strange problem after opamp replacement

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
after replacing the nec4570s with ad8620s in the preamp section of my yamaha receiver i am left with a strange problem: when i turn the volume all the way down i get a strange hiss/hum in the rear channels while in 5.1. this hum goes away as soon as the volume is not at minimum.
there are no problems with stereo operation. i have tried replacing one of the three 8620s with another but that was unsuccessful - when i swapped one for an opa2132 the hum went away even with the volume at minimum. i have two questions:

is there a chance that i will damage the receiver while running with mismatched opamps? ie 2 x 8620 + 1 x 2132. if there is not i`ll probably just leave it alone.

any ideas why this hum might be present only in the rear channels and only with the 8620?
 
RF oscillation?

This sounds a bit alarming. My experience with a strange hum when the volume control was turned down has always been due to HF oscillation . This can result in blown power transistors and blown tweeters depending on frequency ,amplitude and duration. I suggest you look at the channel in question with an oscilloscope. You hear the artifacts in the audio range and not the original oscillation frequency.
You may have to test it with the speaker connected ( potentially dangerous for the tweeter) because it might be a load dependent oscillation. If it goes away with a change in the type of opamp I am more inclined to believe this is the reason. If the problem goes away with some different combination of opamps, use that combination. It shouldn't matter much if some sections use a different opamp.
I must add that RF oscillation causes very rapid heating of the power devices. So if your power transistors on that channel are getting pretty hot ( touch it and see) , you most likely have an RF oscillation. Since it stops with other opamps, you already have the solution.
Cheers.
 
ashok,
thank you for your response.
unfortunately i don`t own an oscilloscope (yet) so can only perform very rudimentary tests.

i will stick with the ad8620+2132p combo for now as your suggestion that there shouldn`t be a problem with this is reassuring and reinforces my own (vague) opinions.

i will definitely follow your suggestion to look for overheating.

thanks for your help. it`s a bit of a relief. :)
 
ashok,
just to let you know that there seems to be no problem with overheating - so i`ll just leave it like this for now.

i am wondering - anyway to get rid of this oscillation - i tried a .1uf film cap from -v to +v but it didn`t help. when i have more time i`ll locate a ground and tie the rails to that with some bypass caps. other than that any way to reduce the bandwidth? or am i barking up the wrong tree?

thanks again.
 
Even though there is a standard pinout, opamps are not all the same. Swapping them is not the trivial exercise that some would have you believe.

The surest cure for the problem is to put back the original opamps. Those are the ones for which the circuit was designed.

MR
 
MRehorst said:
Even though there is a standard pinout, opamps are not all the same. Swapping them is not the trivial exercise that some would have you believe.

The surest cure for the problem is to put back the original opamps. Those are the ones for which the circuit was designed.



But then again, what do I know? I'm an engineer whose mind has been poisoned by too many years of education, and too many years of work experience. All my jobs were given to me by desperate companies who couldn't find enough plumbers, nurses, and aircraft mechanics to fill their open design engineering positions...

MR
 
Joe,

a cause for oscillating could be a too low output resistor directly after the opamp (should be 100-150 ohms). I´ve had this problem when replacing OP27 with AD845 (long time ago) and it went away after I put some output resistors in.

hope this helps,

William
 
If you compare the datasheets of 4570 and 8620 you will see significant differencies. The pcb is designed for 4570 but not for a fast to modern opamp. I suspect that the datasheet tells you all about how important pcb layout and decoupling is. One thing that can cause you trouble is the rather large input capacitance, can cause oscillations if you with take actions.

Max voltage +- 13.65 V. What supply voltage do you have?
 
MRehorst said:



But then again, what do I know? I'm an engineer whose mind has been poisoned by too many years of education, and too many years of work experience. All my jobs were given to me by desperate companies who couldn't find enough plumbers, nurses, and aircraft mechanics to fill their open design engineering positions...

