• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Pre amp

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm thinking about building a tube pre amp (actually only a buffer) and I would need some advice from some tube guys.
I was thinking of using a 12AU7 tube with a 250V power supply and biased at 10mA.
Here's the circuit:
<IMG SRC="http://w1.302.telia.com/~u30208505/tubebuf.gif">
Does it seem right? Any suggestions?


[Edited by tsz on 11-12-2001 at 12:10 PM]
I don't use the 12AU7 much, but shooting from the hip, I'd say that it looks okay. You could reduce R3 a bit if you wanted. I see that you've got both the input and output DC blocking tuned to the same frequency. Should you find that the circuit acts oddly, you might want to tune them to different frequencies.

The circuit looks OK. However, a CF is a tricky beast to achieve good sonics, but when on the money, it is marvellous.

YMMV, but I would not run it so hot. I'd select a much lower voltage from plate to cathode, say around 80 volts, and around 40 volts across the cathode resistor. This effectively halves the present B+ to 120V, and greatly reduces current. Doing this will enhance H2, and add that sweet tube sound.

I'd run it at 2.7mA. Thus the cathode resistor should be 15K. There is no telling what the cathode bias resistor should be; start with 220R and work upwards towards about 350R. It might be necessary to use a composite resistor.

You must gun for 120V on the plate, and 40V at the cathode.

BTW, use a 100nF polystyrene for the input cap, and a Hovland or RTX 1uF for the output. Unless you are driving 10K loads, it will not need a 4.7uF output coupler.


Thanks for your reply.
120V sounds great, that would make it a lot easier to design the PSU. :)
Regarding the current, isn't the tube more linear at 10mA? It looks that way in a grid voltage, plate current diagram I found.
Would perhaps the 12AX7 be a better choice? It seems more linear at lower currents.
Any other tube suggestions?

Any tube can be used for a follower, but low to medium mu tubes are prefered; the 12AU7 is a good choice.
Cathode followers seem to have earned a bad reputation with some people. Just for the record, I don't count myself among them. Followers are fine, just make sure you don't set too low an operating current.

Hi Tsz,

You speak of linearity. This goes to the nub of design.

We have here a vacuum tube, which is chosen because it adds a lovely, tubey sound. This is, in fact, largely low order distortion. You like it, and so you want a tube preamp.

Let's not design a preamp which has NO distortion, or, to put it less controversally, NO harmonic enrichment. You want the tube, therefore, to be a little non-linear. Grey is right; don't operate it too low, because then you have too much of a good thing, and it sounds terrible. Just a little.

In light of your heavy 10K input impedance, I suggest increasing B+ to 150V, increasing current to 4mA, and this will probably not much change the cathode bias resistor, but it will change the cathode tail resistor from its present 15K to 18K. Then, the effective AC impedance seen from the CF cathode will be 6.42K with a 10K load, and this is OK. Once again, vary the size of the cathode bias resistor to achieve 72 volts dropped across the 18K resistor. Then the tube will have just under 80 volts between its plate and cathode, and the sound will be just dandy.


I'm always amused when folks immediately assume that people who want tube circuits <i>obviously</i> like distortion. A well-executed tube circuit has low distortion. Period. It just so happens that the distortion that remains is comprised of lower order harmonics. Those who wish to punish themselves with higher order distortion components are welcome to wear hair shirts in order to proclaim their macho tendencies.
The objective is--or shouldn't be--to design a preamp that sounds 'like tubes.' The objective should be to design a preamp that sounds more like music as heard in a concert hall. I'm partial to jazz and classical, but there are other benchmarks that will do as well.
Given the relatively low input impedance of the amplifier, I'd opt for as high a bias current as possible. The voltage won't be a problem, as the 12AU7 is good for over 300V. The trick will be to balance the rail against the plate current and the plate dissipation. If you don't have the specs for the 12AU7 handy, go to http://www.triodeelectronics.com and download whatever information you need. There are also schematics there, some of which might include a 12AU7 as a cathode follower, which might save you some work or, conversely, stimulate your thinking. There are other places, of course, but there's enough stuff there to keep you busy for a while.
When in doubt, play with it. Start with a cathode resistor you know is going to be too high, and substitute lower values. Resistors are cheap. Have fun with it. Watch the limits of the tube. If you're not using a regulated supply, leave yourself some leeway in case the line voltage decides to act up...but bear in mind that tubes are remarkably robust compared to solid state parts; they can take punishment that would instantly toast a transistor. If need be, take a break and go listen to live music so that you'll know what you're shooting for. That's the step most people skip...and it shows.

