Passive pre amp vs powered pre amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I use a passive attenuator for controlling the output level of my system but am thinking that there would be significant advantages to adding a simple active pre-amp/ buffer.

An active pre-amp would drive the cable with a near constant impedance regardless of level setting and the volume control would operate in a more ideal way, also the source would see a more constant impedance at it's output.

I think an active pre-amp would remove many of the inconsistencies is sound depend on volume setting. A simple transistor design (not op-amp) is all that's required so signal degradation due to the extra gain stage would be negligible compare to potential benefits.

A stepped attenuator followed by a simple active buffer would be a good solution.
 
I've been using a simple passive volume control at the power amp input for years now, and don't feel the need for anything else. A preamp is just more circuitry to worry about, and I can't see what it actually contributes to the system.

I wouldn't use a potentiometer output to drive a long cable, though. The (comparatively) high output impedance may give HF rolloff problems with the cable capacitance, and make it more susceptible to some kinds of RF interference. Putting it at the amp input removes (or at least controls) these problems.

Cheers
IH
 
The reason an active buffer is good is because it eliminates many of the inconsistencies of a passive attenuator alone. Look at the output impedance of a 10k pot for instance: towards the ends of its travel its impedance tends to 0 Ohm whilst in the middle its 5k Ohm! That's a significant ratio.

Also when you put this in parallel with the input impedance of your power-amp you will find that the characteristic of the pot is far from being the logarithmic ideal.

An effective buffer can be as simple as an emitter-follower with a constant current source load, two transistors is not much extra circuitry to worry about.
 
Passive pots are the most transparent way of controling the volume, so long as a few rules are followed, the 100 to 1 rule i i`ve been told to call it.

1: The input impedance of the passive pot should be 100 times HIGHER than the output impedance of the CD player.

2: The output impedance of the passive pot (at it`s worst possition) should be 100 times LOWER than the imput impedance of the poweramp, if not the input impedance (resistor) of the power amp should be changed. If the input of the poweramp becomes greater than 200k to achieve this you should use low capacitance interconects (100pf or less) per foot, and not more than 1 meter long.

If you follow these steps you will have the the best volume control, some people have said that passives are a bit soft in the bass or rolled off in the highs or lacking dynamics, this is because they did not do the impedance matching that`s explained above.


Cheers George
 
You are pretty spot on, but I must dispute the source loading at 100 times. This would give a typical source load of about 2k. You try loading any output with 2k and you won't get very good results.

Also, increasing the resistance to ease the load results in more noise and problems with cable capacitance.

Passive is the most transparant, but in practice it's almost always impossible to implement properly and you will get better results with a buffer.
 
Ok, so a typical CDP output impedance could be 100 Ohm, multiply by 100 = 10k: this is our pot value. At its worst its output impedance is 2k5 multiply by 100 to get a suitable input impedance for our amp =250K.

Now, 250k seems rather high for the input impedance of a power amp, you might be able to change resistors in the power amp but the maximum input impedance is not solely dictated by resistors but also by the active devices in the input stage itself and so it may not be possible to increase input impedance without a major redesign of the amplifier.

worse still is that the output z of many CDPs is >200 Ohm meaning you would require a power amp input z of 0.5M! unlikely and undesirable.

Not only this but the interconnect is now becoming a dominant factor.

I don't see how this is better than a simple active buffer that would work consistently well with any source, any load and any interconnect.
 
Richard C, most GOOD cd players and d/a convertors have out put impedances of less than 100ohms, if yours is`nt then do`nt do the passive thing,
for instance my current d/a convertor is 30ohms, my last one msb link dac was 70 ohms which i changed down to 10ohms with a better output opamp, it`s up to you to look at all this.

Yesthereis an active component after the input resistor of poweramps be the input of a bipolar transistor or a fet, the bipolar would be in the order of megs and the fet in the order of gigs! this would dictate the input loading resistor will be the determining factor,and the input of the active deviceswill have little effect.

Once you do all the right things you will get a better more transparent sound, no matter how good a buffer or gain stage is it will sound coloured in comparison to nothing in the signal path, and if you do like the sound of the buffer or gain stage all your doing is bandaid fixing a problem elsewhere with a colouration.

I have proved this countless times with members of our audio society selling their million dollar pre-amps for CORRECTLY impemented passive volume controls

P.S. the input impedance of my amps are 1meg, i have unsheilded interconets. I have no noise problems and no rf break through, even living on a main road with taxis pulling up outside
 
georgehifi said:
Richard C, most GOOD cd players and d/a convertors have out put impedances of less than 100ohms, if yours is`nt then do`nt do the passive thing,
for instance my current d/a convertor is 30ohms, my last one msb link dac was 70 ohms which i changed down to 10ohms with a better output opamp, it`s up to you to look at all this.

