Biasing battery-powered opamps

I suppose a 1 transistor Cap multiplier would work.

polarity

interesting:

drain

the

provides

I

=20

free=20

It is the way the big corporations want it.

message=20

inputs,

gain

chance=20

knows=20

True, but low noise op-amps are generally not race horses nor = particularly prone to misbehavior.

Very interesting. It'd be great if you had the equipment to take actual measurements of the noise levels. I didn't know about using switching transistors as low-noise amps. Quite logical now that you've brought it up. I did also consider building the front end from traditional low-noise transistors like the BC549, but the simplified process of using opamp ICs is tempting, which, after all, is the whole point of using ICs.

I remember an occasion years ago on which I repaired the mixer console in a recording studio owned by a friend. Two mic input ICs had gone bad and, not having the original type (I forget which type), I replaced them with TL072s. The friend was glad to have his gear functioning again so quickly and didn't notice the slight but definite increase of noise, but _I_ did, and I'm no audiophile.

Maybe! :-)

(I've been mucking around with PIC's programmed in C. I hate C)

That is not in the OP's spec and as you already pointed out, and I agree, if he is using a dynamic mic he is better of using a matching transformer.

You already gave the solution and we agree!

Yeah, but I suggested using them after the first (low noise) gain stage

- and I also suggested, if they were to be used at all, that they needed to be used with care.

Your quote (about the NE5332 and TL-071 and BTY we agree about this) was: "** Either of the above op-amps is more than good enough for professional audio use, with any signal that has gone past the mic preamp stage."

The TL-061 is noisier than the 071 by 9db, so there still may be places it could be used especically if low power is important.

Yes.

Cheers! David

For the OP:

(technicians and engineers do it better anyway - they notice what is there and don't invent what isn't)

You were talking about a spec that was a bit what better than PA quality, yes? (I would take PA quality to be THD 1% and noise @ -46db or better)

On the plus side, all these OP amps are duals and they all have the same pin out so use a socket when prototyping and see if they give an acceptable result.

You mentioned elsewhere going to 2 x 9 volt batteries giving you a split supply. I like that idea. It gives you a longer running time without worrying about if the op amps can run properly on the voltage available.

"David Eather is Off with the Fairies"

** It is the actual application.

" Ah, but this *is* for the input stage of a mic preamp."

.... Phil

Just some info on how much noise is (not) increased with various h*mo/heterogeneous op-amp configurations. As I see it there are three better choices using a mixed configuration gaining between 50% - 90% without appreciably degrading performance. Also even the worst case distortion will be un-noticeable compared to the distortion introduced by the microphone.

It was a fun exercise in using OneNote, Word and Viso - I didn't get it quite the way I wanted.

It's on both these file share sites. (some people can't use file dropper)

Amp with Standard Components.mht

Cheers

"David Eather is TOTALLY OFF with the fairies "

** The link has no diagrams visible.

The NE5533 is obsolete and no dealers (eg Digikey, Mouser, Farnell) have any.

The 3 op-amp " instrumentation amp " circuit using NE5532s is about 12dB noisier than a good, differential mic input circuit. Like this one:

.... Phil

Sure, a good design using discrete components will beat a fairly quiet opamp. I too mentioned the option of using a differential BC549 input stage earlier.

I didn't read through the whole article, just had a quick glance. I did look at the footnote and unless I missed something, there's a fallacy there.

In calculating the S/N ratio at the output, the writer uses the ratio of the UNamplified signal to the amplified noise. The fallacy is that the signal and the input noise are amplified at the same time, and the correct S/N ratio is that of the amplified signal to the amplified noise OR the unamplified signal to the equivalent input noise.

If this weren't the case, there would hardly be any microphone amplifier with a S/N ratio that's acceptable for anything but the crudest application. For example, if the input signal is 1mV and equivalent input noise is 1uV (as given in the said footnote) and the gain is 1000, then the S/N ratio will be 1:1 or 0 db, which is absurd. Any further amplifying stage will increase the relative noise level, actually making it higher than the signal.

If you don't really need differential input, don't use it. It adds noise and parts.

For low-z signals, it's hard to beat a good opamp, like LT1028. Noise is below 1 nv/rthz. There are also some fully differential instrumentation amps with noise down there.

The BF862 jfet is around 0.8. And at audio frequencies, you can parallel a bunch of them. Four will get you to 0.4 nV/rthz.

