• 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

diyAudio Chief Moderator
Joined 2002
Paid Member
Disco's gonna test a 1.5 Vout Mosfet CCS & BJT shunt reg set at 1.2A for #26. 1A available to the heater, 200mA to burn all time at the shunt transistor. I arranged it so it can utilize a 6V3 secondary tap for rectification, so it can work from 6.3-9V DC input. It has a trimmer too. It shows like working on the simulator, I just hope its gonna work properly in real life too, the output voltage you need here for heating up the 26 its just too low and it was a challenge to set it right and find the right transistor even in simulation. If we get good working order and good tone results will let you know. Don't know when he will do it.
 
Hi Sheldon & Rod,

Your circuit diagram is wrong. You show a capacitor in series with the supply output. This will block DC. A feedthrough capacitor would be represented by a small cap from the output to the chassis, at the point where the supply passes through the chassis. A feed through capacitor is a wire that is surrounded by a layer (one or more) of dielectric that shunts RF from the wire to the chassis. DC current passes through the wire from one side of the chassis to the other.

FEeedthough caps are very effective at frequencies that leaded parts can't do much with |(>10MHz), but without a metal enclosure the HF/VHF/UHF may couple electromagnetically from the source of noise (eg rectifier recovery pulses) to the low level audio circuits.

Running the power from the trafo/rec/caps to the preamp via shielded cable can help, when a metal enclosure is hard to do.

Sorry about the delay in reply. Thank you for the corrections and suggestions. I did some hunting in the shed and managed to find a metal enclosure to put the 2 transformers and 2 caps. At the moment they are all laid out on wood (Fig 1).

I’m still waiting for parts to build Rod’s filament psu, so went ahead and built a psu with 7805 regulator and 1086 ccs, sort of like Kevin’s suggestion on his website.
 

Attachments

  • Fig1. Filament psu1.jpg
    Fig1. Filament psu1.jpg
    160.3 KB · Views: 1,611
My 26 preamp

I now have a 3 chassis 26 preamp.

This weekend I finally managed to finish the pre and test it out. One can note the Hammond chokes on the pre under construction (Fig 2). I ended up putting the chokes in an Alu box to shield them, also put some Arcturus tubes in a finished version to make it look sexy (Fig 3).

Voltages were all good, no smoke but I do have this bloody low frequency hum and some hissing noise. I officially feel like I’m part of the 26 problems club. A view of the wiring is also attached (Fig 4).

The B+ psu chassis and filament psu chassis are about a meter away from the actual preamp
I plan on trying the following:
1) Shielding and grounding the umbilical cables a) joining mains 240V from main psu chassis to filament psu and b) from main B+ psu chassis to preamp chassis (Rod's suggestions)
2) shielding the filament wiring in the preamp chassis and grounding it
3) increasing the cathode bypass cap
4) ground the alu box with chokes
5) shield the tubes and ground that (Kevin's suggestions)

Any other suggestions would be most welcome.

Cheers,

Rich
 

Attachments

  • Fig 2. Preamp in progress 1.jpg
    Fig 2. Preamp in progress 1.jpg
    44.4 KB · Views: 1,510
  • Fig 3. Preamp with tubes1.jpg
    Fig 3. Preamp with tubes1.jpg
    165.9 KB · Views: 1,468
  • Fig 4. Preamp inferior 1.jpg
    Fig 4. Preamp inferior 1.jpg
    175.2 KB · Views: 1,395
Last edited:

iko

Ex-Moderator
Joined 2008
Disco's gonna test a 1.5 Vout Mosfet CCS & BJT shunt reg set at 1.2A for #26. 1A available to the heater, 200mA to burn all time at the shunt transistor. I arranged it so it can utilize a 6V3 secondary tap for rectification, so it can work from 6.3-9V DC input. It has a trimmer too. It shows like working on the simulator, I just hope its gonna work properly in real life too, the output voltage you need here for heating up the 26 its just too low and it was a challenge to set it right and find the right transistor even in simulation. If we get good working order and good tone results will let you know. Don't know when he will do it.

Hey salas, good idea. Something to consider, can the bjt take the full 1.2A in case the filament dies?
 
Administrator
Joined 2004
Paid Member
A schematic of the preamp and filament supply is attached below. For those interested there is a description of the power supply in post 90.

Rich out

My one recommendation (unsolicited I'll admit) would be to change out that 7805 for a 7806 (Rohm amongst others) because the LT1086/5/4 etc perform significantly better with >3V across the device itself. These are low drop-out devices but that does not mean their performance does not degrade in subtle ways as you approach the drop out voltage, here the headroom is only 2.3V which is not necessarily enough to assure that all internal circuitry is functioning close to specification. Increasing the voltage across the LT1086 to 3.3V or so puts the device close to the point of diminishing returns as a CCS and is probably good enough - hence the recommendation for a 7806. (Alternately you could stick two 1N4148 in series with the 7805 gnd pin making it in effect a 6.2V regulator which perhaps would be slightly better from a theoretical standpoint. )
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Its alive!

I am just briefed it works right from the start, no noise no hum when on the #26 heater. Disco has to listen in a couple of days in his proper system out of the lab for qualities, let it show if it will present any kind of issues or if its reliable too. He also has a current source special reg so its fortunate that he will be able to compare voltage and current methods at a point, beyond chip solutions.

In the pictures you can see the simple shunt, the trafo and the simple rectification and filtering, 1.5V trimmer set when about 1A was on the load and Vbe was 0.66V on the bootstrapped one Vbe Mosfet CCS controller BJT. 1.18A was running in total. The cables may look more than needed, but remote node sensing is included since heater cabling is always rather long.

''There's millivolt wandering (0,002V max) within 5 minutes, I guess because of parts heating'', ''The old 120MHz scope did not show anything more than its own residual noise'' Disco said to me for now.
 

Attachments

  • 1.5Vreg-foto1.jpg
    1.5Vreg-foto1.jpg
    97.9 KB · Views: 1,435
  • 1.5Vreg-foto2.jpg
    1.5Vreg-foto2.jpg
    95.4 KB · Views: 1,395
  • 1.5Vreg-foto3.jpg
    1.5Vreg-foto3.jpg
    97.6 KB · Views: 1,342
I now have a 3 chassis 26 preamp.

Voltages were all good, no smoke but I do have this bloody low frequency hum and some hissing noise. I officially feel like I’m part of the 26 problems club. A view of the wiring is also attached (Fig 4).

Hum could be because of the way those chokes were mounted close to the 26 tubes. I had a similar issue and it went away when I fix the chokes at normal standing position little away from the 26 tubes. I tried various positions of the chokes (even the way you fixed) but normal vertical standing got rid of this low frequency hum and hiss.
 
Here's some numbers from my latest attempt to get all the voltages right.

PSU HT = 365vDC
Dropper resistor to glow tubes - 3.2K, 9W
Glow tubes - 150v and 105v total 255v
IXYS cascode active load - sense resistor 470R
Anode voltage 178v
Cathode voltage 11.2v
Current 6mA
Cathode resistor 1.85K (2.4K // with 8.2K) wirewound

Seems to sound pretty good - full and detailed.

andy

Andy,

I adjusted my 26 pre with almost similar parameters. Sound fantastic as you said. Big difference from standard 135V-5.5ma config.
 
Hi Coolzero,

Thank you for the suggestion, I’ll make the mods this weekend. I think you may have hit the nail on the head, it may explain the low hum during the testing stage. Also a lot of the other 26 preamps I’ve seen have a lot of space around that tube.

Martyh,

I’m not sure of that method of battery biasing, but as I understand it, one needs to have the potential difference (grid voltage) between the grid and the cathode. Have a look at Kevin’s write up on his 26 preamp there is a schematic of battery biasing there. Also, I have attached a schematic of a battery biased 01A-01A-71A amp by one of the forum members Joel. He uses a neat way of biasing. I hope it helps.

Cheers,

Rich
 

Attachments

  • Joel's 01A c_load 71A SET.png
    Joel's 01A c_load 71A SET.png
    17.1 KB · Views: 1,266
Administrator
Joined 2004
Paid Member
Yes. Make sure the battery can handle the current.

NiCads are the best sounding in this cathode connected fixed bias application, but that said I actually don't like the way they sound. Forget NiMHs altogether.

I have also tried series battery bias in the grid circuit (battery - to grid, battery + to volume pot) but this does not sound all that great with the typical alkaline battery (-9V or so required) and needs to be bypassed with a good quality cap.

I recommend using conventional fixed grid bias on the grid instead and this can be provided by a 9V alkaline battery (+ to ground) and a ~100K grid resistor - of course a coupling cap is required and this should be of high quality.

Sometimes the simplest circuit approach is not the best - I have found many instances where placing a battery in the audio path sounded no better than any alternative and often worse. It is however very dependent on the quality of the battery, and some are much better than others. (I tried a few) I found the Sanyo Cadnica series batteries sonically superior to anything else I tried. (Their internal resistance is lower than most competitors) Note that for cathode circuit battery bias the battery should be rated such that in normal operation the trickle or floating charge current rating is not exceeded. (Typically no more than 5% or so of the MaH rating of the battery)

Note that in the 26 application you would have to stack the equivalent of 6 - 7 cells in series to get the right voltage, and IMO this will sound worse than a comparable number of LEDs in series or my preferred grid bias arrangement.