Amp preamp impedance matching...can anyone explain?


Hi, I currently have vintage tube gear, but want to try a SS amp with my tube preamp, and may try a SS preamp with my tube amps. I have noted there is an impedance matching issue, but do not understand it. Can anybody provide a quick summary?
Thanks
Jim
river251
Ideally the input impedance of the amp should be 10 or more times greater than the output impedance of the preamp, at the frequency for which preamp output impedance is highest. Otherwise audible frequency response irregularities MIGHT result.

If, as is often the case, the highest output impedance of the preamp across the audible frequency range is not known, and only a nominal output impedance is specified (perhaps based on a frequency of 1 kHz), I suggest using a ratio of 50 or more, and preferably 75. Many tube preamps, and some solid state preamps, use a coupling capacitor at their outputs, which can cause their output impedance to be much higher at deep bass frequencies than at higher frequencies.

Impedance incompatibilities are most likely to be encountered when using a tube preamp with a solid state power amp.

If the power amp has an input impedance of around 47K or more, it is unlikely that there will be an issue, even with a tube preamp.

Regards,
-- Al
I too have heard the 10x ratio, but are not there some companies that preach a transmission line approach? I have seen references to a 600 ohm standard in older studio equipment.
Onhwy61, that would apply to situations where the component providing the signal is specifically designed to be able to drive 600 ohms, without a significant increase in distortion or a significant impact on bandwidth or frequency response.

Most consumer gear will not meet those requirements. I believe that Atmasphere is a notable exception. According to posts Ralph (Atmasphere) has made in the past, a benefit that would result from loading the balanced outputs of his preamps with a low impedance like 600 ohms is minimization or elimination of cable effects and cable differences.

Best regards,
-- Al
Onhwy61, I think(?) I remember Jeff Rowland espousing such a philosophy. I also seem to remember(?:-)) past Audiogon contributor Sean suggesting the same thing. Some years ago Dave Wilson writing in The Absolute Sound offered a tiered approach to impedance matching.
Thanks Almarg. What sort of frequency response irregularities?

I have a CJ PV-5 (only thing I found was a post saying 200 ohms output impedence) and Marantz 8B (and Mac MC240 and MC30s but listening to the 8B). I am considering trying a SS amp and considering B&K ST-140 (24Kohm input impedence), Aragon 2004 MkII (22Kohm), and McCormack DNA .5 or 1 (100Kohm).

Regarding:

If, as is often the case, the highest output impedance of the preamp across the audible frequency range is not known, and only a nominal output impedance is specified (perhaps based on a frequency of 1 kHz), I suggest using a ratio of 50 or more, and preferably 75. Many tube preamps, and some solid state preamps, use a coupling capacitor at their outputs, which can cause their output impedance to be much higher at deep bass frequencies than at higher frequencies.

....for the gear I mentioned, do you think I need to worry about the 50:1?

Thanks very much,
Jim
07-11-12: River251
Thanks Almarg. What sort of frequency response irregularities?
If I may be allowed to chime in - like Almarg wrote "Many tube preamps use a coupling capacitor at their outputs, which can cause their output impedance to be much higher at deep bass frequencies than at higher frequencies". So, the preamp output impedance is usually not a constant. It's often lower in the mid/hi freq & higher at bass freq. When higher in the bass freq, you *might* see a loss of signal power from pre to power + the loss of control by the power amp of the bass region. The listener usually perceives this as flabby/bloated bass.
I know that a friend of mine experienced this with a highly regarded preamp that had 600 Ohms output impedance while his power amp had 10K input impedance. It was a bad match & the preamp was sold off immediately.

I have a CJ PV-5 (only thing I found was a post saying 200 ohms output impedence) and Marantz 8B (and Mac MC240 and MC30s but listening to the 8B). I am considering trying a SS amp and considering B&K ST-140 (24Kohm input impedence), Aragon 2004 MkII (22Kohm), and McCormack DNA .5 or 1 (100Kohm).
from my personal experience this should not be an issue. My preamp has 200 Ohms output impedance & my power amp has 10K input impedance. The 2 units mate up very nicely across the entire audio freq range.
FWIW.
Thanks for chiming in, Bombaywalla, and saving me some time :-)

I found some information specific to the CJ PV-5, though. The schematic and parts list can be found here. It can be seen that the output coupling capacitor is two capacitors in parallel, totalling 2.35 uF. At 20 Hz, that corresponds to an impedance of 3388 ohms, based on 1/(2piFC). (That figure will be vastly lower at mid-range and treble frequencies, consistent with the 200 ohm nominal output impedance). It is most likely safe to assume that the overall output impedance at 20 Hz will be just a little higher than the impedance presented by the capacitors at that frequency, so let's call it 3500 ohms.

