Active vs. Passive preamps & volume control output impedances

RonB · « **on:** 1 Oct 2006, 07:18 am »

Hello to all...while perusing a web site forum board on cable differences...a statement was made about active & passive preamps along with volume control impedance's,so I'm writing to ask what are each of these

1)How do they amount to anything in my audio/video system.
2)When reading equipment spec sheets how do I identify if the specific equipment piece has these items & if they do....what do they mean related to audible performance/results?
3)Are there good vs. bad specs...high vs. low....whats more desired & whats not so desired?

Thank You

Bob Reynolds · « **Reply #1 on:** 2 Oct 2006, 02:49 am »

Whether active or passive, I believe it is desirable for a component to have a high input impedance and a low output impedance. The rule of thumb I have read is that the ratio of output impedance to input impedance should be greater than 10.

Daryl · « **Reply #2 on:** 2 Oct 2006, 03:11 am »

Hi Ron B,

Output impedance is the impact load current has upon output voltage.

The lower the output impedance the less output current will effect output voltage.

The capacitance of interconnects causes considerable load at high frequencies resulting in high frequency rolloff.

The problem is exhaserbated in passive preamps.

Active preamps drive their load with an output buffer amplifier of (hopefully) low output impedance.

A passive preamp has no active circuitry and it's output impedance is determined by the volume potentiometer.

The lower the resistance of the volume potentiometer the lower the output impedance of the passive preamp.

However the lower the value of the volume potentiometer the lower the load impedance presented to the source componet.

Load impedance can be only so low before causing distortion in the output stage of the source componet.

Also the output coupling capacitor in source componets (except those without them) forms a highpass filter along with the load impedance.

Lower load impedance results in low frequency rolloff.

So passive preamps are very sensitive to non-resistive load impedance like cable capacitance and present a heavy load to source componets.

What you are looking for is low output impedance and high input impedance from all componets and low capacitance interconnects.

Daryl

amplifierguru · « **Reply #3 on:** 2 Oct 2006, 04:37 am »

There is a further significant issue here - when it comes to interfacing with the power amplifier, for best performance, the input and feddback impedances should be (approximately at least) matched so that the input stage can function properly.

If the feedback impedance seen by the input stage is ~ 1K ohm, then a 50K pot at the input is a serious imbalance.

Cheers,
Greg

Daryl · « **Reply #4 on:** 2 Oct 2006, 05:08 am »

Quote from: amplifierguru on 2 Oct 2006, 04:37 am

There is a further significant issue here - when it comes to interfacing with the power amplifier, for best performance, the input and feddback impedances should be (approximately at least) matched so that the input stage can function properly.

If the feedback impedance seen by the input stage is ~ 1K ohm, then a 50K pot at the input is a serious imbalance.

Hi Greg,

What feedback impedance?

Daryl

amplifierguru · « **Reply #5 on:** 2 Oct 2006, 05:20 am »

Hi Daryl,

If you look at a schematic for a power amplifier, most feature an input differential stage, simple or elaborate. On one side is the input signal, on the other the feedback signal. The impedance of the feedback network as seen by this input should approximately match that seen on the signal input side, from the source component.

If source impedance is low, say, 47 ohms from a buffered pre-out this can be built up to match the other side's ~ 1K (say) by using a R as part of an input filter.

Cheers,
greg

Daryl · « **Reply #6 on:** 2 Oct 2006, 05:46 am »

Greg,

You always want to drive an amplifier with as low an impedance as possible.

A source componet with a very low output impedance such as an OPA27 or OPA37 buffer stage with a short interconnect would be ideal.

You also will want the impedance of your feedback divider to be as low as possible to minimize the amount of bias current noise in the output and to minimize DC offset due to bias current pulling the voltage of the drive network.

Ideally you would want an instrumentation amplifier input section on an amplifier to make it immune to ground troubles.

Daryl

amplifierguru · « **Reply #7 on:** 2 Oct 2006, 06:07 am »

Hi Daryl,

Yes you have it right - except that it is not necessary to go for very low impedances. I used 1K as a realistic example. Bias current noise has never been an issue in any of my designs, and they are renowned for their low noise.

The point I was making was that, even though the input impedance of an amplifier, commonly defined by the input biassing resistor, may be 20K -> 100K this is not carte blancheto drive it from a 2K - 10K source. The drive impedance should match the feedback impedance.

Cheers,
Greg

BobM · « **Reply #8 on:** 2 Oct 2006, 01:47 pm »

Greg,

This discussion leads me to a problem I encountered this weekend - perhaps you can help me get to the bottom of my problem.

I built a basic voltage divider passive device to turn down my bass amp (I'm trying to biamp my system as an experiment and see how it works for me). I have a 100K Noble pot that I'm using as a variable shunt resistor. I used a 1K series resistor (which is where I'm betting my problem truly lies - I probably need a 10K there to get any adjustability).

The problem is I get no variability in the control now, it's either all on or all off. Perhaps I hooked up the Noble pot incorrectly (one line in from the far side of the series resistor, one line out to ground, nothing on the pot's ground lugs - do I need a line to ground from there?). I think I need a larger series resistor though (my amp has a 100K input impediance)? I'm not too worried about top end rolloff, since it is being used to control the bass amp (20Hz-2500Hz approximately).

I did hook up the Noble pot all by itself, as a variable series resistor, and it works fine, but I think a voltage divider style passive would give me a cleaner signal.

Can you please help point me in the right direction?

Thanks,
Bob

amplifierguru · « **Reply #9 on:** 2 Oct 2006, 07:58 pm »

Hi Bob,

If you're using a 1K resistor in series with the variable Noble pot resistance to ground with the junction off to the amp, you will have no variability until you are right at the extreme of the pot rotation.

Q: Do you require a lot of reduction in gain or only a little? If the answer is only a little, of the order of a coupla dB, keep the 1K and use a 10K pot wired as a variable R (pot wiper wired to the CCW grounded end lug). I assume you have a regular low impedance output preamp, so source impedance from the combo will be < 1K which is ballpark for most amps FB loops, so a fair balance.

If more than a coupla dB , increase the 1K to, say, 2K2 and do again.

Hope that helps.

Cheers,
Greg

BobM · « **Reply #10 on:** 2 Oct 2006, 08:20 pm »

Thanks Greg. You're right, I've only got a very little variability right at the last twitch of the pot.

Optimally I wish I had a lower impediance pot ... but I don't. I just orderd a pair of 10K resistors to try out with the 100K pot. Hopefully this will give me some additional variability. I figure I need somewhere between 5-10db reduction in volume. Hopefully this will get me there. I don't think the hight impediance will matter too much to me sonically, since I'm not going above 2500Hz in frequency on this amp. I just wonder if I'm putting any other load problems on the amp.

Thanks,
Bob

NewBuyer · « **Reply #11 on:** 2 Oct 2006, 09:54 pm »

Quote from: amplifierguru on 2 Oct 2006, 06:07 am

...even though the input impedance of an amplifier, commonly defined by the input biassing resistor, may be 20K -> 100K this is not carte blancheto drive it from a 2K - 10K source. The drive impedance should match the feedback impedance.

(Speaking as a layman here): I'm just wondering, are you referring only to balanced (XLR/TRS) connections in this statement? The reason I ask is that I've not previously seen a "feedback impedance" spec provided by amp manufacturers, although the generic input impedance spec is always provided. I too have always been under the impression that with unbalanced connections (RCA/TS) and solid-state amps, you would want amp input impedance to be at least ten times the source output impedance, in order to ensure bridging impedance and max voltage transfer (not power transfer).

amplifierguru · « **Reply #12 on:** 2 Oct 2006, 10:21 pm »

Hi NewBuyer,

The rule of thumb

Quote

you would want amp input impedance to be at least ten times the source output impedance

is fairly close to the truth.

If you take a typical amplifier (mine say) gain is 28 times achieved by a resistive divider from the output of 27K and 1K. The impedance seen by the feedbach side of the input pair is 27K//1K ~ 1K. So a source of impedance 1K to the input side is about right. The input impedance is the same as the feedback resistor so 1/27 would be right.

Yes amplifier manufacturers do not specify the FB impedance as they generally assume the amp will be driven from a low impedance source and have probably buffered the impedance up to equality on both sides using an input R (also used for high frequency filtering in combination with a C to ground). But many audiophile ARE using passives and volume controls into the power amps and this is where an imbalance can arise, which can impact on the performance of the amp itself.

As an example a typical amp might have 1K in both these positions assuming near zero source impedance from the preamp. If you instead use a CD player feeding a 10K pot, for example, the source impedance will vary between 0 and 2.5K (halfway electrically). however, most people use a log pot which is -20dB at 12 o'clock and that is close to max listening volume. This setting is ~ 1K ohm, so the input buffering R filter can be linked out ( or a small value placed there) and let the pot at typical setting do the balancing R.

However higher value pots and fixed high output impedance preamps are a problem as a match is not possible and performance within the amp could be degraded ( in most). A low impedance buffer is then beneficial.

Trust that makes sense. There are generalisations.

Cheers,
Greg

Daryl · « **Reply #13 on:** 2 Oct 2006, 11:49 pm »

Greg,

Why do you keep saying that the source inpedance at the input should match (or be fairly close to) the impedance of the feedback circuit?

The lower the impedance of the source the better.

Daryl

art · « **Reply #14 on:** 7 Oct 2006, 05:47 pm »

Generalisations that I do not agree 100% with.

Forget the input impedance. Only an issue with passives. Let's stick with actives, first.

Output current drive and impedance will have an effect on sound. Even when driving a 100K load. Some of this has to do what is going on, or should I say back and forth, between the amp and the preamp at frequencies beyond the audio range.

As for getting the right value of source resistance to match the feedback resistance........

If one designs amps with bipolar inputs, then there can be offset problems. But since any amp has circuitry to deal with that, this is a moot point. If it is about noise contribution, the noise floor is established further back in the chain, so again, this is a moot point.

Unless, of course, you are using a passive unit. I do not like or use them, so I will refrain from beating up on those who do use them. It is their prerogative.

However, it is common to find some sort of low-pass filter on the input of an amp. There is no way for the amp designer to take into account how high of a drive impedance the user will have if they insist on using a passive driver. Could result in substantial treble loss, but some like that anyway. More power to them if they do.

Pat

NewBuyer · « **Reply #15 on:** 7 Oct 2006, 06:42 pm »

Pat, I'm curious, would your comments about passive preamps also apply to those units using multi-tapped transformers instead of resistors? Such as the preamp shown in the Promitheus TVC thread for example...

amplifierguru · « **Reply #16 on:** 7 Oct 2006, 07:00 pm »

Hi Daryl,

Quote

Why do you keep saying that the source inpedance at the input should match (or be fairly close to) the impedance of the feedback circuit?

The lower the impedance of the source the better.

Of course it is - because the amp input filter R can then buffer it up to balance with the FB impedance.

Hi Pat,

This is de riguer in op-amp design to optimise the common mode and PSRR in the front end - to get the best from it. The reason it is relatively unheard of in audio is because the amp designer has little control over the equipment it is to be interfaced with, even though virtually every amplifier will benefit, unless it has an input buffer.

The easiest way for an amp manufacturer is to put in a input RC filter for approximate balance and specify a low impedance source. What I was pointing out was that you will not achieve best performance from an amplifier by assuming, because it has say 50K input impedance, then a 5-10K source is good. A low impedance source is, but a limited range of impedance can be optimised for by adjusting the R of the input filter so the total matches (roughly) the FB impedance at the other (differential) input.

Cheers,
Greg

Steve Eddy · « **Reply #17 on:** 7 Oct 2006, 08:37 pm »

Of course if your amplifier doesn't have a global feedback loop all the way back to the input...

se

amplifierguru · « **Reply #18 on:** 7 Oct 2006, 08:45 pm »

Hi Steve,

Wouldn't matter much if you're a flat earther, a differential stage is balanced by equal source impedances at both inputs, irrespective of FB.

cheers,
greg

Steve Eddy · « **Reply #19 on:** 7 Oct 2006, 09:08 pm »

Hey, Greg.

Quote from: amplifierguru on 7 Oct 2006, 08:45 pm

Wouldn't matter much if you're a flat earther, a differential stage is balanced by equal source impedances at both inputs, irrespective of FB.

Yeah, but some differential stages don't care so much about equal source impedances.

se