Odd things with single-rail op-amps

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

Samuel Groner

Well-known member
Joined
Aug 19, 2004
Messages
2,940
Location
Zürich, Switzerland
singleNoninv.gif

So you say this works? Where should the noninverting input get his bias current from? :?

Samuel
 
[quote author="Viitalahde"]Still I don't get it why I can't bias both stages the same way (as in the 1st image). Confusing.[/quote]

In the 1st image, the "-" input is actually biased by resistor R2 from output. That's its DC path to signal ground.

Same way a follower works ("-" input ties directly to output, nothing between "-" input and signal ground).
 
Try getting rid of C1 and attaching R1 to the junction of the two 100K forming the voltage divider (now, your virtual "ground"). The startup problem is probably being caused by trying to charge up the capacitor C1 through the feedback network. This capacitor is producing a sizable offset between the + and - inputs when the supply first comes up. You should be able to use a similar feedback connection arrangement for your second stage using this method, too. Be sure to bypass the capacitor attached to the voltage divider with a small good quality capacitor such as a .1uf monolythic capacitor. This will improve the noise performance of the circuit.
 
Think of your "virtual 0V" as a 0V in series with a 10K resistor (if using two 20K's to divide psu).

Explained differently: Your virtual ground don't like having current dumped into it. It's alright for biasing non-inverting inputs, but dumping part of output current into it (through the divider at inverting input) upsets zero-point.

You may want to buffer your "0V" with a single buffer opamp - that'll make it much more stable up to maybe 20mA, depending on opamp.

Jakob E.
 
Here is a possible solution. I wasn't sure what gain you are trying to get out of the circuit so you will need to calculate the input and feedback resistors for the first stage. Keeping stage gain below 10 (~20db) will give you the lowest transient intermod distortion numbers.

opamp_schema.GIF
 
Please publish a complete schematic of what you are doing. I gather the last one is based on a dual rail configuration, so it's not very helpful to diagnosis!

Also it would be good to know what these "class A" op amps are---input bias currents in particular, tendency to polarity reversal when common-mode range exceeded, etc.

The major problem with single supply stuff is the assumption that a simple bypassed voltage divider behaves the same way as a hard ground in a dual rail config. If you pull hardly any current it's a good assumption. If you start loading it a bunch, look out.
 
[quote author="Viitalahde"]I don't have a complete schematic handy right now. It's the same circuit as above with all the necessary capacitors for ac coupling.

I just tested Burdij's suggestion, which didn't work. Then I got fed up and prototyped a simple gain stage from a 5532 in 30 seconds. Worked right from the beginning. :mad: So, there's something wrong with the amps themselves - I thought I checked the PCB layout a dozen times (heck, a 5-transistor circuit found in that op-amp collection Tekay published last x-mas), but I quess its wrong.

Kind of a relief though. I probably did everything right from the beginning, just had the wrong amps. :razz: I don't suck as bad as I thought.[/quote]

Could you post a link or whatever to those amps? It would be instructive to see what was causing the behavior.
 
Thanks!

Wow that is quite a compendium.

One thing I note immediately is that the circuit you used already has a d.c. bias network for the noninverting input. This explains why you could get away with a naked coupling cap in some of the versions of the circuit, which would not wash with a typical op amp.

I'll give the rest of the bits a ponder in a while.
 
Back
Top