• 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.

Tube preamp with current mirror

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi!

I would like to ask the tube experts about the attached tube preamp with current mirror:

How much gain has it?
What is the output impedance?

I would like to use it for a Mosfet power follower amp, so I need a lot of gain.....and nice sound too! :)

Thanks and greets:

Tyimo
 

Attachments

  • hybridpreamp2.gif
    hybridpreamp2.gif
    4.1 KB · Views: 1,298
I don’t think you want to do that! In a current mirror like this, Q4 is configured as a diode. The plate of the 6922, which wants to see a HIGH resistance plate load for lowest distortion and highest gain, will instead see a perversely low and non-linear resistance of a few tens of ohms. Voltage gain to the plate of the 6922 will be fractional (negative dBs of gain – call it “loss”), and distortion high.
 
On top of the previous posters' remarks is the (IMO) negative of a volume control in the O/P circuitry. Also, the I/P impedance of the supplied schematic is VERY low.

I think you can achieve your goal with a constant current source (CCS) loaded triode. CCS loading allows you to get the full mu (amplfiication factor) of the tube along with minimal distortion.

Think about a CCS loaded ECC99 section DC coupled to a power MOSFET source follower. You'd get lots of gain and bulldozer drive capability. "MOSFET Follies" shows how.
 
Tyimo,

The best low voltage tube I'm familiar with is the 6GM8/ECC86. 6GM8 data sheet here.

Given the 6GM8's low max. of 30 V. on the plate, I have doubts about your being able to swing the signal voltage you want.

JMO, if the B+ rail voltage is to be about 60, you will have to get part of the necessary voltage gain from a step up trafo. Plan on having more than 1 FET voltage follower in the final design.
 
Hi everybody!

Chris!
How much drive voltage do you want out??

I need min. 7V RMS output, but 15 would be the best.
max. 1 K output impedance.


Rocky!
http://digilander.libero.it/paeng/projects_frame.htm

Thanks for the tipp, but I know it already. I am afraid it hasn't got enough gain for my power amp. Have You got any infos about it?


Eli Duttman!
The best low voltage tube I'm familiar with is the 6GM8/ECC86

Thanks, I kow the 6GM8, but I have some nice Siemens and Mullard E88CC tubes and I would like to use them.



Greets:

Tyimo
 
I made that circuit back in my student days and it worked as my computer preamp. But at such low voltages, Vgk is very low, and it has limited input range. I tried it to even less than 40volts mostly for experimentation. Soundwise it is very good.
Eli, I'm surprised you dont see how this works. The tube doesn't see any load. Ideally it should be 0ohms into the mirror. The tube has a constant anode voltage. When the grid swings, the cathodecurrent swings almost as much as the tubes transconductance. At 10mA idle, the 6922 will swing almost 10mA for a change of 1volt at the grid. The mirror does just what it is called, mirrors the change in current into the load, the output resistance. and now you can have gains of much more than the tubes mu. The most awesome preamp I made with this topology was with a variable load resistance. (R4 in the circuit was a stepped attenuator).
The output impedance is totally dependent on R4, so Tyimo, hard to get gains of 15 and lower than a few kohms output impedance. But really, have you tried with more than 1kohms? I think most people are so into textbooks that they just assume it must be low. I have tried (real high output impedance) and it sounds great.
 
Hi Semper Fi!

But really, have you tried with more than 1kohms? I think most people are so into textbooks that they just assume it must be low. I have tried (real high output impedance) and it sounds great.

You are right. 1Kohms input imp. is not so important.

I think the attached schematic is also yours. Is it possible "to cut half" and use only one half of the preamp? I don't need balanced pre, only a SE.
What Mosfets did you use for the CCS?

Greets:

Tyimo
 

Attachments

  • tubepr1.jpg
    tubepr1.jpg
    20.4 KB · Views: 790
If you cut it in half you must also cut away the tail part of the long-tail-pair. in other words make it a common cathode stage. So its a totally different circuit. The MOSFETs are IRF610, but any MOSFET will work as a CCS. The anode resistors of only 1500ohms actually works very well for a low gain preamp, but I made this in my student days, and wanted the textbook low output impedance, hence the low values. I now use higher resistances such as 5kohms with two parallell sections of 6922. The resultant output impedance is still an acceptable 2kohms or so. (The tube rp is in parallell with the anode resistor). This requires higher voltages of course.

Btw, I never intended to have 1kohms at the input in the first circuit. That is the default velue of resistors in the program used to draw circuits, and I guess I forgot to select a new value for that part. I usually use 100kohms for HiFi stuff.
 