22K/3.5K and 24K/3.5K are ratios that are between 6 and 7, lower than the ideal of at least 10. So there may be a slight impact on the deepest bass frequencies, which may or may not be perceptible depending on the deep bass extension of your speakers, and on room acoustics. If your speakers have good deep bass extension, my instinct would be to play it safe and avoid the B&K and the Aragon, although it's a close enough call to be debatable.

Regards,
-- Al
Even with an ideal pre-amp to amp impedance match the B&K ST 140's sonic signature (as is the c-j's too) is a bit soft in deep bass, compounding it with a less than perfect impedance match might not be the best choice. I wouldn't recommend a low impedance speaker (<4 Ohms) with the B&K ST 140 either. Which is not to say that in an appropriate system the B&K ST 140 is not a good amp, as it certainly is, especially for the money. Keep in mind that the B&K ST 140 is getting a bit long in tooth and B&K is no longer in business. Though I would imagine that having the B&K ST 140 serviced by an outside source wouldn't be too difficult should the need arise.
I would like to add another factor to the discussion.

The ratio of 10/1, while ballpark, may not be the only factor if the solid state device uses operational amplifiers in the output stage. Many opamps have output impedances well below 1K, but if you try to drive a 10K load with these (particularly at voltages over 1 V RMS), you can often get flat, lifeless sound performance. Of course, the results also depend on the type of opamp used. Some are better than others.

For opamp driven output stages I suggest an imput impedance of 47K or higher, with output voltages below 1-2V if possible, irrespective of output impedance.
We have been talking ratios of pre-out to power-in as 1:10, 1:50, 1:75 or absolute numbers as 47,000 ohm.

Is there a limit to how high these ratios or the power-in should go, before it sounds bad?

I am a layman asking this... Thank you
Is there a limit to how high these ratios or the power-in should go, before it sounds bad?
No, not that I can envision.

Regards,
-- Al
As a simple rule of thumb most any transistor preamp can drive any tube amp.

Its not always the other way 'round as pointed out above. (FWIW we get around the problem by direct-coupling and using paralleled tube sections to obtain a low output impedance that can drive any amp around.)

But if there are coupling caps its a different ballgame. For example, ARC recommends no less than 30K as the input impedance of the amplifier with which the Ref 5 is used.

10:1 is a safe minimum value, but you can go 100:1 and its fine. So in the ARC example above, standard engineering practice suggests that the output impedance of the Ref 5 is 3K ohms. However they are fairly conservative and are showing 600 ohms on their website, although that might be at 1KHz. IMO the 20Hz output impedance tells the real story.

The lower the output impedance, the more the line section can control the interconnect- by that I mean the less sonic attributes the cable will have. So the lower you can get the output impedance, the better (all other things being equal...).
I am really learning a lot from you guys. Thanks.

I was just playing with the thought on the different types of power-amps that my preamp of 20 ohms output, can be connected to! Seems like it will be alright. ;)
I have a tube DAC that uses a single 6SN7 tube with 1 microfarad coupling caps. I was told it has 8 kohm output impedance. I wanted to get this integrated SS amp but the input impedance is 10 kohm. So this is a horrible match, so I stayed away. My tube amp with passive stepped attenuator at the input of the amp has 100 kohm impedance. This DAC/amp combo sounds dynamic as hell with deep bass extension. But I really wanted that SS integrated.
Reading this as well and I wanted to add my specs from my Marantz PM-KI-Pearl Integrated that I would like to entertain using as a preamp and then adding an amp...tube or ss still not determined.

The testing of this Pearl unit shows "The output impedance from the Preamp Out jacks was 217 ohms at high and middle frequencies, rising to 495 ohms at 20Hz."

Based on this information, does one still look for the above in the input impedance of 47k or higher?
Vineman, since the worst case (20 Hz) output impedance is known, just multiply that by 10 to calculate the minimum desirable amp input impedance.

495 ohms x 10 = 4.95K (minimum)

Nearly all power amps will meet that criterion.

