• 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.

#26 pre amp

Coolzero,

You could also try a simple DN2540 / IXCP10M45S cascode ccs on top of the 26 with the output taken from the top of the current set resistor.

Easy and effective with low output impedance.

If you search source follower in the tube forum you will find lots of stuff. Also search Mosfet Follies on the web for some basic circuits. Tubelab also uses a source follower in his Tubelab SE.

Here's a simple circuit. The plate of the 26 would be attached to the FET gate. The FET is an example of one that will work. You can also use the IZXS parts or the DN2540. You want a part with low crss (reverse transfer capacitance). If you direct couple (no cap between plate and gate), you want to make sure that your you have at least 20V or so across the FET. This is to make sure that your crss is low and constant.

The source follower has very good PSRR so you can tap your HV supply at a point where the voltage is higher but maybe noisier.

If that's not convenient, you can cap couple to the FET, but you will need a voltage divider to the gate to bias it, let's say two 500k resistors between B+ and ground.

It's a good idea to include a safety diode for the gate. We can get to that if you want to go down this route.

Thanks Sheldon/mach1. BTW, I have few questions. What would be the most suitable type of FET sound wise? Dep. mode or Enhance mode? Is there any real difference? Will this type of addition/change to my 26 tube pre do any harm to the nice 26 tube sound? What are the pros and cons?

I found the attached schematic on one of the articles. Will the same work if I replace the FET with IRF840 or DN2540 (Both I have on hand)? What should be the Zener to use for protection in this case? Still learning to understand the math behind all this hence appreciate your advice.
 

Attachments

  • Follower.jpg
    Follower.jpg
    24.6 KB · Views: 954
Either device will work. An FET follower such as this will be one of the most transparent active circuits you will find. The signal from the 26 plate will be very accurately followed. The DN250 has lower capacitance, the IRF device has higher transconductance. I'd lean toward the DN250 for low current applications such as this. But I don't know if you would hear a difference.

This circuit will work fine with one change. The source resistor will need to be larger (R5). Let's say that you want bias to 15mA, and you adjust the divider so that the gate is at 100V. The source voltage will be about 3V higher than the gate for the depletion mode device, and about 3V lower for the enhancement mode device. The difference is small compared to 100 volts, so let's ignore it and just say the source will be at the gate voltage for our purposes here. At 100V, a source resistor of about 6.8k will be required to set the current at about 15mA.

Remember that the 26 plate will see the load of the divider resistors in parallel. Not required, but I'd probably double those.

The 2540 gate will tolerate about a 20V swing either side of the source. The IRF FET will be similar. So any zener from 5-15V will work. The zener is connected from gate to source. For the depletion mode device, the zener cathode is on the source side. For the enhancemode device, the cathode is on the drain side.

Sheldon
 
Last edited:
BTW, if you bias with a divider, you probably cannot connect it directly to the B+ supply. The PS ripple will be divided and applied to the gate too.

There is really no need for a pot here, unless you want to fiddle with bias current. So, for example, use a string of three resistors, 100k, 100k, 200k from B+ to common (100k to B+, 200k to common). Connect the gate between the bottom 100k and the 200k resistor. This will set your source at about half the B+ voltage. Connect 5-10uF capacitance from the middle of the two 100k resistors to common. This cap will filter away the PS ripple to the gate.

Sheldon
 
I've been following the thread for awhile and am really inspired. You guys are being super helpful even for general questions here and you don't always see that. I just finished a 417a to 2a3 and then 211 amp with switchable cathode or fixed bias for the output tubes. It's an amazing sounding amp and I am 100% hooked on SET now. That amp isn't mine as I built it for a friend and in the meantime I've been collecting tubes to experiment with.

Today I bought 4pr of type 26 (2ST and 2 globes) and then 2pr of type 112 which are rather similar to type 26 I believe? I like the 211 a lot and the power and dynamics even when using my efficient Klipsch Forte II's at around 98db I believe. Listening at less than 1 watt means THD with this amp is well below 1% and on peaks it can rise up to 25 watts at around 6% THD and in class A2.

I like the idea of matching gain better and don't need the 417a for gain. 26 with a plate choke or a FET there will be perfect. Drivers I will try other than 2a3 are the 45 and 50. I wonder how these tubes will play together?? Isn't that awesome to try combinations that you have no idea if they have been tried before?

The OPT for the 211 is a 10k:8R, so a 35:1 turns ratio. This seems common for a 211 amp, but wouldn't something like 15k:6R be a much better match and load? That's a 50:1 turns ratio and maybe it's that it is too difficult to wind.

BTW I'm going from my DAC into a TX-102 TVC. I can adjust my DAC's output to add just 1db of gain so I would be able to fully drive a low gain input tube for an amp like this. It's a very nice TDA1541A-based DAC - Audial AYA.

Keep the type 26 work coming!! :D
 
BTW, if you bias with a divider, you probably cannot connect it directly to the B+ supply. The PS ripple will be divided and applied to the gate too.

There is really no need for a pot here, unless you want to fiddle with bias current. So, for example, use a string of three resistors, 100k, 100k, 200k from B+ to common (100k to B+, 200k to common). Connect the gate between the bottom 100k and the 200k resistor. This will set your source at about half the B+ voltage. Connect 5-10uF capacitance from the middle of the two 100k resistors to common. This cap will filter away the PS ripple to the gate.

Sheldon

Greatly appreciate your support over this Sheldon. This is very informative and educational. I have redraw the schematic according to the points you have given and attached if you could verify. BTW, what wattage should be the Zener? And the FET's require heat sinks?
 