To follow up in more detail on my earlier post about the first circuit: This circuit is essentially a folded cascode design, where the lower tube sees a very, very low resistance load. Voltage “gain” from grid to plate is therefore almost non-existent, but the second transistor adds gain so that the total circuit produces about 6.3X (16dB) of voltage gain in the end. In effect, a folded cascode takes the plate current and folds it upside down, dumping it into the 3.3K resistor. The trouble with this kind of circuit, IMO, is that it makes the gain depend solely on the triode’s transconductance, gm (and not on its mu). The gm parameter is far more non-linear than mu. Therefore, this kind of circuit will generate far more distortion than would a standard triode gain stage with RL>rp, where instead mu dominates in determining gain (and distortion). Most tube cascode circuits (non-folded) use an identical triode in the upper position, and offer the lower tube at least some non-zero load to work against: (rp+RL)/(mu+1), typically about 1K. In this hybrid circuit, the plate sees only the dynamic (and non-linear) resistance of a silicon diode plus 56 ohms, probably totaling less than 75 ohms.

FWIW, I felt this circuit deserved a quick simulation. These days I can bang a circuit into the SPICE program (I usually use MicroSim) in about 10 minutes. Yes, this circuit does “function” – it makes gain of about 16dB. But simulated distortion with 2 Vrms output was around 0.56%. Most well designed triode line stages can output 2 Vrms with ten to a hundred times less distortion. Predicting distortion from simulations is fraught with uncertainty and inaccuracy, so the measured distortion will be more or less, but the point about this circuit’s relative distortion disadvantage remains valid. Not counting the output pot (which should be on the input and using a higher value), the output resistance is determined primarily by the 3.3K resistor – so let’s call it 3K. Could someone build this circuit and say it sound great? Sure. That always happens. That doesn’t mean you can’t do better. Why bother with tubes if you hamper them?

An alternative is just to use a straightforward common cathode circuit with RL>rp. Gain will be higher with a 6922 this way than in your original circuit. If you want to keep a low B+ supply with the 6922, use a CCS plate supply or a choke in the plate circuit. The output resistance (if you bypass the cathode) will be around 3K, about the same as in your hybrid circuit. If you can afford more supply voltage, there are numerous lower mu tubes that will give you a little less gain but still a reasonable output resistance. I’m thinking of a 5687, or even a 12B4A for less gain. If you want really low voltages, you could use a 6GM8/ECC86, or just live with a decent opamp gain stage like a OPA627.

You need to define how much voltage gain you’ll need, what kind of impedance you’ll be driving and how much ultimate voltage swing it will need. Let us know more about your exact situation and needs, and someone will have a better-tuned suggestion for you.
 
Hi Brian!

Thanks for the nice and "heavy" comment! I am an amateaur and a realy novice in the Tube world. :)


You need to define how much voltage gain you’ll need, what kind of impedance you’ll be driving and how much ultimate voltage swing it will need. Let us know more about your exact situation and needs, and someone will have a better-tuned suggestion for you.

My Mosfet Follower requires a preamp that can swing +/-10-12V.
Output imp. is not so critical, because the Power amp has 250K Zin.
I don't know how to calculate, to define how much voltage gain I need.
If it is possible I would like to use and operate my E88CC tubes under 100V or even near to 60V.

Waiting for your suggestion!
Greets:

Tyimo
 
Yes this is a compromise. Its also a true hybrid where the mirror plays a bigger part than say a CCS loaded tube, where the CCS is SS. Two main reasons why I built this in the first place: 1. I needed a tube circuit that worked with low voltage rails, and didn't afford chokes or interstage trannys at the time. 2. This was something I hadnt seen before, I simply find it boring to build what everyone else builds.
However...I have since built ccs loaded stages, aikido stages, choke loaded and most other types of gain stages. I kep coming back to this mirror gain stage, it just sounds so good. Very live, super dynamic. Perhaps because there is no anode swing that otherwise provides local feedback? This circuit is very wideband, if that may have something to do with it? The distortion may be around 0.5%, but it is mostly second harmonic. And just about every single research done on human perception of distortion finds we need more than 2-3% to even sense it. So no worries. Sure, it can be done better, all circuits can. But try to make a better circuit that runs off 50-60volts. What is better? One person will like one type of sound and someone else likes another. So who is right? Which one is best?
Heres a suggestion for a simple PP MOSFET amp. Run it classA and itll blow you away! R21 finetunes the idle and R15 helps null the offset. It only uses DC feedback, none at frequencies. I find it sounds best that way. Yes it is load dependant, but in the real world hasnt shown any bad sides when listening to music.
 
Since I had the scope available, I did some checks. This with a 10x probe, so the voltage swing is +-9.5V, not 950mV.
Second harmonic is -75dB and third -81dB. And no higher order harmonics at all. Measured with PICO ADC216 PC scope. Is this bad?
 

Attachments

  • mirrorfft.jpg
    mirrorfft.jpg
    35.8 KB · Views: 531
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.