Regards,
-- Al
Thanks for clearing this up here...
07-19-12: Dracule1
I have a tube DAC that uses a single 6SN7 tube with 1 microfarad coupling caps. I was told it has 8 kohm output impedance.
that is correct. impedance at 20Hz = 1/2*pi*20Hz*1e-6 = 7957.7Ohms, which could be rounded up to 8KOhms. The manuf wasn't lying! ;-)


I wanted to get this integrated SS amp but the input impedance is 10 kohm. So this is a horrible match, so I stayed away.
right again! a horrible match down the in bass region. Your system would have sounded tinny. So, good thing you stayed away.

But I really wanted that SS integrated.
better put a lid on that desire while you have this particular tube DAC! ;-)
Does the voltage gain play into this at all? The amplifier I am using is a McCormack DNA-125, which has 30db of gain. Its input impedance is 100 kOhms. I am using an Audio Research SP8 MkII preamplifier with it which has an output impedence of 1,000 ohms.

I had Audio Research lower the gain in my SP8 recently, but needless to say, the volume through my Vandersteen 3A Signatures still gets loud in a hurry.

I was wondering if someone could please recommend a low gain, solid state Audio Research amplifier that is rated between 100-200 watts per channel? I do not want a tube amplifier, and I do not want to get rid of the wonderful SP8 MkII.
07-24-12: Distortions
Does the voltage gain play into this at all?
No, gain and impedance compatibility are two separate issues, although if the amplifier's input impedance is too low in relation to the preamp's output impedance (which it is not in this case) there will be a very slight reduction in overall gain.

I see that in stock form, prior to the recent modifications, your preamp had a line stage gain of 26 db. That in combination with the amp's 30 db gain, as you realize, is way too high, especially if you are using a digital source having the typical full scale output of 2 volts or so.

I see that you have another thread in progress on amplifier replacement and other possible approaches, and you've already received an intriguing input from Steve McCormack himself!

Regards,
-- Al
I have a Pass X250.5 amp whose input impedance is 22 kOhms. I've been reading about tube preamps, specifically the ARC Ref 3 (output impedance 600 ohms balanced) and Cary slp-05 (output impedance 500 balanced), both of which seem to draw comments that they wouldn't be good matches with Pass amps because of the low input impedance of the Pass amps.

So I went on the Pass website and see that the XP10 and XP20 preamps have output Impedances (for balanced) of 1,000 ohms! Presumably, these two preamps are ideal matches with my X250.5 amp, but if I do the math, how is an XP10/XP20 with 1000 ohms output impedance a good match for the X250.5 when an ARC Ref 3 with 600 ohms output impedance is considered by some to be a bad match? 22000/600=37 for ARC and 22000/1000=22 for Pass XP10/XP20

What am I missing or not understanding?

Lpw
Lpw, as I indicated earlier what is important is the preamp's output impedance at the (audible) frequency for which that output impedance is highest. That figure will often occur at 20 Hz, especially in the case of a tube preamp (due to the coupling capacitor that most but not all tube preamps have at their outputs), and in the case of a preamp having an output coupling capacitor will very often be much higher than the specified value. The specified value can be assumed to be based on midrange frequencies, unless explicitly indicated otherwise.

According to Stereophile's measurements:

Ref 3:
The output impedance was also to spec., at 635 ohms balanced and 326 ohms unbalanced in the treble and midrange, but rose to 1437 ohms and 625 ohms, respectively, at 20Hz.
SLP-05:
The SLP 05's output impedance is specified as a usefully low 400 ohms. However, I got significantly higher values for the unbalanced output of 1500–1600 ohms in the midrange and treble, rising to 3400 ohms at 20Hz, with similar if slightly lower figures for the balanced output.
The specs for your X250.5 indicate an input impedance of 30K balanced and 20K unbalanced, which are the same as the input impedance specs for many of the other Pass amps. At some point in the past, though, their unbalanced input impedance may have been spec'd at 15K, lower than it really was and is. See the specs and measurements in Stereophile's review of the XA30.5. That discrepancy perhaps contributed to some misconceptions.

The XP-10 and XP-20 are spec'd as having output impedances of 1K balanced, 150 ohms unbalanced. Presumably balanced connections would be used between all of these particular components.

