Help in choosing a potentiometer as a "Passive preamplifier"

Help in choosing a potentiometer as a "Passive preamplifier"
============================================================

This is more of a guide as I already found the answers I need. To begin a potentiometer as passive preamplifier does not amplify anything. It is just a potentiometer used as a potential divider in a box between the source and the amplfier. It is really just an attenuator but becuase of its position and the box it is called a passive preamplifier.

You might think, well I have a single high voltage source with 2.1 Volts, no interest in microphones, vinyl, balance or (dread) tone controls - so I think I should just use a 10K logarithmic potentiometer to get a nice pure sound, because after all - less stuff in the signal path must be better than more stuff.

But if you try it, you may not get the sound you expect - it might exceed your expectations but could possibly sound terrible. If you ask about it, you are met with claims of high frequency attenuation, lack of dynamics, lack of top end, lack of bottom end, lack of sleep!, balance problems, changes in the sound as the volume is changed, too low volume at max and that by violating a 5 times or 10 times voodoo source and load impedance rule can make the sound gods angry and get you punished with bad sound.

Before I dispel some of these claims and explain the validity of the rest let's get some terms out of the way and see what a preamplifier actually does.

When you see the word impedance it is usually accompanied by one of the words input, output, load or source. In the case of a preamplifier the input impedance refers to the inputs where you connect your sources, for example your CD player. The output impedance refers to where you connect your amplifier.

Great now what is the load and what is the source?

Looking at the chain left to right CD > Preamplifier > Amplifier, you can say that the source impedance is the output impedance of the box to the left and the load impedance is the input impedance of the box to the right. But you have to qualify this with whose source or whose load you are actually referring to.

So the source impedance of the preamplifier is the output impedance of the CD player and the source impedance of the amplifier is the output impedance of the preamplifier.

A preamplifier serves several purposes:
1 It amplifies low voltage sources such as microphones or record players
2 It fixes that audio curve thing for record players
3 It provides a balance and (dread) tone control
4 It provides low output impedance
5 It provides a high input impedance
6 And it looks nice and impressive, gives a warm and fuzzy feeling inside, has a headphone output and a remote control too if you are lucky then some blue LEDs....

1 Low voltage sources must be amplified before sending the signal to the amp, otherwise they may not be loud enough. An amplifier has an input sensitivity value in Volts. A signal with this voltage will be amplified to the limits of the amplifier i.e. the rated power in Watts.
2 Look elsewhere for an explanation
3 I try to sit in the middle anyway and my sources are usually have a good balance,
4 The output impedance refers to the ability to drive a heavy load. A heavy load is one with a small impedance. The smaller this figure the better. More on this later.
5 The input impedance refers to the load imposed on your source by the preamplifier. The impedance should be a high as possible (= a light load) so it does stress on your source too much.
6 Look elsewhere for a warm and fuzzy feeling inside.


High frequency attenuation
Capacitance in the interconnect cable combined with the output resistance of the potentiometer make a low pass filter. There is a good explanation on the DACT website in the technical section two PDFs and also an Excel file to calculate it. This can be a problem if the output resistance of the potentiometer is high and or if your interconnect cables have a high capacitance and or are very long. If you choose a 100K potentiometer and 10 meters of high capacitance interconnect cable then you will get some high frequency attenuation and hear it too. If you have a 10K potentiometer and a short run of interconnect cable then you won't. This low pass filter problem does not affect preamplifiers as they have a very low usually 50-100 Ohm output impedance. With preamplifiers you can use meters and meters of interconnect cable and not fear this high frequency attenuation low pass filter problem. My L-25D amplifier has a 1 nano Farad capacitor connected to signal and ground as part of an optional RF Filter and the capacitance of the RCA interconnect cable is probably around 300 pico Farads. Plugging these values into the DACT Excel file shows that capacitance is not a problem. Be sure to read the PDF which describes how to use the DACT excel file.

Apparent lack of dynamics

Well if you are experiencing the attenuation above that would go some way to explaining the lack of dynamics. The rest of the way is found by calculating the load imposed on the source by your potentiometer and the input impedance of your amplifier in parallel. Let's look at how a potentiometer is wired and do the maths.

Figure01 shows how a potentionmeter is wired to make a passive preamplifier. The ground is the outside of the RCA interconnect cable and is usually connected to one of the outside pins on the potentiometer. The inside conductor in the RCA interconnect cable is the signal and is usaully connected to the other outside pin of the potentiometer. The signal for the amplfier is usually taken from the middle pin of the potentiometer. This is the wiper part of the potentiometer and its position determines the values of R1 and R2 in Figure01. The formula on Figure01 is used to calculate the output voltage from the potentiometer. This formula is used to calculate the volume curves seen in the figures of graphs seen later. Figure08 is a LTSpice file showing the simulated circuit. R1 and R2 are represented here as variables with a step to simulate the rotation of the potentiometer. Simulate the circuit using a 300K RLoad and you see that the VOut becomes very linear.

