What you need to look at is the input impedance (the higher the better) of the power amp and the OUTPUT (not input) impedance (the lower the better) of the preamp. The common consent is at least 10 to 1 ratio, i.e. you need the input impedance of an amp to be at least 10 times higher than the output impedance of a preamp.
21 responses Add your response
With a 15kohm input impedance, I would look at an output impedance of 200
ohms or lower, if possible. 15kohm input impedance is on the low side.
Personally, I think the 10 to 1 ratio is too low. When matching for my system, I
now look at a ratio of at least 50 to 1 in order to not have rolled of bass or
Keep in mind that amps and preamps do not have flat impedance curves. A
preamp with a rated output impedance of 200 ohms can easily have impedance
spikes well above 1000 ohms, and two preamps with nominal 200 ohm output
impedance specs can have wildly varying impedance spikes. Therefore, IMO
the value of the impedance spike should be carefully considered...even more so
than the rated output impedance...when matching gear. John Atkinson's
Measurements section of Stereophile equipment reviews is very helpful when
examining a component's impedance curve.
Both Sidssp and Tvad have fairly well hit the nail. FWIW many tube pre-amps have a 600 ohm output impedence (my ARC SP10 does, and recommends a minimum amp input impedence of 20K ohms). As common are some with output impedences of 3500 ohms. (Three other tubed pre-amps I have have output impedences of 3500 ohms.) None of them really sound all that good with amps with less than 47K input impedence and one even likes 100K best.
Tvad's comments about impedence being a moving target is right on, and you need to start by knowing the output impedence of your Bottlehead (I tried to look it up but couldn't find it).
05-05-09: SidsspI'm afraid that I do not quite agree with you. For example, here is one implementation of a resistor attenuator:
The only places where the impedance will be high & is likely to affect the sonics is at the -0.5dB, -3dB, -5dB, -7dB settings. Of these settings the -7dB setting is probably the best as the effective impedance the power amp would see in the 500 Ohms region. This is still high for a 10K input impedance power amp but it just *might* work - on the cusp.
(BTW, a digression: my calculations show the 1M resistor provides -0.008dB attenuation, the 4.99K provides -1.5dB attenuation, the 1.69K provides -4dB attenuation, etc. Am I missing something here? Is he adding some other resistor in series w/ the 1K such as the source output impedance (which is not shown)?)
More often than not, the user is going to be in the -31dB range where the effective resistance of this attenuator will be defined by the resistor to ground - which is a very small value. Thus, for even a 10K input impedance power amp this should not be an issue.
(this is all electrically speaking, which is the focus here. No idea what the in-line attenuator does to the overall sonics - my intent is *not* to discuss that here).
05-05-09: Bob_reynoldsBob, 1stly, who is the question directed to? The forum in general OR to any particular member?
2ndly, not sure what you are asking - you mean what are the ill-effects of adding a SERIES resistor between pre & power?
You wrote "source" which I interpret as CDP/tape-deck/reel-reel/TT, computer, etc but the author was talking between pre & power. So, you mean between pre & power OR source & power amp, as you wrote?
First, start with the source impedance. Then add to it the value of your passive. That is now the source impedance (much higher) at low volumes. A tube amp has a high input impedance- 100K is common.
The result is that the high output impedance of the volume control, with the high input impedance of the amp, and the character of the interconnect cable itself will result in drive difficulties at low volume, because the source impedance will be primarily the value of the control. At higher volumes this effect is reduced. Thus it is to your advantage to not have a whole lot of voltage output (3 volts max) out of your source, so that you can run the volume control near the top! Midway down, and the control is already degrading the signal.
Most CD player manufacturers don't get this simple point, and so put **way** too much output from their players. I've seen them as high as 15 volts- making a passive control useless (if you want it to sound right).
The result is that the high output impedance of the volume control, with the high input impedance of the amp, and the character of the interconnect cable itself will result in drive difficulties at low volume, because the source impedance will be primarily the value of the control.
Atmasphere - the resistor attenuator made by EVS seems to have exactly the opposite problem. I.E. a higher volumes the sound might be degraded. Take a look @ the link I provided in my prev post.
The EVS attenuator has a 1K in series & uses shunt resistors to provide the attenuation. So, for any attenuation setting, isn't the effective resistance of the attenuator the parallel combination of the 1K & whatever attenuation setting resistor?
So, for normal volume settings, I see that the shunt resistors are very small in value thereby making the effective resistance this shunt resistor value.
It appears to me that this EVS resistor attenuator is correctly made & will avoid the lack of drive levels at low listening volumes.
Bombaywalla, it will help a lot, in one way, but now you have a different problem- drivability. There are very few sources that can drive 2.2K without distortion and/or bandwidth loss!
There really is not a good way to execute a passive volume control without running into some sort of problem like this. TVCs work much better, but have their own issues as well, although IMO far less severe. IMO/IME experience if you really want to do it right, an **properly designed** line stage is the only way to go.
Hi Bob, to really get a passive to work you get into trouble when you get over 10K for a value, even with short cables. That's why you are better off having the control inside the amplifier, or else inside the source (both inconvenient), and buffered from the cable. Even with 10K, the effect can still be heard.
There was a lot of talk about buffered PVCs a few years back, but what we are talking about then is an active without gain.
There are some preamp designs that put volume pots at the output. What do you think of them?
I just bought a preamp that uses 100K volume control at the output. It works well with my power amp which has a 500K input impedance. However, it sounds thin with another power amp that has 60K input impedance.
Please comment. Thanks.
Hi Bob, no worries! The higher the value of the control, the greater its artifact (and that of the cable) is going to be. If you have a higher input impedance, like Vett93 above, the problem you get into is high frequency roll-off.
The lower the input impedance of the amplifier, the harder it will be to get bass at low volumes.
Essentially you can regard a passive system as a sort of tone control that also affects the volume. Because TVCs have greater control of their output impedance they can offer greater performance in this regard if other factors are also taken care of.