[Q] OPA2604 bypass caps

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Enchilada

Well-known member
Joined
Aug 24, 2011
Messages
323
Location
Australia
Hi guys,

I've recently replaced eight channels of opamps in my Tascam M-3700 from JRC 4580D to OPA2604 on the input A PCB. I've read that to prevent to 2604 chips from oscillating, I would need to connect a 0.1uF bypass cap to pins 4 and 8 and then to a common ground.

There appears to me to be 0.01uF bypass caps on pins 4 and 8 already (see the bottom right corner of the linked schematic). Why then should I add another 0.1uF capacitor as from my limited understanding this would total the capacitance to 0.11uF. Is the difference between 0.1uF and 0.11uF going to be that big of a deal?

Here's the schematic:
http://img11.imageshack.us/img11/4559/tascaminputa.jpg

Thanks,

Kris
 
Enchilada said:
There appears to me to be 0.01uF bypass caps on pins 4 and 8 already (see the bottom right corner of the linked schematic). Why then should I add another 0.1uF capacitor as from my limited understanding this would total the capacitance to 0.11uF.

The 0.1uF is just the habitual value to put in these bypass caps. You don't need to change anything since you already have those 10nF caps in.

Enchilada said:
Is the difference between 0.1uF and 0.11uF going to be that big of a deal?

These caps are used to filter PSU crud right at the source for the opamps. I suppose the guys at tascam might have actually measured that they need these 10nF caps to remove some higher frequency crud and that 100nF would have been unnecessary, or even work less optimally.

Often designers just put those 100nF caps as a habit, instead of truly needing them.
 
I've been told by people who have replaced these chips themselves that if I just leave the circuit as is and replace the 4580s with 2604s then they will oscillate up in the MHz region. They also said something about blowing up speakers  :eek:. The general consensus  appears to be to add the 0.1uF caps. This was the reason I initially thought to just replace the 0.01 caps with 0.1 and save having to add extra caps.

I don't have an oscilloscope to test my console with either.
 
I see. If there's a proven case of oscillation then replacing the 10nF caps might not be enough, but it really depends how far (physically) they are from the opamps. Combating oscillation you really should solder the extra 0.1uF caps directly on the opamp (slot) legs. That way you can at least make sure the rail resistance, or perhaps parasitic capacitance will not have any effect on the opamps.
 
Thanks Kingston.

I have one more question.

I replaced C111 470uF 6.3V (at the top middle of the schematic where it says "MIC") with a 470uF 10V Panasonic FM electrolytic. I connected the "-" of the capacitor to the 5k potentiometer and the "+" side to R121. Is that correct? Although I am tired of soldering, I don't want to pop a cap in each channel  ::)
 
The 470uF 10V standard electrolytic is not going to blow up in that position, no need to look for bipolars just yet. The cap is there to prevent DC voltage going to the pot which would make it scratchy when turning. Fire up the channel and listen to it when moving the pot. If it makes scratchy noises then find some bipolars, or put a 1 - 0.1uF film cap in parallel with the standard electrolytic to make a pseudo bipolar.
 
I have just seen the schematic , I have not done any calculation but the problem could given by frequency compensation missed for the opamp changing, so the circuit needs to be compensated again, but this way it's a bit more complicated than adding some caps to the power supply pins.
However, it's possible to do, is a thing than an electronic engineer could do. 
 
IMHO to recompensate it C110 needs to be changed, but could be necessary adding two caps too and eliminate it.
However, probably with 2.2nF the preamp could be recompensated (I have not done any calculation, I have just followed nose)
 
So changing the 1nF to 2.2nF might eliminate the need for adding 100nF caps on the supply pins?

What calculations would need to be done to determine the exact correct value to re-compensate the preamp?
 

Latest posts

Back
Top