Figure03 shows that the maximum impedance seen by the source (the load imposed by the potentiometer and amplifier) is the impedance of the 10000 Ohm potentiometer. This occurs at the lowest volume setting -infinite dB.

Figure02 shows that the minimum impedance seen the by source is the impedance of the potentiometer and the amplifier in parallel which is calculated 1 / (1/Resistance of potentiometer+ 1/Impedance of amp) = 3230 Ohms. This occurs at the highest volume setting -0 dB. This is a rather heavy load for your source to deal with.

Well not all sources are equal. My DAC has a bunch of NE5532 opamp chips in its output stage. These chips are rated to drive audio loads all the way down to 500 Ohms. One review tortured the output stage with a 600 Ohm load and it performance was barely affected. Your mileage may vary - it is unfortunate to say and I hate to say it depends, but it does. Some sources have highly integrated chips that also incorporate the output stage opamps and a bunch of other stuff into one chip. These are the bad guys that you want to avoid if you venture into the world of passive preamplifiers. They are not designed to drive heavy loads and likely will cause you problems.

So now you should be thinking: high Ohm potentiometer = potential low pass filter problems, low ohm potentiometer = too heavy a load for the source.

Balance problems
These are caused by poor matching of the two traces in the potentiometer for each channel. This effect is magnified in the low volume area of the potentiometer where the traces approach 0 Ohms. Here it is very difficult for the manufacturers to get good matching. The matching varies with potentiometer position. This is not really a problem with good quality pots, but might still be a problem if you have a low input sensitivity high voltage gain amplifier with high sensitivity speakers. Good quality potentiometers from TKD have a 0.1 dB usual to 0.3 dB maximum deviation for a stereo potentiometer. This corresponds to a 1 - 3.5% deviation in the levels between channels. Good luck hearing that difference.

Changes in the sound as the volume is changed

This is due one or two of the two problems explained above. It could be that at high volumes, you are stressing the output stage of your source too much. Also if you have a symmetrical source the balance problems are magnified. If you are worried about your four trace potentiometer for your symmetrical source then used a stepped type with make before break resistors.

Too low a volume at max

This is a problem of a source which has an output voltage which is too low or the input sensitivity of the amp is too high or both. Not really a problem for me with my DAC as a source with 2.1 V and an amplifier with an input sensitivity of 1.6 V

Hour of power
The next problem is not listed above but I will explain it anyway. It is the effect of turning the volume control and it stays really quiet until close to the end where all the power seems to be in the last "hour" of the volume control between 4 and 5 o'clock. The problem is the input impedance of your amplifier in relation to the value of the potentiometer and how this affects the logarithmic(ness) of the curve of the voltage output from the potential divider. Using the Excel file in the attached zip you can plug in some values and understand the problem better than I can explain in text how it manifests itself. The Excel file will show you the minimum and maximum load imposed on the source.

Figure04 shows my system, details of which can be gleaned from this text or seen in Figure01. The output Voltage of the potentiometer is on the left and the -decibels(attenuation) on the bottom. The red line represents the ideal curve and the blue line is the actual curve.

Let's try a 100 KOhm potentiometer instead in Figure05. That is just terrible, all the power is in the last hour as I mentioned above.

Let's leave my potentiometer as it is and increase the input impedance of my amplifier from 3.23 KOhms to 100 KOhms in Figure06. Much better, but I don't feel like buying another amplifier with higher input impedance and I cannot modify my amplfier to have a higher input impedance.

Let's leave the amplifier as it is and change the potentiometer to 1 KOhm in Figure07. That looks much better than my 10 KOhm potentiometer, but now I am possibly stressing my source with 764 Ohm load at maximum volume(formula is above) Most sources will have given up at this stage and be on their knees begging for mercy.

For my system a 5 KOhm potentiometer would have been better, but the 10 KOhm is quite acceptable. A 50K potentiometer or heaven forbid a 100 KOhm potentiometer would be no fun at all with the capacitance problem and the last hour of power problem explained above. With a input sensitivity of 1.6 V and a maximum power of 350 watts @40hms, a source with 2.1 V and 200 Watt 4 Ohm speakers - my maximum power is achieved @ 0.7 Volts which is at the -4.5db mark. The perfect result here would have been -9dB.

