Mix and match impedance

January 21, 2018
 by Paul McGowan

Popular audiophile lore has it that a mismatch in impedance between the preamp and amplifier will result in bass loss. Yup. If the amp's input impedance is too low weaker bass will be the result. That's true in the same way it's important to pump the car's accelerator before starting it.

In the days of carburetors, a cold engine started easier when you primed the pump with a few squirts of gas. I remember a manual choke on cars that had to be pulled when they were first started. But no modern car has a carburetor just like no modern amp has impedance so low it would impact a preamp. But it was once true.

Preamps of old, especially tube preamplifiers, were capacitor coupled (many still are). On preamplifiers where sound and performance mattered high-quality film capacitors were used at their output, which meant values were small. And, small capacitors result in rolled off bass in the presence of low impedances. This is because the preamp's output capacitor and amplifier's input resistor form a high pass filter (reducing low bass).

But today we need not worry about such matters. Most power amplifiers have input impedances of at least 20kΩ and higher. Our own power amps, for example, have 30kΩ input impedance, high enough to never reduce bass levels of capacitor coupled preamplifiers, even if they are the vintage kind.

So yes, matching impedances used to be a thing to worry about. Modern designs have obviated the need for hand-wringing.

I put together a small video on the subject you can watch here.

Subscribe to Paul's Posts

23 comments on “Mix and match impedance”

  1. I had two cars I remember with manual choke. They were both terrible. The idea that anyone has to put up with such technology or anything incompatiable in any area of consumer technology defies belief. I remember in the 1980s loading drivers to get a third party printer to work with a PC took hours. Now you don’t even notice.
    I had a strange one yesterday. A modern UV filter was preventing a 1970s Leitz Minolta CL’s sensor from taking a light reading. The sensor is a huge thing, the size of a fingernail. Just one of those incompatibilities when you get when you muck around with old technology, which is totally avoidable. For me photography is a hobby so I don’t mind, but audio is about listening to music, so a modern AIO avoids such issues entirely.

  2. The Sumo line of amplifiers designed by J. Bongiorno (RIP) have input impedances of 1 MegOhm in their unbalanced inputs, while in the balanced inputs they have 10 KOhm, and a total capacitance in their P: S: of 60,000 uF
    The Delta 250 Mono PS Audio, have input impedance 30 KOhm and a
    a total capacitance of 240,000 uF in P.S.?

    Why then, in either case, (considering today's topic) both brands of amplifiers have an impressive bass response?

    Can someone explain it?

    1. Because bass response doesn't really concern itself with input impedance anymore. Even Bongo's 10K isn't too low for most preamps though a few tube jobs might be offended.

      The real trick to bass response, surprisingly, is in the amp design itself. One of the craziest secrets to excellent bass we discovered is lowering the bottom high pass filter of the amplifier to at least 10X below where you think it ought to be. And I know James felt the same way. Most design guidelines suggest placing the pole pretty low, perhaps 1Hz, but that tends to give you wimpy bass. Lowering the pole to 0.1Hz offers surprising good bass results and a factor 10 below that even better.

      1. Paul, I find that a fascinating statement. You are setting the amp to a level where the effect on the audible bass must be a tiny fraction of a percent but are getting a discernable improvement. This confirms my suspicion that audiophiles listen at a level of detail which is unsuspected by the general public.

        1. Chris - not so sure about that. You seem to be translating what is on paper a vanishingly slight number to an equivalent percentage change in audible program.

          This is one of the things us hearing-based, part-time-measurementalists keep going on about. Putting a percentage on it does not correspond to the perceived result.

          1. Beef, not related to today’s post but I’m starting to plan Axpona. When my daughter was in town and it was at the old venue it was easy to use Southwest points to fly into Midway and stay with her and then get the train/Uber out to the venue. Now that it’s at Schaumburg there’s no subway that goes there and my daughter no longer lives in town so I have to figure out accommodations, plus very inconvenient to Midway so I can’t use Southwest points LOL. How easy is it to get from O’Hare to Schaumburg and are there any reasonable accommodations up that way? I haven’t looked into Airbnb. Feel free to just PM me if you don’t want to fill up the blog post!

          2. BB, I agree with you. I think small differences (on paper) can make a big difference (in sound). My original post may not have been as clear as it should have been. Since I have 'flu at the moment it was made from a tablet rather than a PC, which led to it being rather terse, since I find typing on a tablet a bit of a nightmare.

  3. Aside from the bass roll off issue, aren’t there other influences of the impedance differential on the character of the sound that may be important to discerning ears? Isn’t there an optimal impedance ratio unique to our individual systems that most of us will never know exactly what it is because we cannot adjust our output impedancess easily?

    Before I got the DS DAC with its output volume control and then the BHK preamp, I drove my amps with my DAC, which like the DS DAC and the BHK preamp had a very low favorable output impedance. But it’s output was a fixed 2v, requiring an external volume control (attenuator) in order to modulate the strength of the signal driving the amps. I used a Tortuga “passive preamplifier” (essentially an attenuator volume control using LDR technology). The Tortuga attenuator had an unusual feature—an adjustable output impedance which allowed the output impedance to be adjusted by simple remote control to achieve the most optimal sound (including bass response, frequency range, clarity, “warmth”, etc.) The best sound in my system was not achieved with the lowest output impedance, but somewhere in the range of impedances offered. This leads me to ask, are there other important aspects of component to amp impedance matching other than bass roll off? The Tortuga experience led me to conclude “yes”. There is an optimal ratio in many systems.