Attachments

  • FET follower.jpg
    FET follower.jpg
    127 KB · Views: 894
That's the general idea. Couple of comments:

1. C3 is unnecessary

2. You can double your divider resistors. This will allow a smaller C1 (and C2, although 10uf is fine if you've got it). When you calculate the low frequency corner for C1, R1//R2 for the RC calculation.

3. You have plenty of voltage across the FET. You can decrease this if you want to lower the dissipation for that part and burn off the power in the source resistor. Just keep the drain to source at about 20V or better. So you can adjust your divider accordingly. At 15mA and 25V, you would dissipate about 0.4W over the FET. If you have T220 device you can use just a small sink. You'd need to resize R5 to keep your current where you want it.

4. C4 is way overkill. Even with a 5k load, 4.7uF would give you a corner at about 6Hz.

5. Zener will never carry much current, as it will be in series with R3. A 1/4 watt device would be more than enough, though you can use larger if that's what available.

Sheldon
 
Last edited:
That's the general idea. Couple of comments:

1. C3 is unnecessary

2. You can double your divider resistors. This will allow a smaller C1 (and C2, although 10uf is fine if you've got it). When you calculate the low frequency corner for C1, R1//R2 for the RC calculation.

3. You have plenty of voltage across the FET. You can decrease this if you want to lower the dissipation for that part and burn off the power in the source resistor. Just keep the drain to source at about 20V or better. So you can adjust your divider accordingly. At 15mA and 25V, you would dissipate about 0.4W over the FET. If you have T220 device you can use just a small sink. You'd need to resize R5 to keep your current where you want it.

4. C4 is way overkill. Even with a 5k load, 4.7uF would give you a corner at about 6Hz.

5. Zener will never carry much current, as it will be in series with R3. A 1/4 watt device would be more than enough, though you can use larger if that's what available.

Sheldon

Thanks Sheldon, I'll try this and let you know how it goes.
 
Sheldon,

I tried just a simple en. mode FET follower (similar to the schematic you provide). But results were not that impressive to me. Certainly it reduces the distortion but I felt that it reduces the liveliness of the music as well. The instruments I could hear before seems diminished or sound thin. Not sure this is because I had set the FET to run in low current (7ma). Any suggestions you could give?
 
Sheldon,

I tried just a simple en. mode FET follower (similar to the schematic you provide). But results were not that impressive to me. Certainly it reduces the distortion but I felt that it reduces the liveliness of the music as well. The instruments I could hear before seems diminished or sound thin. Not sure this is because I had set the FET to run in low current (7ma). Any suggestions you could give?



If you are comparing it to a heavily loaded 26, then it will have much less distortion, and probably reduced top end. A lot of second and third harmonics will make the sound seem full. You can try higher current, no harm there. Listen for a few sessions before you decide. Make sure you can listed for a long time without tiring.

Sheldon
 
If you are comparing it to a heavily loaded 26, then it will have much less distortion, and probably reduced top end. A lot of second and third harmonics will make the sound seem full. You can try higher current, no harm there. Listen for a few sessions before you decide. Make sure you can listed for a long time without tiring.

Sheldon

Hmm..what I felt was bit confirmed. Top end had lost it's strength Which I didn't like. I personally felt it is one of the strength of 26 tube of having the smoothest and live top end. However FET follower improves the low end a bit. Anyway, I will try with higher current and also the dep. mode FET follower as well. I will post my findings as it progress.
 
Hmm..what I felt was bit confirmed. Top end had lost it's strength Which I didn't like. I personally felt it is one of the strength of 26 tube of having the smoothest and live top end. However FET follower improves the low end a bit. Anyway, I will try with higher current and also the dep. mode FET follower as well. I will post my findings as it progress.

Also may try feeding the #26 with more current. Guessing you may be running it too lean?
 
Hmm..what I felt was bit confirmed. Top end had lost it's strength Which I didn't like. I personally felt it is one of the strength of 26 tube of having the smoothest and live top end. However FET follower improves the low end a bit. Anyway, I will try with higher current and also the dep. mode FET follower as well. I will post my findings as it progress.

I didn't say it very clearly, but I would expect the loaded 26 to have more roll off at the high end. But with more distortion products, the top end could be perceived as more pronounced. In the end it comes down to taste. For me, a certain kind of distortion sounds nice, but over time I don't like it as much.

Sheldon
 
I didn't say it very clearly, but I would expect the loaded 26 to have more roll off at the high end. But with more distortion products, the top end could be perceived as more pronounced. In the end it comes down to taste. For me, a certain kind of distortion sounds nice, but over time I don't like it as much.

Sheldon

Agree with you complete. It is all about the personal preference of the sound. I liked a bit of forward sound and it well mate with my other equipments use, especially with my mellow sounding speakers. That could be the reason the loaded 26 sound is more pleasing to my ears. This may not be true in case the other equipment use with 26 pre is behaving differently I believe. This is why I like DIY. :) You can always get what you like, not someone else prefer. But in the meantime you could experience both the sides in case if you want to compare. Thanks Sheldon for all your education provided. I actually learnt something which I didn’t know much before.
 
Quanghao, thanks for sharing! looking good!

You have Fans on the filament circuit heatsink?

Please post the filament circuit, if you changed anything.

Yes! because i buye heasick from degike, but now not arrive!
It is very hot! but it is exelent, no hum no noise.
in my Amli, I put my handphone to head tube 26, so i can listen some microphonic!