advertisement


Valve pres with solid state powers , what works

I use a VTL pre with a XTC POW1 which replaced a VTL ST85 power. The only mismatch was too much gain from the pre which necessitated attenuators. It sounds much better that the all valve combo because the power suits the speakers better.

I am not sure how often these mismatches occur because every time I have bought something it works with the rest. Sometimes not as well as I would hope.
 
I got great enjoyment from an Audio Research LS26 and a pair of Pass Aleph 2 monoblocks. In the summer when it got too hot for the Alephs I used a Plinius SA50 Mark III power amps. I've disposed of all of them in favour of an all solid state set up of Pass Labs SP20 pre-amp driving Pass XA100.5 monoblocks. All these options were used to drive a pair of ART Alnico 8 loudspeakers.
 
Obviously, again, there is far from a simple answer, but modern SS power amp's have input impedance values of 20K and thereabouts, as you have said, but what would most valve pre's be happier with? OK, more, but realistically?

Where were/are valve power amp' input impedances pitched? That is assuming that they are aimed at operating with a valve pre'. Or is the "matching" achieved between valve pre' and power achieved in some other way?

I can only generalise in answering this as there are so many variables.

Valves are high input impedance (Zin) devices and it is hard to imagine any reason for a valve power amp to be less than say 100K Zin. 1M is not unusual... Almost all then will be 100K or more and very easy to drive. They tend to not have input capacitors/RC network to ground as well and the few that do tend to have very small caps only. I'll come back to this.

SS power amps can vary from 10K or less and up to about say 100K for ones with bipolar transistor inputs but ones with FET inputs an be as high as valves. Most (big generalisation here) SS power amps are between 20K and 68K with around 50K being about the most common. And yes folks I'm sure you can find examples of 5K and 2M if you look hard enough...
Many, but not all, have an RC filter at the input which rolls of HF above say 35KHz - 100KHz. This is a resistor in line with the input socket and a capacitor to ground after it.

The very high Zin of valve power amps puts very little load on the pre amp, it draws very little current from it and doesn't "pull down" the output of the pre amp.

Imagine the waveform from the pre amp having to carry a weight up and down with it due to the load of the power amp and you're getting there! Lower Zin of power amp means "more weight" so the signal is reduced in amplitude.

Now this is where it (hopefully) gets interesting:D If the pre amp was a simple passive then it is nothing but a resistance and won't generate distortion in driving a load. The load of the power amp will drag it down in amplitude as described above and that is all it will do.

A typical valve line stage though, and especially ones such as the Croft 25 and Audionote, not only have a Zout around as high as a 20K passive but their ability to "push" is quite a bit different to their ability to "pull".
Now what this means is that when the load of the power amp input pulls down the waveform from the pre amp it pulls down the top of the waveform more than it does the bottom (think of a sine wave with an imaginary centre line here. above the line = pull and below it = push)... hence it distorts the shape of the waveform.
In this most basic form I'm describing 2nd harmonic distortion, which will usually predominate in most simple valve stages (but doesn't have to).

So..... the lower the Zin of the power amp the higher the distortion.

The input RC filter on many SS amps can is some cases be severe enough that some valve pre amps will be "pulled down" more at HF as this RC filter "adds more load" as frequency increases, this is how it works as a filter in fact but with too high Zout from a pre amp the filter may be starting down as low as say 14Khz rather than the intended 40Khz when its used with a preamp that has a Zout of only 1K or less, as is typical with an active SS pre amp.

Hence "how low Zin a SS power amp can I use with a valve pre amp?" is kind of "how much distortion do you want/can you tolerate?"
To be less flippant and give at least some guidance on this then, 100K is great, more can only be good but is pretty academic by the time it is 100K. 50K will be fine with most valve pres but distortion will typically have risen from say 0.02% with 100K to say 0.04 - 0.1% but should be fine still. By 20K distortion will often be up at say 0.2 - 0.7% ish... which some may like!