Yesthereis an active component after the input resistor of poweramps be the input of a bipolar transistor or a fet, the bipolar would be in the order of megs and the fet in the order of gigs! this would dictate the input loading resistor will be the determining factor,and the input of the active deviceswill have little effect.

Once you do all the right things you will get a better more transparent sound, no matter how good a buffer or gain stage is it will sound coloured in comparison to nothing in the signal path, and if you do like the sound of the buffer or gain stage all your doing is bandaid fixing a problem elsewhere with a colouration.

I have proved this countless times with members of our audio society selling their million dollar pre-amps for CORRECTLY impemented passive volume controls

Totally agree with you. Personal experience with implementing a buffer. It definitely changed the sound in a wrong direction. I personally subscribe also to "less is more" in the way of the signal. Especially adding non linear components like transistors. Than to make it more linear you have to ad current source and then really good and well filtered power supply for that stage and then.......

Properly design passive is better. Improperly implemented one, than who cares, replace it with whatever.

As far as pot value, 100k, if the input impedance of the amp is 33k, the total imp. that the interconect will see is 25K. You can use a lin. pot (better match between L and R ch) that will be close to a log pot (because the 33k in paralel) and your sorce out impedance have to be 300 Ohms. If that's hard to do start asembling transistor preamp. Even if my Source out is 2K I'd still use passive preamp and low cap. interconnects.
 

PRR

Member
Joined 2003
Paid Member
If the wire from Volume-Pot to power amp is very short, and you do not need gain, there isn't any gread difference. After all, what is the difference between the preamp amplifier input and the power amplifier input? Mostly cable length: 3 feet (or 1 meter) for external power amp, 3 inches (or 0.1 meter) for a preamp circuit mounted right behind the volume pot.

Taking the 10K pot, max output Z about 2.5K, if you hang 10,000pFd on it the response will droop 3dB at about 4KHz, pretty awful. But 10,000pFd is about 300 feet of cable! Use 3 feet, 100pFd, -3dB response is 400KHz, -1dB at 200KHz, -0.5dB at 100KHz, better than many active preamps, not a real issue.

However if you need 30 feet of cable to reach your amplifier, you are -1dB at 20KHz, not great. This will be at worst-case, Volume pot set 6 dB below full-up, and response will be better at other settings. (But with a no-gain "preamp", the -6dB position is likely to be your "pretty loud" position and shouldn't also be your worst treble droop.)

If you need 100K volume pot to avoid loading a source (or to reduce its distortion), then at just 3 feet you are into significant droop (-1dB at 20KHz worst-case) and you should be looking for some shorter low-C cables and jam the amp against the Volume-pot box to reach. (But beware of hum from amplifier inducting into volume-pot!)

The 100X rule is generous. It is always nice to recover 99% of signal voltage, but often not essential (and rarely practical!). A 100K pot with a 25K load (power amp input) will work fine, except if it has dB markings the "-6dB" marking will really be -12dB. The "error" is less as you move away from the "-6dB" mark, reducing to zero error at full-up and full-down. Unless you are a numbers-head who goes by dial-markers instead of by loudness, this is probably no problem at all. (However, in my recording work, I calibrate my knobs because I do think in numbers as a guide to gain structure and to rough-out a mix.)

Since nearly all modern sources (and certainly all home hi-fi CD players) can drive high level into 10K, the decision is fairly simple. If you drive long lines to a power amp, over 20 feet (6 meters), you almost certainly need a buffer at the volume-pot end of the line to push the cable around. If you run very short lines, under 10 feet (3 meters), you can almost certainly be happy with a passive 10K volume pot. If the particular pot you favor only comes in 50K or 100K impedance, you better keep your cables well under 3 feet/1 meter, or use a buffer between the hi-Z pot and the cable.

The above applies about the same to stepped attenuators. The simplest ones are just pots, but with switch contacts instead of a wiper. For double the price you can get a ladder attenuator: the output Z tends to be more constant but not necessarily lower (can be higher). As a rough guide, treat a switched attenuator like a pot but keep the wires a little shorter. The makers of these high-price items should be able to give specific guidelines for how much cable they can drive sweetly.

There are some tapped-transformer attenuators. I think Lundall lists the trannies, not wired to a switch. VERY expensive, and not wide-range at high gain-resolution, but the output Z can be lower than the source impedance at all positions below full-up. Of course even these very good transformers impart a "sound", and possibly more "sound" than a resistive pot/switch attenuator.
 
georgehifi said:
Richard C, most GOOD cd players and d/a convertors have out put impedances of less than 100ohms


I've got the schematic for a Marantz CD6000 in front of me and that has around 250 Ohms of resistance in series with the output and this is not uncommon.