I too am always amused by the obvjectivists, notably Doug Self, and perhaps also Grey, who loftily shoot for the 'straight wire with gain'.

Grey, I am very glad you are so amused. Clearly you have quite a sense of humour.....

Of course 'well executed tube circuits have low distortion'. Ca va sans dire..... However, a tube is nearly always audible, and the good implementation is patently not highly distortive, but neither is it zero distortion. H2, and to some extent H3, are musical in very small quantities. Hence the truism that many amps measure badly but sound great. Such amps do rather well in the market, too, so there must be something in it.

I agree with Grey that you must build it and listen to it. Not just for ten minutes, but a couple of days.

All this is clear; it's called 'voiceing the circuit'.

I hope this does not start an absurd flame session between the 'euphonic distorters' and the 'straight wire with gain' set. If it does, I'm outta here..........


Egad...that I should be lumped in with Self/Slone and their ilk! I must be living wrong. I trust that those who have been around here for a while got a good chuckle out of that one.
Hugh, it would take a minute or two (or a few thousand...) but if you go back through the multitudes of things I've said here (try the infamous Opti-MOS thread, for starters) you'll find that I'm clearly:
A) <i>Very</i> favorably disposed towards tubes. I've got a tube preamp (Conrad Johnson Premier III), and a pair of 130W monoblock tube amps of my own design. That's not to say that I don't like solid state, as I've got several representatives of that species, as well. But solid state and tubes tend to do different things well. One of my current projects is a balanced tube phono stage, but I ran into a shielding problem. I'll get there.
B) Not obsessed with distortion measurements, having found that they correlate poorly (verging on not-at-all) with sound quality.
That said, I don't hold with designing tube gear with an eye towards having any more distortion than necessary. Other factors being equal--as though that would ever happen in the real world--lower distortion is good. In that sense, yes I'm interested in a straight wire with gain. Given my druthers, I'll take lower order distortion products, which even the specs uber alles tribe admit are less obnoxious.
More subtle points intrude, however. Why is it that tube gear nearly always images better than solid state? I dearly love a good solid image, so that tends to lead me to tube front ends and amps for the midrange. Generally solid state does a better job at low frequencies--at least that makes sense, due to damping factors, etc. The high end is up for grabs. I've heard good high frequency reproduction out of both tubes and solid state.
Hope this clears my kharma...


Thanks for your post; gracious, a measure of the man no doubt.

I think we agree; there is no finer voltage amplifier than a triode, and no better current amplifier than a bipolar transistor. The trick is to use hybrid, to exploit both.

Distortion is spectral. Its perception is psycho-acoustic. To narrow it to a THD or IMD figure is off target, like IQ. High order sucks, Low order can be quite nice in small quantities. Transistors when switching abruptly produce a spray of the former; tubes, when gracefully handing over the musical baton, produce almost no artefacts, and thus sound great in Class AB.

I believe we should design tube gear with just the right amount of low order distortion. It sounds more 'natural', in the same way that no one goes on TV without makeup. Here's my take on this..... (ready???)

Microphones are highly directional, even the omnidirectiona ones. What is not understood well is that they are not particularly linear, and often register different frequencies erratically. Put another way, they can de-emphasise certain frequencies, just as speakers do. Inevitably, this changes the spectral distribution (cf. harmonic structure) of the sound, and our ears hear it. It comes across as slightly dry, a little lean, lacking in fullness, or emotional communication, if you will.

The vacuum tube amplifier adds a little H2 and H3. These are both musical, particularly H2. This adds in some of the harmonic richness lost during recording. Ergo, you hear something which sounds more 'natural', even though the harmonic additions must surely differ from the original.

This theory is reflected in the wide variability of modern recordings. Mike placement is crucial, and the type of microphone is important; it is very significant, in my view, that the best of them are the German Neumanns, which all feature an integrated vacuum tube.