John

I do need a differential input for this design as it's to be used mostly with balanced sources.

It adds

Yup, especially to the parts count. It's not the cost - a discrete assembly can cost the same as or less than a good opamp, but I usually like to aim for a minimum of clutter, both electrically and physically.

I'm familiar with the LT1028 but had completely forgotten about it. Thanks for the heads-up, but it doesn't seem to be the best solution. Pros: Very low noise and excellent linearity. Low voltage operation - noise levels given down to +/-3.5V, gain to +/-2.5V Cons: High supply current. High component count and high cost. For a balanced-to-unbalanced preamp, three LT1028s and six resistors are needed to do the function of one INA217. A single LT1028 costs 50% more than an INA217 at my source.

Haven't really investigated this.

...and then you said you wanted an all op-amp solution using only common op amps.

The results were presented as a ratio between otherwise similar designs and the only point I was illustrating was noise was predominately determined by the noise of the first gain stage. Had I made that mistake the answer would not have been affected because it was a ration between two designs whatever was multiplied by what cancels out when converting to a ratio. A/B is exactly the same as(A * 10)/(B*10) and I didn't even make that mistake!

No I don't.

The

I did not do that at all. In my calculations I specifically used the amplified signal to the amplified noise as it applied over two stages. I am not happy about you misrepresenting me.

There is no such thing in any footnote of mine.

and

Since that is based on something I didn't say *DON'T* blame me for it. These mistakes are all yours.

"David Eather"

"David Eather is TOTALLY OFF with the fairies"

** FFS - work out who wrote what and don't answer the WRONG post.

"Pimpom"

** The you need to correct that error - f****it. ** That bit has been added recently by the guy who owns the web site.

I suppose you never noticed who the author was.

.... Phil

Oops My Bad.

Just checked it.

The top of the post was

The top line. The post I answered was from Pimpom.

Top posting 'cause I feel like smacking someone

Now, Pimpom,

You made the spec that you had an 8 op amp solution and you wanted some ideas on what op amps to use in a battery powered design. Right?

You limited those sections to old established designs that were widely available. Right?

Your suggestion was to use NE5534 all through. Someone suggested this was a high current load and I suggested a mixed op amp strategy using low power op amps for later gain stages. Right?

It was claimed that this would not work because it was too noisy (i.e. bad signal to noise ratio although no one use the s/n term). Right?

I then posted my diagrams and calculations for all to see. Right?

The SECOND paragraph says this: "Consensus was reached that dynamic microphone amplifier built with NE5533 would be low noise and suitable for professional applications. This is fortunate, considering that this exact configuration has been built, sold and used in that task for decades. What was disagreed upon was if a suitable dynamic amplifier could be built using noisier components and this is the question examined here." Right?

Notice the last to lines. The question I answered was how much noisier some other configurations were.

Now if you care to indulge in some math rather than in speculation, you will see that design 2, 3, 4 are 2dB, 1.5dB and 0.9dB (respectively) noisier then the reference designed and this is true *REGARDLESS* of the signal input. No matter the size of input or if the reference design is super good or super bad design 2, 3, 4 are 2dB, 1.5dB and 0.9dB (respectively) noisier then the reference design.

I did not answer how they responded to a specific input level because I answered the question for *every* input level. Had I been stupid enough to calculate S/n ratios I might have got something like (for a 1 mV signal) a reference design with 68dB s/n and designs 2, 3, 4, 66dB,

66.5dB, and 67.1 dB s/n respectively. If I wanted to continue being stupid I could calculate s/n with 0.1mV input and get 48dB, 46dB, 46.5dB, and 47.1 dB s/n. Or for 10mV input, 88dB, 86dB, 86.5dB, and 87.1 dB s/n. Depending on my level of stupidity I could keep going on but I might be smart enough to realise that design 2, 3, 4 are 2dB, 1.5dB and 0.9dB (respectively) noisier then the reference design. Which is exactly what I calculated in the first place!

If your design made with NE5534 all the way through works well enough (and it should it has been done many times before) then replacing some of the NE5534's with lower power devices would result in a degradation of the s/n ratio (in the dynamic mic preamp section) of 2dB, 1.5dB or

0.9dB for design 2, 3, or 4 respectively for all input levels - which is an *inaudible* change but which increases battery life by 50% - 90%.

Don't call my work flawed because you didn't bother to understand it. I answered the question emphatically. You didn't understand it. Right?