MR


mrehorst,
sorry if i reminded you how much you hate your job:)

wuffwaff,
thanks for the suggestion. i have read that a resistor on the output can reduce 'ringing' - is this what you mean?

peranders,
i have compared the datasheets of the two opamps - unfortunately i was unable to pinpoint any differences that may be leading to the problem - mainly due to my own inexperience in reading datasheets. do you have anything specific in mind that i should be looking at? you speak of is taking actions against the relatively high input capacitance - perhaps you`d be willing to explain what you mean a little fuller?
if i`m not mistaken the voltage is around +-12V.

i was a little surprised that i had any problems (although i realize that the 8620 is a very different chip from the 4570) - i have used the 8620 in a few headphone amps and always have found it to be robust and easy to work with (it would have to be to survive my ham-handedness).

i guess i was wondering if you had found one or two things in the spec sheet that i should look at very closely? if at all possible i would like to keep the 8620s as i have always really enjoyed them. and so far they make a considerable improvement over the 4570s. i can`t hear anything that speaks of distress but i am a little worried that something may be going on 'behind the scenes' that wouldn`t be immediately obvious. also, that something bad may happen when i turn things up!

thanks guys.
 
mrehorst,

i always found the yamaha a little 'schreechy'. i always liked the sound of the 8620 in the headphone amps i have built. i was hoping that peanut butter and chocolate might make reese`s peanut butter cups.:) sonically, the change seems to have made a difference -

also, i just like screwing around with stuff.
 
joe dick said:



i was a little surprised that i had any problems (although i realize that the 8620 is a very different chip from the 4570) - i have used the 8620 in a few headphone amps and always have found it to be robust and easy to work with (it would have to be to survive my ham-handedness).

i guess i was wondering if you had found one or two things in the spec sheet that i should look at very closely? if at all possible i would like to keep the 8620s as i have always really enjoyed them. and so far they make a considerable improvement over the 4570s. i can`t hear anything that speaks of distress but i am a little worried that something may be going on 'behind the scenes' that wouldn`t be immediately obvious. also, that something bad may happen when i turn things up!


If you already used the 8620 for a different application, first I would start seeing what was different in your implementation and Yamaha's.

Second, even if I don't know the 4570's specs, they are probably quite different than the 8620's. Slew rate to start with, which makes the AD chip more critical to use.

A place I would look on to improve things, if you did not already, is supply: local cap bypass close to the chip and a small size film cap (around 1uF or so) soldered right between the V+ and V- legs are some suggestions.

Also a 1K resistor before the input and a 100 ohm resistor at the output might make it more manageable for the rest of the Yamaha to assimilate it, which seems to be the problem.

The rest is layout, which is difficult to solve, but the resistors might help you there.

How did you assemble the 8620 in order to put into the 4570's place?


Carlos
 
ashok,

you win some you lose some was pretty much how i looked at it. i always thought that the yamaha was ok - but just ok. if i could make it better - why not try?
i don`t seem to be able to learn from electronics books - for me it`s pretty much a process of breaking things then getting them to work again - and searching around and asking questions on forums like this one.

carlmart,

i think those are all good ideas and i`ll probably try them out one at a time.
i`ve used the 8620 in simple circuits like the cmoy headphone amp and in the more complicated buffered multiloop configs of walter jung - something that was popularized by ppl for use in headphone amps.
yeah, the slew rate on the 8620 is like 5-7x that of the 4570. but i have tried with a (not exhaustive) combination of inner and outer loop configs (changing bandwidth, changing gain) and always found it easy to work with.

unfortunately, at present, the ad8620s are soldered to a socket which is plugged into another socket which is soldered onto the board. i realize that this is problematic in itself but felt i had little choice - what if the 8620 hadn`t worked at all? so, i wanted a setup where i could try a variety of opamps. i guess i got the worst of all possible results - the ad8620 works but i am left with niggling doubts about it`s suitability for the circuit.

but probably you are right. i will have to closely examine the circuit and see if i can modify some of the resisitor values to something that more closely resembles something that i know works. to bad i am so lazy....

something i am worried about - if i am still getting oscillation from the circuit - would i know? would there be some obvious audio clue that i would hear. as it stands the mod sounds really rather good - in fact kinda impressive. nothing seems to be overheating (i`ve been checking this obsessively) and there are no other obvious signs. so, would i know?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.