I must stress that valve pres vary considerably in Zout and load driving ability though. The above is true for simple one valve line stages that are not cathode followers. For most cathode followers (and probably the majority of single valve line stages are a cathode follower) it works exactly the same as above but Zout is usually low enough to drive even 20K with little issue. If the cathode follower is an ECC83 then 20K is pushing it but will probably still be fine. Generally if your one valve line pre has no gain then you can assume it is a cathode follower.
 
Do you know what? I followed 90% of that, with ease, first read through (even without analogies :) ) . Brilliant, very easy to follow, thanks. I will go through again to get the rest (and then forget it, I have to admit, ashamedly).

My first thought about impedance mismatch (SS power amp too low for valve pre) was "ah - attenuators", thinking that they would be reasonably simple resistive, non-inductive devices. So I searched to see what they consist of, and unless I am missing something (I worked through only a T section, but see no major change in logic for other circuits), they actually present the pre with a LOWER impedance, the complete opposite to what intuition would have you believe.

So, realistically, there is presumably no practical, cheap, simple way to have a valve pre "see" a SS power as having significantly higher input impedance?

Incidentally, I very strongly suspect that my current Super Micro, with 2 off 12BH7 on the output/line stage, is push-pull. Whatever, a very serious improvement (to my ears) over the Micro with one ECC82.
Maybe, just maybe, I like LESS distortion!!!! LLLOL
 
Do you know what? I followed 90% of that, with ease, first read through. Brilliant, very easy to follow, thanks. I will go through again to get the rest (and then forget it, I have to admit, ashamedly).

My first thought about impedance mismatch (SS power amp too low for valve pre) was "ah - attenuators", thinking that they would be reasonably simple resistive, non-inductive devices. So I searched to see what they consist of, and unless I am missing something (I worked through only a T section, but see no major change in logic for other circuits), they actually present the pre with a LOWER impedance, the complete opposite to what intuition would have you believe.

So, realistically, there is presumably no practical, cheap, simple way to have a valve pre "see" a SS power as having significantly higher input impedance?

Incidentally, I very strongly suspect that my current Super Micro, with 2 off 12BH7 on the output/line stage, is push-pull. Whatever, a very serious improvement (to my ears) over the Micro with one ECC82.

I did simplify it so even you could understand it Vinny:D

Attenuators can be made for any impedance but the most useful in hi fi are usually too low yes.

A transformer could do it but would need to be correctly matched so not very practical. Some valve pres in fact have a transformer at the output to give a low impedance. Some of the expensive Audionotes for example.

Yours is likely a SRPP if it has gain or a White Follower if no gain. Both should drive pretty much any SS power amp fine:)
 
SRPP - shunt regulated push-pull

White follower - a cathode follower circuit named after the man who came up with the idea.

Just for the uninitiated, including me, until I looked them up.
 
Incidentally, I very strongly suspect that my current Super Micro, with 2 off 12BH7 on the output/line stage, is push-pull. Whatever, a very serious improvement (to my ears) over the Micro with one ECC82.

Interesting thread. I went from a Croft 25R:

6843836877_69b1958576_c.jpg


Which as I understand it has 1 ECC83 in the line stage and two in the phono (the other valves on the left of picture are to do with the PSU IIRC, 2x83s and a little Mullard rectifier I think).

To a JC Verdier Control B:

15961345743_f8d955f3a8_c.jpg


Which has 3xECC83 on the phono board (left) and 3xECC81 in the line stage (right). The overall output level can be pre-set with jumpers which is neat, AIUI this is achieved by varying the negative feedback level (I do have a schematic for it but feel it would be unfair to publish without permission). I have it on its lowest setting (which I assume is the highest negative feedback), purely because of level matching and efficient speakers. It just keeps the volume controls in a useable range! The PSU is external and solid-state, just a transformer, bridge rectifier and a couple of caps in a plastic box (I don’t have the uber-fancy add-on valve PSU). I really like this preamp and can’t imagine changing it. It is beautifully simple too, to the degree I can service it myself.

