26 preamp

[w00t]

Hi all,

I'm interested in building the 26 preamp found on VT52.com but am not sure about some things. The simplicity of the circuit both attracts me and scares me.

http://www.vt52.com/diy/myprojects/amps/26bat/26bat.htm

1. There is no value stated on the input pot, so I will be using a 100K stepped attenuator i have lying around. Is this fine?

2. There are 2 preamp stages running parallel to each other. I am only interested in the 26 bit and the comments on the webby says that the 2C51 bit is to be used for LF work, so will I still get full range sound if I just remove the 2C51 bit and just build the 26 bit?

3. I have a 330-0-330 80mA tranformer lying around (leftover from previous projects) and would like to know if it will be able to meet the power requirements of this project. I'll only be building the 26 bit. I'm quite sure it would, but seeing that vt52.com used a 100mA spec-ed tranny for EACH channel makes me feel small. I'll be modelling in PSUD to adjust for the right B+.

4. The schematic lists the current draw for each part of the preamp to be 8mA (so total current draw = 16mA?), so I presume I can use just one of my 330-0-330 80mA transformer to power 2 channels? My 5V transformer will be separate (my current tranny only has 6.3-CT-0 windings ) and I'll probably go for GZ34, 5Y3 or 5U4. Maybe I'll try my luck with some mercury vapour 83.

5. Would the 1:1 line stage transformer do the job properly? Having seen some other with output transformer preamps with different winding ratios makes me feel uneasy. Tranformers aren't cheap.


I'm hoping the simplicity of this project will let me use high end components without burning too big a hole in me pocket... and contribute towards the transformer cost. Currently thinking of using 6N OCC silver wires and Kiwame resistors from AHF. (http://www.ahfartaudio.com/components/resistors.html)

edit:
On another note, how do you guys think a RF remote control would do for this preamp? Would it interfere with the audio signal? Probably getting some generic TX/RX RF modules operating in the range of around 300-400MHz to control power, volume and input selection. IR requires me raising the remote and isn't really much simpler than RF to build, though part sourcing is much easier and costs lower.

Thank you for reading.

 

[dsavitsk]

1. There is no value stated on the input pot, so I will be using a 100K stepped attenuator i have lying around. Is this fine?

Should be.

2. There are 2 preamp stages running parallel to each other. I am only interested in the 26 bit and the comments on the webby says that the 2C51 bit is to be used for LF work, so will I still get full range sound if I just remove the 2C51 bit and just build the 26 bit?

Yes

3. I have a 330-0-330 80mA tranformer lying around (leftover from previous projects) and would like to know if it will be able to meet the power requirements of this project.

Yes

4. The schematic lists the current draw for each part of the preamp to be 8mA (so total current draw = 16mA?), so I presume I can use just one of my 330-0-330 80mA transformer to power 2 channels?

Yes

Maybe I'll try my luck with some mercury vapour 83.

Sounds like a bad idea to me.

5. Would the 1:1 line stage transformer do the job properly? Having seen some other with output transformer preamps with different winding ratios makes me feel uneasy. Tranformers aren't cheap.

No, they aren't. but a 1:1 is a bad choice here unless your next stage has a very high input impedance. 15k:600 is more standard, though you'll lose your gain -- maybe use a step up transformer on the input. If you need a lot of gain, though, the 26 may not be the best choice.

I'm hoping the simplicity of this project will let me use high end components without burning too big a hole in me pocket... and contribute towards the transformer cost. Currently thinking of using 6N OCC silver wires and Kiwame resistors from AHF.

I'd get it working first, then use good parts. 26's are notoriously tricky. Check out Edcor for some cheap transformers. once it works, then go for the expensive ones.

 

[Burnedfingers]

I would suggest that if you haven't built many projects to leave the 26 alone because its a humming bast**d. Check out kevinKr because he has made a very nice 26 that is very workable.

dsavitsk suggested not using the 83 and he is 100 correct on that.

 

[316a]

I would suggest that if you haven't built many projects to leave the 26 alone because its a humming bast**d. Check out kevinKr because he has made a very nice 26 that is very workable.

dsavitsk suggested not using the 83 and he is 100 correct on that.

DC filaments kill the hum at a stroke . It's the microphony I found to be a problem with the 26 , backing off the filament voltage slightly can help here . Also if the chassis rings , the valve rings with it so the chassis should be damped and braced .

cheers

316a

 

[arnoldc]

Try Kevin's version of the 26 preamp.

 

[w00t]

I would suggest that if you haven't built many projects to leave the 26 alone because its a humming bast**d. Check out kevinKr because he has made a very nice 26 that is very workable.

dsavitsk suggested not using the 83 and he is 100 correct on that.

I've seen kevin's 26 preamp but was put off by its complexity and the need for a tube reg supply. Maybe I'll just CLCLC to get the right B+. The use of a capacitor between the grid and pot is also putting me off. Can I remove this cap safely?

DC filaments kill the hum at a stroke . It's the microphony I found to be a problem with the 26 , backing off the filament voltage slightly can help here . Also if the chassis rings , the valve rings with it so the chassis should be damped and braced .

cheers

316a

Yep. I'm definitely using a voltage then current regulated supply for the filament. Only thing I'm really worried bout is the microphonics.

Try Kevin's version of the 26 preamp.

I'm currently looking at it... is there anyway to do a fixed bias version?


Thanks!

 

[arnoldc]

Hi, Kevin's pre *is* fixed biased.

 

[w00t]

Hi, Kevin's pre *is* fixed biased.

http://www.kta-hifi.net/projects/pre_page/26DHTPre/26dhtpre-1.pdf
http://www.positive-feedback.com/Issue16/kennedy.htm

Isn't R1/R4 for biasing?

 

[kevinkr]

http://www.kta-hifi.net/projects/pre_page/26DHTPre/26dhtpre-1.pdf
http://www.positive-feedback.com/Issue16/kennedy.htm

Isn't R1/R4 for biasing?

I think you are confusing fixed bias with cathode bias. This is a fixed bias design and all that really means is that the bias voltage is fixed externally and not generated by the current flow through the tube itself. This has the significant benefit of allowing you to easily tune operating point to the transformers, tubes and supply voltage you are using.

You can't remove the coupling capacitor on the grid in this design because it sits at about -9V relative to ground. You can use a 9V battery to supply your bias. (You can omit the pots and associated circuitry in such a case, just connect the - end of the battery to the grid resistor, + end to gnd.) Cathode bias requires a large cathode bypass capacitor and separate filament supplies for each channel. Good cathode bypass caps (they are very audible in this sort of application) are quite costly and take up a lot of space. A good film cap in the high impedance grid circuit will handily outperform a not so mediocre cathode bypass cap.

Your supply transformer is barely workable, the problem is that you have to drop about 250V prior to the 26 plate to line transformer. The ideal operating point for the 26 is around 6mA and 150V (or somewhat less) This transformer would be ideal for use with the regulator circuit however.

You can use unregulated plate and bias voltages, and although the design does look complex it is actually pretty simple to get working.

Very quiet supplies are a must, and the filament supply must be ripple free dc.

Magnetic and electrostatic coupling to the 26 is a big issue, shields around the tubes are a must.

I recommend a separate chassis for the power supply to keep magnetic fields away from the plate to line transformers. It is very difficult to eliminate all coupling, but an over size chassis can help in the case where a single chassis is desired. Use aluminum to minimize magnetically induced currents in the chassis.

Heavy, solid construction is a must, and mounting the 26 tube sockets on a suspended subchassis while not absolutely necessary does help. Mine are mounted directly on the chassis which is mass loaded with a lot of material. (resilient floor tile works well to do this)

All of that said, I was most sceptical about this design when I first started working on it, over the course of years it has supplanted every other line stage I have designed. Still evolving it continues to get better.

I would be happy to advise simplification where possible. The earlier version uses a simpler B+ supply using a single 6EM7 - rectification can be tube or SS. I can furnish a schematic on fairly short order i f you want one. Battery bias will work fine.