The 10 times or 5 times myth
Well this is just a general rule to keep the load impedance high - there is no real maths in there. If you read the above then you know already keep the load impedance high so as not to stress your source. You might get claims of: well your potentiometer has a output impedance of 10 KOhm and your amp has an input impedance of 20 KOhm and that this is bad because it breaks the rule!!. These people don't understand that the potentiometer is not a source. The potentiometer and the amplifier ARE THE LOAD.


I hope my little rant here is helpful to someone. Pointing out mistakes I made is welcomed, pot shots (pun intended) and hit and runs are not welcome. Claims of: "Well I tried one and it sounded terrible" without any details of the components involved and their input and output impedances are not welcome. If you read the above then you will now know that you cannot just drop in a passive preamplifier into any system and expect good results!! I tried to deal with all the factors involved and did not intend to scare people off. Armed with this knowledge you can put togehter a system with a passive preamplifier and get better sound than any preamplifier.
 

Attachments

  • Figure01.png
    Figure01.png
    36.7 KB · Views: 8,831
  • LogCurvePassiveExcelLTSpice.zip
    27.7 KB · Views: 849
  • Figure08.png
    Figure08.png
    9.8 KB · Views: 2,559
  • Figure07.png
    Figure07.png
    15.7 KB · Views: 1,618
  • Figure06.png
    Figure06.png
    17.1 KB · Views: 1,682
  • Figure05.png
    Figure05.png
    19 KB · Views: 8,199
  • Figure04.png
    Figure04.png
    19 KB · Views: 8,316
  • Figure03.png
    Figure03.png
    25.1 KB · Views: 8,335
  • Figure02.png
    Figure02.png
    25.5 KB · Views: 8,515
Last edited:
deChrLam said:
These people don't understand that the potentiometer is not a source. The potentiometer and the amplifier ARE THE LOAD.
The pot and the amplifier are the load for the source. The source and the pot are the source for the amplifier. Assuming a very low source output impedance, then the source impedance seen by the amplifier varies from zero (zero volume) up to a quarter of the pot resistance (half volume i.e. -6dB) and back down to just the origonal source impedance (max volume).

The '5 times' or '10 times' rules are useful provided people don't treat them as laws of nature. How important they are will depend on how non-linear is the output impedance of the source and the input impedance of the amplifier.

Apart from that, a good summary!
 
These people don't understand that the potentiometer is not a source.
I see the point you are trying to make and I also see how DF & Pano are expanding on your statement.

I don't think they make their clarification strongly enough.

The volume pot is the load for the source (transmitter).
The amplifier (receiver) is the load for the volume pot.

Take these two statements and swap around the words while maintaining the meaning.
The Transmitter is the source for the volume pot.
The volume pot is the source for the receiver.

No matter what one "thinks" The source/transmitter feeds the receiver/next stage. That is ALWAYS the case. I cannot think of any exceptions.

These people do understand that the potentiometer is a source.
 
Account Closed
Joined 2001
If you disconnect the source from the load then you prevent transmission and prevent reception.

You do?

If I turn my FM radio (receiver) off, I'm preventing the transmitter from transmitting? :)
Can't you have a source without a (viable) load?

Sorry, but you can see how the labels might get confusing for non-techies.


DF96,

Totally agree with your last paragraph. Well described.

Cheers,

Dave.
 
Ok guys, thanks for your efforts. Been doing some reading: Impedance bridging - Wikipedia, the free encyclopedia and Playing in excel and with simulations in LTSpice. Hey I can hear a a penny dropping slowly.

I will fix up the rest of the text tomorrow. @Davey you are right, it did get a bit long - someone said once if you want to truly understand something, then write a book about it.

Cheers.
 

Attachments

  • LTSpice.png
    LTSpice.png
    14.5 KB · Views: 2,391
  • OutputImpedance.PNG
    OutputImpedance.PNG
    26.1 KB · Views: 1,566
You need to alter your excel model if you want it to more closely predict what really happens.

eg. add in the effect of the load impedance.

say your load impedance is 1M0 // 47k, then at the -6dB position shown in the model, the actual lower leg resistance is 1M0//47k//5k = 4k499
The attenuation becomes
20*[log{4499/(4499+5000+50}] ~-6.54dB

As you move the pot around the rotation the load has a greater or lesser effect.
As you change the load it has a greater or lesser effect.
The effect is quite small as long as the load impedance >> source impedance.
In this example we are looking at Rs (of the pot) <=2k5125 and Rload = 44k89 The Rload:Rsource ratio is >17.87:1
Getting very close to the 20:1 that you will see I recommend fairly regularly.

BTW.
change the lower leg resistance from 5000r to 5025r leaving the upper leg @ 10000-5025 and you will find that the maximum Rs does come out at the predicted (back of a fag packet) result of [Rs+Rpot]/4 = 2k5125
That would be the true half voltage output rotation of the pot, when the load effect is ignored.