FWIW I felt both were fine driving the Quad 303 (22k), though I do prefer the Verdier overall, which is fair enough as it would have cost a fair bit more (I think they were >€2k, though I paid a small fraction of that used).

PS As an IT-geek it makes me chuckle that the Verdier uses old PC serial port cables for internal connection. I guess they are just as good as any internal wire being a decent solid-core and it makes it fully modular and easy to strip down.
 
^^ Agreed. Basic nuts and bolts stuff is always interesting.

The only hassle with Croft kit is that n+1 circuits appear under any one label/name.

Ignoring any power supply valves (which some people reckon affect sound....) my Micro has two ECC83 in the phono stage and one ECC82 line/output/cathode follower. The Super Micro substitutes two 12BH7 for the ECC82 of the Micro (and changes PS valves as well).

Fabulous phono using an SUT (BTW) - many thanks @Dowser for prompting me on that one.

I have no idea at all of vintages for the two (both bought second-hand), but if anyone wants serial nos. to try some detective work, drop me a PM.
 
So a Quad QC24 will not match a QUAD 909 well?

Nope it's not a good match. 909 is 20K Zin and IIRC the pre uses a very similar topology to the Audionote M Zero and so would likely give maybe 0.7% ish distortion.

Now as I said earlier (and as Tony L noted about Croft 25 with Quad 303) all this doesn't mean you can't use such a combo... just that distortion will be higher than optimum. The THD will increase fairly much in proportion to output voltage and so although at the clipping point of the power amp the pre maybe making 1% THD, at your normal listening level, and especially with high efficiency speakers, it may well be down at 0.2% or less. Obviously power amps with high sensitivity need a lower input level for a given volume and this will have a large bearing on this as well. With a SS power amp of only 20K Zin AND it needs 3V for full output things would be much worse! I'm assuming 1V ish sensitivity in my "guesstimate" distortion figures.

I can't stress enough that in all I've been saying on the matter, I'm generalising, and showing the trends that will happen. All sorts of combinations of valve pre and SS power will be fine and some a real no no with many more "theoretically not optimum but it's close enough for jazz":)
 
Just a thought...……

Glenn makes no valve power amp's, as such. Power amp's with valves certainly...………………………. So what are his pre's aimed at? Series 7 input impedance - 470K.
 
Now as I said earlier (and as Tony L noted about Croft 25 with Quad 303) all this doesn't mean you can't use such a combo... just that distortion will be higher than optimum. The THD will increase fairly much in proportion to output voltage and so although at the clipping point of the power amp the pre maybe making 1% THD, at your normal listening level, and especially with high efficiency speakers, it may well be down at 0.2% or less.

Also worth noting that the Quad 303 is just a half volt input for full output too. My Croft was actually the gain-reduced version, which I bought initially to use with the Leak, but it was still far too twitchy with the 95db Tannoys. It worked far better with the Quad gain-wise. The Verdier on its lowest setting is more usable again, but still too hot for the Leak (which is great with a passive so no issue).
 
I briefly ran an ifi tube buffer between my chord DAC and Bel canto power amp, I find it strange that people talk about the distortion of a tube, all the ifi did was increase space and sort of define highs in their own place. Lovely!
 
I briefly ran an ifi tube buffer between my chord DAC and Bel canto power amp, I find it strange that people talk about the distortion of a tube, all the ifi did was increase space and sort of define highs in their own place. Lovely!

See above. Distortion happens, in the cases here, when you "ask too much" of the valves.
What does the ifi actually do? Output impedance of the ifi? Input impedance of the power amp'?

Given that the ifi is a very recent design, it would be logical to assume that it was designed to be connected to a modern amp' of something like 200K input impedance. Although using logic may be a bit presumptuous.
 


advertisement


Back
Top