So that is the reason I like my vacuum tube circuits to add a little distortion. I like it, my customers like it, and I gun for it. Not in big quantities, mind you, but just enough that people say, 'Wow, that sounds really warm and emotional!'

Your kharma is fine.......



PS: Grey, your fully balanced phono stage will cancel any even order distortion, allowing only odd order to pass through. A nice, symmetrical clip is a thing of beauty on the CRO, but an excrescence to the ear. Why not go for a single ended phono design, like the OzLoesch, a design Mick Maloney and I put together a couple of years back. You can see it here: http://www.cantech.net.au/~supra/preamps.html
Naming equipment after wines? You're making me thirsty.
As for the balanced phono...because it was there. A phono cartridge, by its very nature, makes an excellent case for using balanced circuitry for noise cancellation, etc. Better still, you get a 'free' 6 dB of gain.
If it clips, I need to go back to the design phase. That's what the R&D part of the whole deal is for. If I'm still so stupid as to let the thing misbehave, then I'll pay the price with my ears, but I'm hoping that I'll be able to avoid that.
If all else fails, I can always do a single-ended circuit.

Hello tsz, Hugh, Grey,

just my 3 cents :)


the 12AU7 / ECC82 is a wonderfully sounding audio tube, it was designed for audio frequencies and i heard it in several fine sounding amps,


looking at the schematic, i would replace R2 by a solidstate constant current source (see Valley/Wallman, Vacuum tube amplifers, p438) and see that the CCS runs at 10.5 mA. My Ratheiser tube manual proves that you were right, tsz, for B+ @250 V and bias @ -8.5V. Ip is at 10.5mA; for B+ @100 V and bias @ 0V Ip is 11.8mA; 120V still will not have much bias for around 10mA of Ip.

The CCS makes the CF work more linear and to my ears, this is clearly audible.

Allen Wright always recommends to run a tube with lots of Ip; he got some experience. And, recently another buddy of mine demoed me how a tube sounds with low Ip and high Ip. High Ip sounds fast, vivid, juicy, low Ip sounds beautiful but lame. Not my taste.
(for the record, with high Ip i do not mean the plate glows cherry-red :) )

I would go with B+ @250 and bias @ -8.5V as then you have heaps of headroom. True, a CDP outputs 2V RMS max, this is 5.65 V p-p. So a bias of -3.4 should be enough. But more headroom always is better and what if you want to use a RIAA phono preamp one day which has 1V RMS for 0db, (and 7.1V RMS @+17dB which can happen on an LP) then you would need bias @-8.5V for the CF to work properly.

hope then you have not a Decca London cartridge as this monster has 8.5mV @5cm/sec instead of the usual maximum of 5mV, and then your heardoom is not enough

Consider, these are worst case considerations: volume control full open.

Methinks, the usual opinion, a CF is an impedance follower and will follow any signal provided it doesn't bounce into the B+ limit, is not thought to the end. A CF used like this will draw grid current for very short time if the grid tries to climb over the cathode. Nasty dynamic behaviour will occur although the tube appears to work ok when a steady signal is displayed on the scope. Grid current mode never sounded pleasant to my ears.


please do not use junk tubes. Use Valvo, Siemens, Telefunken, Philips/Amperex "bugle boy" if you can get your hands on them, but a General Electric 5814A also sounds very fine and should be easy to obtain.

tsz, maybe you find it interesting to cross-check the differing philosophies and report about changes :) mehopes, i did not stir up too much after you finished your prototype.
If you're going to get into working with tubes,you might want to get a copy of TubeCAD,which is a great program for working with some basic tube circuits.You would be able to tweak this circuit virtually and probably save some time and frustration.Another great source is Bruce Rozenblit's book, "Beginner's Guide To Tube Audio Design".Lots of good stuff in there and very helpful.
E82CC go for it !

hello every body,

i do agree with all af you : a sufficient current draw, is'n it

tsz what you got to do now it try to elainment the circuit,
change the bias around 4mA until 10mA, if your amplifier have a lower input impedance ussually it will suitable with a much more current draw, and remember to keep grid bias in a sufficient headroom too.

you could do much more simpler by software program as john sugesstion, i have try the software and its userfriendly

and i would like to hear the result,
do you agree all folks?
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.