Another lesson from the model.
note how small the pot lower leg resistance needs to be to approach zero voltage at the output. Most pots cannot get this low. Quite often the minimum volume is of the order of -70dB to -90dB. Few pots can better -100dB, very few. And even with audio taper, just the tiniest rotation gives a big jump of the order of 10dB to 20dB increase in output.
 
Last edited:
Potentiometers are not the most transparent device to tune the volume. Mobile contact+carbon resistance are far to be ideal. (noise, grainy sound)
Better to go for metallic resistance arrays, mechanically (rotators) or electronically switched. You can even set that way a constant source impedance for your amp, where it behave the best. (low pass filter in it, square waves overshoots etc.)
 
  • Like
Reactions: 1 user
Much of the music we hear will have already gone through a number of carbon pots and carbon film resistors. The contact resistance is not too much of a problem provided the next stage has a high input impedance, and no DC current draw. Switches can introduce clicks unless DC is kept well away.

Some people prefer switched attentuators and they may in theory be best, but many have also been disappointed to find that they can't have them without also introducing extra coupling capacitors and ground resistors. As in all engineering, you have to choose your compromises carefully.
 
I think I should just use a 10K logarithmic potentiometer to get a nice pure sound
Can you please explain why you chose a logrithmic pot, over a linear pot? A linear pot gives you more precise control of the output level.

Japanese components typically use logrithmic pots, because they give the false sense of a much bigger amplifier. American/European components typically use linear pots, because they are more acurate.

MLStrand56
 
Can you please explain why you chose a logrithmic pot, over a linear pot?
A linear pot gives you more precise control of the output level.

Hi,

A linear pot will give you awful control of volume levels, not precise at all.

The best way is a linear pot with a log law faking resistor.
It will have far better channel matching than a dual log.

An externally hosted image should be here but it was not working when we last tested it.


example from ESP of a linear pot + log law faking resistor

Channel matching at low levels can be improved further by adding
a resistor that gives a minimum gain, rather that zero, at zero.

rgds, sreten.
 
Last edited:
In a conventional pot (whatever law) the worst piece of resistance is the wiper-track interface. Therefore you want as little current as possible through this resistance, so any resistance variation does as little damage as possible. Adding a law-changing resistor to a linear pot does the very opposite. OK in an emergency if the right pot is not available, but I would avoid it when possible.
 
Can you please explain why you chose a logrithmic pot, over a linear pot? A linear pot gives you more precise control of the output level.
Japanese components typically use logrithmic pots, because they give the false sense of a much bigger amplifier. American/European components typically use linear pots, because they are more acurate.
The sensibility of our ears is logarithmic. A perfect logarithmic pot will gives-you the feel of the same volume increment for each same angle increment, while a linear pot will up most of the power at its beginning.
A pseudo logarythmic pot (they are all pseudo) is make of several superposed carbon slices at various places. It is more prone to generate parasitic noises when you move-it.

Much of the music we hear will have already gone through a number of carbon pots and carbon film resistors. The contact resistance is not too much of a problem provided the next stage has a high input impedance, and no DC current draw. Switches can introduce clicks unless DC is kept well away.
I have mitigate comments about you entry. In fact, in professional mixing desks, we use plastic Penny and Gilles linear potentiometers. And more and more records are produced in the Digital domain.
Despite i'm very suspicious about "audiophile" differences, i found carbon resistances to sound a little grainy, comparing to metallic ones.
Too the relative precision between the two channels of a traditional potentiometer is far to be perfect, and can lead to a little change in the balance between the two channels. reason why we used rotators and array of resistance in our amps.
Not to forget you can find very good sounding cheap circuit we can use as remote control resistances array based. As an example: http://cgi.ebay.fr/CS3310-Crystal-Volume-Remote-Control-Preamplifier-Kit-/260706392408.

Did-you agree ?
 
Last edited:
In a conventional pot (whatever law) the worst piece of resistance is the wiper-track interface.

I built a remote control using an Analog Devices AD5206 RDAC, and while it was inititially intended to drive an LDR (which would pass the audio signal), I connected one up to pass the audio signal through the resistor in the AD5206.

It is a linear pot, but the firmware that controls it uses a logarithmic function to control the pot, so the taper is truly logarithmic. As an aside, there are no mechanical potentiometers that have a true log taper - all are approximations.

The AD5206 device sounds very, very good to me. No noise, just purity. Don't know if it was a "proud papa" thing or not, but I ended up selling all my preamps and going solely with this type of arrangement instead.

Here's a write-up of my experience with this approach: