DC offset and such VS. FET input opamps..

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

Svart

Well-known member
Joined
Jun 4, 2004
Messages
5,134
Location
Atlanta GA USA
In a post that I can't seem to find for the life of me now, there was the normal discussion about recapping and reamping boards but one point stood out in my mind.

Sorry whomever you are, I can't seem to remember who brought up an idea that I was interested in before that thread started. the idea was getting rid of interstage coupling caps. I had tried to just remove them and see/measure what would happen. nothing shocking happened, just a lot of DC offset and clicking/crackling when stages were switched in/out. the DC offset was even noticable on the wave editor program i use to master. viewing a channel with the coupling caps installed showed nothing out of the ordinary, viewing one without the caps showed terrible offset.

mind you, this was with NE5532s in most of the signal path, and around 10 coupling caps.

ok enough background. In this thread a forum member mentioned that FET input opamps do very well with DC offset and other assorted things that BJT input opamps would have trouble with. I thought about it and since my testing sans caps was done with BJT opamps and I had since replaced several channels with high-end(yes that's subjective but i DO like the sound better..) FET input opamps I decided to do another round of testing.

absolutely NO DC offset detected without caps. the only caps in the channels are the phantom blocking caps and a few in the outboard gear.

I know this isn't as technical as most would like to see but removing 10+ caps per channel does add up to a considerable amount.. however the costs in moving a console over to FET input amps is also very high.

I invite conversation over this topic because it's one that i have rarely seen here.



:thumb:
 
> my testing sans caps was done with BJT opamps and I had since replaced several channels with high-end (yes that's subjective but i DO like the sound better..) FET input opamps I decided to do another round of testing. ... absolutely NO DC offset detected

Well, and you were comparing the old-tech 5532 against the new-tech amps.

Back In The Old Days, BJT had several-mV offset, high but well-matched bias current. FET had zero bias current but many-mV offset. Which was better (for DC offset) depended a lot on your resistances.

Then came zener-zap and then electron-beam. It now costs "nothing" to zap an FET input pair down to microvolt offset. If your gain is not huge, your DC offsets may be much-much lower than your signal voltages and thus generally negligible.

If you use 5534 with low-value matched bias resistors (which can be a tough design issue) and use the offset-trim pins on a 5534 (not on 5532), or if anybody sold a pre-trimmed 5534/5532, you could get down to negligible DC offset.

However the resistances may have to be 10K with 1K match between inputs, whereas FET input current is so low that 100K resistances don't screw-up DC offset. While the trim on a BJT can get zero offset at one point, the bias currents drift with temperature and the trim will wander away from zero.
 
Thanks PRR, i knew you'd be around sooner or later.. :green:

I figured as much but I don't think I could have put it as plainly and straight forward as you have!

One thing i did not mention as i think it's a result of the "way" a specific opamp "sounds" was something i noticed on the wave viewer. With the caps in place there were peaks many times(don't know exactly how much though..) higher than much of the rest of the transient. usually just one cycle positive and negative then back to a normal *looking* transient. it was also noticable to the ears. since removing the caps the signal now looks completely normal. I have not compared with the original opamps yet.

any ideas?
 
I have used both OPA627 and AD8610, 8620. I have got offset voltages in the range of 70-90 uV! The AD8610/20 is extremely good for being a JFET opamp.

In the picture below you can see an AD8620 located near the green caps. The picture shows a "Gainclone" with extremely low output offset, 70-80 uV.

qrp02r0_topview.jpg
 
i am also using the opa627, opa2604 and opa404. I like the way the opa404s and the opa627s sound! they have taken the harshness out of the signal that was there before with the 5532s/tl082s and tl084s.

upgrading like this costs a lot but I have been more than happy with the outcome.

I have not measured the selfnoise and induced noise of the unit since i removed the caps, but will do so shortly. Also, a lot of the stages are using these opamps in unity mode as little more than buffers. most of the opamps with gain don't have feedback caps either.

i really need to do sweeps of the frequency spectrum on this console..
 
Back
Top