Tube amp power supply test build

[Kennyg2209]

I have my power supply set up on the test bench. I think I'm on the right track. I'm hoping the experts will take a look and see. I think the current draw with two 2A3s and two 6sn7s will be between 120 and 132ma. As you can see my transformer voltage on my schematic is a little high so I need to drop it some. I put the 75 ohm resistor to limit the current flow and drop the voltage some going into the first caps. Plus I had it and I'm on a tight budget. I've been trying to use components that I have . I don't have a cap after that because it seemed to raise the voltage too much and I'm trying to use 450v caps that I have. I went with the two smaller chokes that another AK member suggested. I've read a lot of other theads when used with good luck. And there cheap . The next drop resistors to get me close to below the max voltage for the 2A3s. They'll be two more drops for the 6sn7s. My main concern is my capacitor values. If someone knows better values to use, please let me know .

I found a good procedure for calculating the ripple on the B+ from Elpaso Tube amps. I used 100ma load for now. My load will probably be a little more than that and bring the B+ down to a good voltage. I can always change my drop resistors some if needed. I also used Duncan amp psu2 and came up with the same results as on the bench.

If you see in the one photo that my voltage and ripple(28.5) are a little lower. In that photo I was connected to an isolation transformer and my line voltage was a little below 120v. The other photo, no isolation .

Could everyone take a look and tell your thoughts.

Are the capacitor values okay?
Is 29.3mv of ripple good enough for 315v B+?

Any help is greatly appreciated, Thanks!


Kennyg

 

[mroboto]

I would not put a resistor and choke between the rectifier and the first cap, you risk hum, I am not an expert, but this is my first thought when I see your diagram.

Single ended tube is sensitive to ripple.

 

[mroboto]

You can series two 47uf 450V to have ≈ 23uf @ 900V

Then put resistor and choke or choke and resistor and choke after that.

 

[Kennyg2209]

You can series two 47uf 450V to have ≈ 23uf @ 900V

Then put resistor and choke or choke and resistor and choke after that.

I kinda thought of something like that. I would need to parallel resistors with the series capacitors at 50/C , I think. Maybe I'll give that a try tomorrow and see how it changes. Does 29mv of ripple seem like to much? Is the ripple not the only thing that can produce hum? Thanks for your reply!


Kennyg

 

[mroboto]

I don't know, I never made measurements like that.

But as I added to my first reply, "Single ended tube is sensitive to ripple" , much more than Push Pull , this is the reason why choke is not that much needed in PP amps while always present on SE amps.

 

[kward]

Those chokes would need to be used with a cap input power supply. They are not spec'd appropriately for a choke input supply. (Cap input means the cap comes first after the rectifier.) With a 380-0-380 secondary and cap input, solution is perhaps a bit cumbersome but the overall approach will work as you've shown. You should use at least 450V rated caps, and at that, it will be pushing them right to their voltage limit. I'd opt for higher rated caps or connect caps in series with voltage equalizing resistors as shown below.

There's another approach you could take. Use an active device to drop the voltage. Here I've shown the use of a 6550 power tube for this purpose. It will require silicon rectification otherwise there's not enough voltage to bias the tube properly. This approach has some advantages--it puts the heat generated from dropping that much voltage above deck, so the power supply under the chassis runs quite cool. It also offers slow warm up. Not shown, but so you don't exceed max heater-cathode voltage rating, you will need to bias up the filament voltage of the 6550 to the cathode voltage, so it will require a dedicated 6.3V filament winding. Not saying you should do this, just offering another idea if you wanted to study it.

The output resistance of the 6550 is kind of high (for a power supply) in this configuration, but since it will be used with a class A power amp design, it shouldn't be an issue, but I wouldn't use this exact approach in a push pull amp power supply. Tweak the grid voltage divider resistors (R1 and R2) to slightly adjust the output voltage if you need, but make sure the tube still operates within its safe region. As depicted, the tube operates at -5V grid with respect to cathode, 170V across the tube, and 140 mA current. Plate dissipation is 60% of design center max for a 6550, so the tube should last years. This approach is somewhat more complicated than the 5U4 approach, but you get slow warmup as a benefit.



In summary, no matter which way you go here, you've got to do something different than what's shown in your PS schematic to get this to work properly.

Easiest approach (my opinion) would be to obtain a 250V to 260V secondary and use silicon rectification, or a 300V secondary and use 5U4 rectification, do cap input filtering, and use an R/C filtering section after the second choke to drop it down to the 315V you need.

 

[Kennyg2209]

Thanks for your information. I'm going to look into using a 6550 for that purpose. I'll like to explore all my options and learn all I can . I've done some changes since my last post. I had an extra 6.3v tap on the transformer and wired it with the primary as a buck which dropped my voltage down to 368v. I removed the first resistor and replaced it with a 10uf cap. I did a couple other changes also. I'm going to post another photo later today. Thanks again for your help!

Kennyg

 

[primosounds]

It seems to me that the easiest solution is to get a proper input choke and make an LCRC or LCLC ps filter. A choke input supply is regarded as one of the best types for single ended amplifiers. It just requires the space.

 

[kward]

I wouldn't disagree with Primo's comments. Choke input will allow you to use 350V rated caps and not need to connect them in series for higher voltage capacity, so the power supply would be simpler. However, you'd need silicon rectification or else you'd be lower than your targeted 315V output.

 

[Kennyg2209]

I wouldn't disagree with Primo's comments. Choke input will allow you to use 350V rated caps and not need to connect them in series for higher voltage capacity, so the power supply would be simpler. However, you'd need silicon rectification or else you'd be lower than your targeted 315V output.

When I first started planning this I was going by another schematic and the 2A3 tubes must be new run tubes that can operate at higher voltage. The tubes I have are NOS RCA. The typical operating voltage for them is only 250v. 300v max. So I will be operating at a lower voltage. My main problem right now is, I'm kinda on a limited budget and can't afford a larger choke. I have everything I probably need other than that. I've read in some other forums of guys using small value chokes with good results. Is the problem using them poor voltage regulation? The psu2 doesn't show any problems. Is there something that it doesn't account for?


Kennyg

 

[Kennyg2209]

Don't get me wrong. I appreciate all the information that you guys are giving me . I want to stick with building tube amps and eventually this will all come naturally to me. I have some health problems and there's not a lot I can do physically. This is something that I can do and so far enjoy it very much.

Thanks again!

Kennyg

 

[kward]

Chokes need current to do their work. If current falls below critical for the given choke size, the choke will become invisible to the circuit. For a choke input PS, having the current fall below critical will cause a voltage spike (power supply will suddenly become cap input until the current again rises above critical). Class A amps are more straight forward because they have constant current. Working backwards, given a known current, critical inductance for a choke input power supply can be estimated as:

L = V/(I*1200) for 60 Hz mains

or

L = V/(I*1000) for 50 Hz mains

In your case:
V = 350V (approx)
I = 170 mA = 0.17 A

L = 350/(0.17 *1200)
L = 1.7H

That's the critical inductance. I'd probably 1.5x to 2x that value to give you some headroom, so a 3H choke would be indicated here.

Construction of a choke for choke input power supplies is important. They can buzz like a band saw sometimes. Many people building choke input supplies will seek out surplus or salvage units from the golden era which were built like tanks, many were potted.

PSUD may not be able to simulate these boundary conditions appropriately. PSUD works for main stream stuff, far away from any edge cases.

 

[Kennyg2209]

Chokes need current to do their work. If current falls below critical for the given choke size, the choke will become invisible to the circuit. For a choke input PS, having the current fall below critical will cause a voltage spike (power supply will suddenly become cap input until the current again rises above critical). Class A amps are more straight forward because they have constant current. Working backwards, given a known current, critical inductance for a choke input power supply can be estimated as:

L = V/(I*1200) for 60 Hz mains

or

L = V/(I*1000) for 50 Hz mains

In your case:
V = 350V (approx)
I = 170 mA = 0.17 A

L = 350/(0.17 *1200)
L = 1.7H

That's the critical inductance. I'd probably 1.5x to 2x that value to give you some headroom, so a 3H choke would be indicated here.