Based on the foregoing numbers for balanced connection, the Cary's output impedance rise at deep bass frequencies results in it not meeting the ideal minimum ratio of 10 that is applicable if the preamp's maximum output impedance across the audible frequency range is known (30K/approx. 3.4K = 8.8). Although it comes close enough to probably be marginally acceptable in some systems (depending on the deep bass extension of the speakers, for one thing).

The Ref 3 should be no problem, in terms of impedance matching. 30K/1.437K = 21.

Although I couldn't find measurements, it can be presumed that the specified 1K output impedance of the solid state Pass preamps does not rise significantly at 20 Hz (or other) frequencies, so it too should have no problem working into a 30K load.

Regards,
-- Al
Al, isn't it the case that ARC does not recommend less than a 30K load with their preamps? I know that is true of at least one of the recent Ref series.
Hi Ralph,

Yes, I seem to recall that being true of some of their later preamps and/or phono stages. For the Ref 3, though, the specs shown here, and also in the Stereophile review I linked to earlier, indicate 20K. Although the wording leaves it unclear whether that applies to the balanced or unbalanced outputs, or both.

In any case, given the 30K impedance of the balanced inputs on the Pass, and the measured worst case (20 Hz) output impedance of 1437 ohms for the Ref 3 (which it should be noted is only a bit more than a factor of 2 greater than the midrange output impedance), it would appear that there shouldn't be any frequency or phase response issues. And I'd be surprised if there were any audible distortion issues either, although you can speak to that possibility more knowledgeably than I can.

Best regards,
-- Al
Al and Ralph, both the ARC Ref 3 and 5 literature recommends that the minimum combined load on the Main outputs is 20K Ohms. My amp is the ARC VS-115 which has an input impedance of 300K Ohms (balanced) -- so no problem there. If you pull my threads, you'll see where I got into a problem was when I asymetrically loaded the 2nd Main output with a 20K Ohm load. Main 1 -- VS 115 (balanced) 300K Ohms; Main 2 --self powered subwoofer (SE) 20K Ohms.

As I mentioned in my old threads, the solution was that Tom Tutay designed and built an impedance buffer device that summed the left and right channels without shorting the Ref 3/5 output Mains and enabled me to load Main 2 (self powered woofer) with a balanced input having 330K Ohms impedance. So my overall preamp output impedance load is 157K Ohms -- well above the recommended minimum.

Result -- tighter and more extended bass. Possibly more open midrange. Per ARC, symetrically loading the Ref 3/5 will also extend tube life -- but what do I know???? FWIW.
Al, thanks so much for your explanation - I get it now, although coupling capacitors and such are still beyond the scope of understanding for me :)

The only regrettable part in all this is that it seems impossible to determine good impedance matches based on the nominal output impedance spec that the preamp manufacturer will show. Everyone would have to rely on magazines like Stereophile or audio websites to actually review a certain product and show readers the detailed impedance measurements across the whole spectrum in order to do the correct calculation. This is not a problem for popular brands like ARC or Cary, but the lesser known brands might not get a review, and then we audiophiles are taking an educated guess as to whether a preamp is a good match impedance-wise with a power amp.
Bifwynne, that is a common problem if you are trying to use a balanced preamp with a subwoofer. Many subs only have SE inputs.

Another solution is that Jensen transformers makes a transformer that is optimized for subwoofers. It has a fairly high impedance input that can be run balanced or SE, and has bandwidth to less than 2 Hz so it does not mess up the bass.

But I think you are lucky you ran into Tom Tutay. He does good work :)

I don't know if this thread is still active (if not I'll have to find another forum).

I'm driving a Quad 303 power amp from the headphone socket of my Yamaha Amp.

From the published specs:

The Quad Power Amp i/p is 0.5V @ 22K ohms impedance

The Yamaha Headphones o/p is 0.51V @ 330 ohms

It looks like the impedance match is ok, but I had to attenuate the signal with 1MΩ resistors which I soldered into the DIN plug (Quad i/p).

What I want to do now is connect to the aux i/p of a Cambridge amp; the impedance is 47KΩ (the signal requirements aren't mentioned in the specs, but I would think it'll be around 15omV). From what I read here this will be a mismatch one option is to solder 47KΩ resistors across the i/p's (sig to scr) in the phono/rca plug; I don't really want to go inside the amp and change the resistors to lower values.

I'll also need to attenuate the signals perhaps with higher values than 1MΩ.

I wonder what the experts think.

Regards:
-Pete