    1. As long as the device feeding the amplifier has reasonably low output impedance you're in good shape unless that device was this passive thing.

      A passive attenuator is an entirely different beast altogether. Here, a low output impedance is going to work against sound quality because what you're doing is hurting the performance of the actual attenuator. Building an attenuator with too low or too high values is a dicey proposition because passive attenuators will always be dependent on everything else around them.

      For example when we built our earliest preamps we had a switchable passive attenuator we labeled Straightwire. In that mode it was a simple pots in a box. We worked hard at selecting the right attenuator. Too high of an impedance device and the sound became flaccid, soggy, wimpy. Too low and it took on an unwelcome stridency with sharp edges (because it was stressing out the circuitry before it). We eventually wound up with 10K as the perfect balance. That said, even then, the straight attenuator without benefit of an output buffer had weak bass.

      1. Your explanation supports the rationale of having an adjustable output impedance on a passive attenuator. With my passive, I was never bass starved at most impedance settings.

        There appears to be an optimal preamp output impedance for each preamp and amp. The advantage of going with the same manufacturer, such as PS Audio, for amp and preamp is that they have been designed for optimal synergy, and that would include optimized impedance relationships. The interconnect used must also have something to do with the synergy, which I understand is why some cable manufacturers include on their interconnects some kind of impedance influencing module customized for the particular preamp and amp being used.

  4. Paul, I thought feeding an amp with a low input impedance, using long ICs , caused a loss of highs.
    The bass roll off is something I did not know, so I learned something today.

    In the tech specs in Stereophile, John would sometimes tell us that using a tube preamp being reviewed that had somewhat high output impedance should not be used with long ICs or a power amp with an input impedance on the low side.

    I also thought there was a basic ratio of something like 10 to 1[?], or you would have a mismatch that would effect the frequency response especially when using long interconnects.

    1. Well long interconnects are a horse of a different color. There we can get a treble roll off because of the capacitance of the cable. If the preamp has high output impedance, the cab;e's capacitance becomes a low pass filter and you'll lose top end. That's different than the low impedance amp input of which I wrote originally.

  5. This is a technically complex concept for anyone who does not have a background in electrical engineering but it is important to equipment performance. Impedance is the resistance to the flow of electrical current. But resistance has a technical meaning too. It is the resistance to the flow of current regardless of frequency and is therefore often measured with DC voltage and current where the frequency is zero. Impedance has another component as well. Reactance is the resistance to the flow of current that IS dependent on frequency. So Impedance is the sum of reactance and resistance. To make matters worse, they don't coincide in time. Sounds strange. You turn on a water tap and water flows immediately. Imagine if it didn't flow for awhile when you turned the tap on and when you turned the tap off it continued flowing for awhile. Or more bizarrely it began to flow before the tap was turned on and stopped before the tap was turned off somehow knowing what you were going to do in advance. Welcome to the strange world of AC electricity. Even more bizarrely you can combine all three at the same time.

    When you connect two pieces of electrical equipment you are dealing with three impedances, the output impedance of the source, the input impedance of the load, and the impedance of the wire connecting them. Here's a bizarre fact, the maximum power is transferred from the source to the load occurs when the input and output impedance are an exact match. But in audio that is generally not the goal. Instead the goal is to get the signal at the output of one component to match the signal at the other component with as little distortion as possible. If the output impedance of the source were zero and the input impedance of the load was infinite then the impedance of the wire would have no effect. But that is not the real world. Getting real circuits to be as close to this ideal as possible should be the goal of electrical design engineers. But everything has its limits including engineers. So we come up with a lot of designs by engineers and wannabe engineers that ignore this principle. Here's an example.


    Look at the measurements in figures 1 and 2. The red and blue curves are for this preamplifier output connected to a 100,000 ohm load. The curves in cyan and magenta are the performance connected to a 600 ohm load. What a difference. This signal circuit is built around three very common vacuum tubes, one 12AT7 and two 12AX7s. These are widely used dual triodes. The signal circuit is simpler than a 1950s table radio that might have cost $20. Price for this preamplifier is $12,500. If I bought one of these at some absurd price, someone get me to a psychiatric hospital fast. So is 5,000, 10,000, or 20,000 ohms good enough for the input of a power amplifier? Is using long interconnects between a preamplifier and power amplifier a good idea? Why did I endorse the use of vacuum tubes for the input of the BHK amplifier? Because their input impedance is typically 100,000 ohms or more. Even with MOSFETs the solid state counterparts didn't come close.