However, if people are prepared to restrict their choice of source component to only those that work ok with passive attenuators then there isn't a problem.

Yes there is an active component after the input resistor of poweramps be the input of a bipolar transistor or a fet, the bipolar would be in the order of megs and the fet in the order of gigs! this would dictate the input loading resistor will be the determining factor,and the input of the active deviceswill have little effect.

This depends entirely on the particular amplifier topology, changing resistors will not work on all amplifiers, and who wants to modify their amplifier for the sake of an attenuator.

no matter how good a buffer or gain stage is it will sound coloured in comparison to nothing in the signal path.

A passive attenuator is not "nothing".

There seems to be an assumption that just because something is passive it is intrinsically better, quieter, more linear or less coloured than an active device.

, and if you do like the sound of the buffer or gain stage all your doing is bandaid fixing a problem elsewhere with a colouration.

A simple emitter follower with a constant current load will be contribute very little to the sound. And yes it is fixing a much bigger problem elsewhere: the problem of output impedance changing by orders of maginitude over the travel of the pot that an un-buffered attenuator suffers from.

I have proved this countless times with members of our audio society selling their million dollar pre-amps for CORRECTLY impemented passive volume controls

I don't advocate expensive parts or complexity, just a simple impedance converter.

P.S. the input impedance of my amps are 1meg, i have unsheilded interconets. I have no noise problems and no rf break through, even living on a main road with taxis pulling up outside

I haven't mentioned rf breakthrough but unless your hearing extends to rf how do you know it isn't there?

Clearly, the biggest problem with a passive attenuator is that it's performance is largely depedent on the characteristics of the source, interconnect and amplifier. That's a lot of variables, variables that can all be vastly reduced by an active buffer.
 
Sorry Richard C, but if you have`nt heard it yet it` looks as though you never will, and as for a cd6000, if it has 250ohms output impedance you really should look at changing the output opamp as it needs to be better if they used 250ohm output resistor.
And Just to let you know the CD6000 standard or limited edition model is a midfi product.
Marantz comissioned Ken Ishiwata to do a CD6000KI model which costs a lot more and sounds much better, and guess what the differences are, yes a completely new output stage to lower the output impedance, so it can drive better.

(the only 2 things a preamp is good for is for analog lp amplification, or a boat anchor if it`s one of the heavy expensive ones)
Cheers George
 
georgehifi said:
Sorry Richard C, but if you have`nt heard it yet it` looks as though you never will, and as for a cd6000, if it has 250ohms output impedance you really should look at changing the output opamp as it needs to be better if they used 250ohm output resistor.
And Just to let you know the CD6000 standard or limited edition model is a midfi product.

It's not the op-amp that dominates the o/p z of the CD6000 its the series resistance (does the KI model not have this? according to this schematic it does), and I don't own one or believe it to be high-end I merely use it as a popular example.

As for op-amps in the output stage of CDPs: my CDP doesn't use any. As many people on this forum have discovered a simple circuit implemented with a hand-full of transistors makes a much better job of it.


:)
 

Also when you put this in parallel with the input impedance of your power-amp you will find that the characteristic of the pot is far from being the logarithmic ideal.

Well, numbers are typically 2.5K worst-case for the pot and 50K for the input, meaning there's a 5% or < 0.5dB absolute error in the volume - but this is on a control you twiddle by hand anyway. Specifically, this issue doesn't affect frequency response or L/R balance.

I'll agree that 2.5K is too high for driving a substantial length of cable - 1nF cable capacitance has an impedance of 8K at 20KHz. However, 10-20cm cable between the pot and the power amp input will be orders of magnitude below this and (measurably) won't affect the frequency response.

Cheers
IH
 
I think my Naim cd 3.5 has an output impedance of 10 Ohms, i say think because i can only find the spec's of the previouse model the "cd 3"

My Naim nap 250 power amp has an input impedance of 18K Ohms.

I have a 20K DACT stepped attenuator, would this work?

What spec's should i look for in a cable?

Do i risk blowing my amp up?
I only ask this because Naim amps have to have speaker cables at least 3.5 meters long and using the wrong type can ruin the amp.
 
Fat Marley if your Naim cd has an output impedance of 10 omhs that`s great for use with a 10K passive, but your input impedance of ypur poweramp is too low it needs to be >100k if your not comfortable reading a circuit diagram to see if it`s safe to do then get a qualified technician to do the job. (should only take 15 min or so) and you`ll hear sounds that you never think possible from your Naim cd player.

Cheers George

PS if you must use the 20k dact then the input of the amp should be higher 200k or more.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.