Output transformer quality is paramount to the end result, spend the most money here, everything else can be upgraded down the road. (15K primary Z is the minimum I would recommend)

The line stage is used with a 300B SE amplifier with about 24dB of gain into a speaker system that with room reinforcement is exceeding 100dBspl/w/m efficiency and is extremely quiet. (Line stage gain is about 4dB - all I need, but with an output Z of only 300 ohms allowing me to use unshielded silver wire in my interconnect cables.)

 

[w00t]

I think you are confusing fixed bias with cathode bias. This is a fixed bias design and all that really means is that the bias voltage is fixed externally and not generated by the current flow through the tube itself. This has the significant benefit of allowing you to easily tune operating point to the transformers, tubes and supply voltage you are using.

You can't remove the coupling capacitor on the grid in this design because it sits at about -9V relative to ground. You can use a 9V battery to supply your bias. (You can omit the pots and associated circuitry in such a case, just connect the - end of the battery to the grid resistor, + end to gnd.) Cathode bias requires a large cathode bypass capacitor and separate filament supplies for each channel. Good cathode bypass caps (they are very audible in this sort of application) are quite costly and take up a lot of space. A good film cap in the high impedance grid circuit will handily outperform a not so mediocre cathode bypass cap.

Your supply transformer is barely workable, the problem is that you have to drop about 250V prior to the 26 plate to line transformer. The ideal operating point for the 26 is around 6mA and 150V (or somewhat less) This transformer would be ideal for use with the regulator circuit however.

You can use unregulated plate and bias voltages, and although the design does look complex it is actually pretty simple to get working.

Very quiet supplies are a must, and the filament supply must be ripple free dc.

Magnetic and electrostatic coupling to the 26 is a big issue, shields around the tubes are a must.

I recommend a separate chassis for the power supply to keep magnetic fields away from the plate to line transformers. It is very difficult to eliminate all coupling, but an over size chassis can help in the case where a single chassis is desired. Use aluminum to minimize magnetically induced currents in the chassis.

Heavy, solid construction is a must, and mounting the 26 tube sockets on a suspended subchassis while not absolutely necessary does help. Mine are mounted directly on the chassis which is mass loaded with a lot of material. (resilient floor tile works well to do this)

All of that said, I was most sceptical about this design when I first started working on it, over the course of years it has supplanted every other line stage I have designed. Still evolving it continues to get better.

I would be happy to advise simplification where possible. The earlier version uses a simpler B+ supply using a single 6EM7 - rectification can be tube or SS. I can furnish a schematic on fairly short order i f you want one. Battery bias will work fine.

Output transformer quality is paramount to the end result, spend the most money here, everything else can be upgraded down the road. (15K primary Z is the minimum I would recommend)

The line stage is used with a 300B SE amplifier with about 24dB of gain into a speaker system that with room reinforcement is exceeding 100dBspl/w/m efficiency and is extremely quiet. (Line stage gain is about 4dB - all I need, but with an output Z of only 300 ohms allowing me to use unshielded silver wire in my interconnect cables.)

Hi,

Thank you for clarifying my confusion.

With regards to the cathode bypass caps, what value would be optimum? I'm thinking of feeding the filaments of each tube constant current. (LM317? Or a higher powered version)

For my transformer, I'm actually thinking of using some high H chokes (with rather high DCR) in multiple LC configurations to drop the voltage, as compensation for not using a regulator circuit. As much as I like seeing tubes in action, I'm trying to not use additional tubes due to the year-long hot weather here in Singapore. Is this advisable?

As for chassis design, I'm thinking of internally sub-dividing a main chassis into 2 separate suspended subframe, shielding these 2 sections (power and signal) by means of copper or aluminium plates. The main chassis will house the subframes to give an overall view of a single chassis. The remote control circuit would probably be mounted directly on the main chassis.

Thank you!

 

[arnoldc]

@W00t,

I also have a 26 preamp, and while it is not an all-out, "statement" design like Kevin's, it does sound great. I am posting the picture of the internals so you'll have an idea.

An externally hosted image should be here but it was not working when we last tested it.

I will emphasize again what was posted before- microphonics. Since this preamp was made like an ordinary tube project, and since it sits on an acrylic shelf

An externally hosted image should be here but it was not working when we last tested it.

It's susceptible to external interference, like myself plugging in an interconnect wire!

This is a usual cathode bias design and uses LED (no cathode capacitor isssues to worry about as Kevin mentioned). But since LEDs are happy with some 10mA flowing through it, and the 26 requires only 6mA, there is augmentation to make it so. The plate resistor is those two Riken, and cap coupled to output using Auricap.

The power supply is solid state, with "brute force" filtering using those large electrolytics. The filament is single stage, current regulated.

I will be making Kevin's pre amp, but with 01A as the tube, as I would like a statement DHT preamp and I have huge respect for Kevin's designs.

Cheers!

 

[garbage]

dsavitsk suggested not using the 83 and he is 100 correct on that.

is the 83 that bad a rectifier?

 

[kevinkr]

Hey Arnold,
Thank you for the kind words, it is nice to hear when someone appreciates the work you do..

is the 83 that bad a rectifier?

The issue is that the 83 contains mercury.. Electrically other than the need to preheat and deal with the RF noise issues it is a pretty nice rectifier. They're also kind of pretty to look at..

One major caveat is that you should never use mercury rectifiers made with quartz glass as it transmits the UV generated with little attenuation which is a great hazard to vision.

The other extremely serious problem with mercury rectifiers involves the risk of breakage, should you break one you now have a "mini superfund" environmental disaster on your hands. It is extremely toxic and readily absorbed through the skin, it's also impossible to get out of carpeting and probably wood flooring as well. (Costly professional cleanup in such a case is advised.)

I have used them in the past, but I don't recommend them for the above reasons. It's your choice though, many do still use them in audio amplifiers. Make sure that you will never break one - a very tall order.

 

[garbage]

hi kevin

thanks for the reply.

i'm aware of the issues involved when using the 83 or other mercury rects, however i'm not aware about the one on quartz glass.

how can i tell if it is not quartz? does RCA manufacture their 83 in quartz?

btw, there surely are advantages to using the mercury rectifiers? the bass response is suppose to be better, but in my d3a/300b DRD configuration, i cannot tell if it is due to the choke or the rectifier.

 

[Salas]

The 83 sounds tuneful and open, third best sound ever, after AZ1(pic) and WE274B IMHO.

 

[Burnedfingers]

The AZ1 is a nice looking tube. I have seen several in the last 10 years or so but haven't ever come across a base for one. Can't say I have ever heard a sonic difference between different rectifier tubes with the rest of the supply staying the same. I would suggest staying away from the 83 tube. Mercury is not our friend.

 

[arnoldc]

Hi Kevin, I just bought the filament regulator parts. Be prepared for a lot of questions

Here's the first one-

1. I don't have those UTC transformers. Can I use choke load (I have a 5000H plate choke) plus output capacitor and still use 150V for the 26?

 

[dsavitsk]

Can I use choke load (I have a 5000H plate choke) plus output capacitor and still use 150V for the 26?

Shouldn't matter. You might at some point pick up a pair of parafeed transformers and add them, unless you need the gain.

Also, keep in mind that the output impedance is on the order of 7K5 with that setup. The down side of this is that your next stage should probably have an input Z of 100K or so. The upside is that you can probably get away with a 0.47uF coupling cap.

 

[arnoldc]

dsavitsk, excellent. My amp has 100K at the input and I already have 0.47 coupling caps



The hard part will be the voltage regulator! Can this be powered up without a load?

 

[kevinkr]

Hi Arnold,
You can load the filament regulators with 1 ohm resistors, that should put you in the right range to test the filament regulators without jeopardizing your 26's.

The choke load will work fine with all of the caveats Doug noted in his response. Later you can convert to parafeed or conventional plate to line transformers if you so choose.

When using chokes or transformers on the output it is very important to make sure that there are no near by sources of ac magnetic fields. (plate power, filament transformers or mains wiring) Some provision for magnetic shielding should it be needed is a good idea. (Leave a little space around the chokes basically.)