    The output stages of CD players, tuners, preamplifiers are often configured as what are known as cathode followers for tubes and their counterparts for solid state circuits called emitter followers. IMO this is NOT a good design to connect to most solid state power amplifiers. This is why the output impedance is relatively high. A better design would be a small class AB power amplifier whose output impedance wouldn't be 600 ohms but close to zero ohm. When tube amplifiers were expensive, this was not practical and tube circuits that were going to be the connected load already had very high impedances. In the solid state world things changed but electrical engineers didn't. Electrical engineers are not necessarily the sharpest knives in the drawer even if they have a piece of paper that tells you they graduated from school.

    With power amplifiers things can get much worse. This is because the load can become the source and the source can become the load. This is how it happens. When a speaker, especially a woofer is pushed to its maximum excursion by an amplifier, especially a long throw acoustic suspension woofer with a heavy cone, it has a lot of potential energy. A motor and a generator are actually the same machine and a dynamic woofer is a linear motor. When the voltage of the amplifier drops to zero to cross to the opposite polarity the potential energy stored up is converted to kinetic energy generating electricity back at the amplifier. We call this reverse EMF. The speaker becomes an electrical generator driving the amplifier. The amplifier and the wire convert this energy into heat to dissipate it. The lower the impedance of the wire and the amplifier output stage the faster and more efficiently this energy is dissipated. We call this electrical damping. For rotary motors we call it dynamic braking. It's the same thing. Therefore heavy gage wire and very low output impedance are desirable. If the amplifier and wire impedance are high, the speaker will oscillate for awhile at its natural resonant frequency regardless of what the driving frequency was. (don't feel too badly if you don't understand this, I had an argument with an electrical engineer who was a graduate of MIT and designed speakers. He got it wrong too.) the amplifier with the highest electrical damping factor which is the speaker impedance divided by the amplifier output impedance was the Crown Reference 1 and Reference 2 models whose claimed damping factor was 20,000 for an 8 ohm speaker. That's 0.0004 ohms output impedance. It's practically a dead short. There are other impedance anomalies at high frequencies for tube amplifiers at high frequencies probably due to characteristics of the magnetic material in the impedance matching output transformers. Also the output impedance of a vacuum tube can generally be in the range of 15,000 ohms, much too high to drive a loudspeaker without an impedance matching transformer. You may like the distortion, I don't.

    1. ... "Therefore heavy gage wire and very low output impedance are desirable" ...

      I use my 2 Crown Studio Reference "1", with balanced signals, in vertical bi-amping to feed through 4 double cables 8TC 1.5 m long (Kimber) (p / ch)

      Each of the Focus, have 3 woofers of 12 ", so you need an amplifier with the highest possible damping, and those were the Crown, I do not know that there is an amplifier on the market with a damping of 20,000.

      But more than one manufacturer, has told me that the damping of my Crowns (made in the USA) is obtained thanks to an excessive negative feedback, which is not correct. I have taken this with skepticism because the Crown figures for the different Distortion classes in those models, are very low, and what is more important: for me they sound very good.

      I ask: Do the opinions of these manufacturers make any sense?

      1. They are jealous as hell. For what you paid, you've got some fantastic amplifiers. I'd buy them myself. If you use too much negative feedback, distortion will rise. Negative feedback compares the input and output waveforms and pre distorts the input by the negative of the difference so that the output and input are an exact match. Crown amplifiers are known for super low noise and distortion and wide flat frequency response. Compared to so called audiophile amplifiers, it just blows them away. It is also super reliable.

        Sadly Harman Industries was bought by Samsung and they have dropped the Crown product line. Professional audio engineers almost always had Crown on their bidder's list. Great value, great performance, super reliable backed up by great service and an excellent warranty. What more could you want?

        1. Thanks SM, for your explanation on the subject of feedback.

          What I think is that they can not make an amplifier with those characteristics, which include a sophisticated protection system for the large amount of current that these amplifiers handle, which does not compromise the purity of the signal, and they can not do it at that price. , due to Crown's experience.

          The only argument that has been expressed in this forum, is that they do not sound good, which I do not care .

  6. From what I gathered..... When the problem with a particular tube preamp has that capacitor placed at its output designed to prevent DC passing? Is that not in effect creating what we might call a first order crossover as seen with speakers? Thus causing a bass roll off?

    1. It's a first order high pass filter. It's more commonly referred to as a DC blocking capacitor. It prevents hundreds of volts from reaching the next stage that could blow it up. By making the capacitor value larger the cutoff frequency is lower for a given circuit. Here are some images. Note that every one of them has a capacitor connected between the cathode and the output terminal except for a few where it was inadvertently omitted but it should be there.


  7. Paul, like jeffstarr, I've always been told a tube preamp driving a SS amp which has too low an input load would roll off the treble. Was I listening to the wrong people?

      1. Thanks, yes I've heard about high frequency loss with long cables so that is part of the story.

        However, at one time I had a pair of Boulder amps (memory tells me their input was 10K, but that was many, many years ago) and my tube preamp did not work well driving them. So it was from my own experience as well as word from others.

  8. So I am late to this. I have an amp w 27k ohm input impedance and a Cary slp98 w 440 ohm output impedance which rises to 6k ohm at low frequency. I think I was losing low end. I am getting the Cary direct coupled and the output cap is being increased in size. Also having hexfreds put in the power supply. Do you think this will fix the issue? I am told this improves performance. Any input?

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram