 10000pF polystyrene 160volt caps that may also be usable(if a bit big).

pinkmouse said:


If the opamp was on the edge of oscillation, it might have been ringing on transients. This would produce more high frequency distortion, and could have given an impression of brighter, clearer sound.

When you addded the cap this may have been reduced, leading to a more accurate, but subjectively duller sound.


Now, this is possible. Swapping in the opa627s in place of opa2134s was a total upgrade. Clarity, air,detail, soundstaging etc. all improved noticably. Tonal balance was very similar. If this was distortion then I like it
:D
However, as I was expecting oscillation, I was also listening for signs at high freqs, and heard none.

Maybe its just that aural memory is unreliable. I am trying to honestly assess my "improvements", and not fall into the "I modded it so it must be better" mentality.
I think I will refit the 2134s and compare again, and if the difference seems less than before then I will have to disconnect the bypasses just to see if they did have an effect or not. Hopefully I'll get round to this tomorrow:(

Thanks again for your help
 
Update:

Well, I swapped back and forth between the 627s and the 2134s with a couple of test tracks and -lo and behold- the 627s are better again. The difference was not subtle. The system (or my ears) has lost the messy closed in sound it had the day I added the bypass caps.
I guess the caps just needed to burn in ;)

I haven't decided whether or not to disconnect or adjust the biasing resistors.

Next step is the clock. I ordered an XO module from Guido last week.
Hurry up Postie!!:D
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.