Construction of a choke for choke input power supplies is important. They can buzz like a band saw sometimes. Many people building choke input supplies will seek out surplus or salvage units from the golden era which were built like tanks, many were potted.

PSUD may not be able to simulate these boundary conditions appropriately. PSUD works for main stream stuff, far away from any edge cases.

Great, I appreciate you explaining it to me. I'm actually going to pick up a

Chokes need current to do their work. If current falls below critical for the given choke size, the choke will become invisible to the circuit. For a choke input PS, having the current fall below critical will cause a voltage spike (power supply will suddenly become cap input until the current again rises above critical). Class A amps are more straight forward because they have constant current. Working backwards, given a known current, critical inductance for a choke input power supply can be estimated as:

L = V/(I*1200) for 60 Hz mains

or

L = V/(I*1000) for 50 Hz mains

In your case:
V = 350V (approx)
I = 170 mA = 0.17 A

L = 350/(0.17 *1200)
L = 1.7H

That's the critical inductance. I'd probably 1.5x to 2x that value to give you some headroom, so a 3H choke would be indicated here.

Construction of a choke for choke input power supplies is important. They can buzz like a band saw sometimes. Many people building choke input supplies will seek out surplus or salvage units from the golden era which were built like tanks, many were potted.

PSUD may not be able to simulate these boundary conditions appropriately. PSUD works for main stream stuff, far away from any edge cases.

I really appreciate your reply and that you gave me reasoning behind it. So I should be okay using a 3h, 200ma choke. Using that would that be as a choke input? I would not need a cap input. Also , could I still use one of the 500mh that I have after that? Again, Thanks!

Kennyg

 

[primosounds]

When I first started planning this I was going by another schematic and the 2A3 tubes must be new run tubes that can operate at higher voltage. The tubes I have are NOS RCA. The typical operating voltage for them is only 250v. 300v max. So I will be operating at a lower voltage. My main problem right now is, I'm kinda on a limited budget and can't afford a larger choke. I have everything I probably need other than that. I've read in some other forums of guys using small value chokes with good results. Is the problem using them poor voltage regulation? The psu2 doesn't show any problems. Is there something that it doesn't account for?


Kennyg

There are those that believe this type of ps is better because of its supposed ability to deliver current quicker than a more conventional ps. Thus components are smaller and of lower value. The biggest drawback is that you might have a higher noise floor than with a choke ps with chokes that exceed the critical level and caps of higher capacitance. I suppose it depends on how much hum you are willing to tolerate in the output from the speakers. Of course, it is hard to know until you build it.

 

[kward]

So I should be okay using a 3h, 200ma choke. Using that would that be as a choke input? I would not need a cap input. Also , could I still use one of the 500mh that I have after that?

You've got some extra current needed to run the frontend tubes, so I'd plan on a 3 Henry 200 mA choke minimum. Get one that is potted to minimize mechanical buzz. Don't forget to include the DC resistance of the choke in your voltage calculations. For a perfect choke with no loss, the choke will pass 0.9 of the RMS input voltage from the secondary. But in reality, at full load, you will see more like 0.85.

So reasonable estimate of output voltage of a choke input PS, with a 40 ohm DCR choke, a 380V RMS PT secondary, and 200 mA current draw total, would be:

(380 * 0.85) - (200 mA * 40 ohms) = 323 - 8 = 315V DC. (this assumes silicon rectification, if you use tube rectification you need to subtract the voltage drop across the tube rectifier).

You can add as many L/C or R/C sections after that that are called for in your design.

Also the choke will throw off a magnetic field and could interact with the magnetic field from the power transformer and introduce hum. Good construction techniques are essential--place the choke as far away as possible from the power transformer and output transformers and rotate the choke's axis 90 degrees from the axis of the power transformer. It would be a good idea to look at the layouts of other's work to get some ideas here. The place I'd start though as an example, is on a rectangular chassis, place the power transformer in the back left corner, the choke in the front left corner, and the output transformers on the far right side. Put the audio circuit and tubes front right on the chassis. If you want to build mono blocks, it becomes a bit easier since iron can be a tad smaller.

You can (probably) get away with using a new construction choke (like Hammond or Triad) for this purpose because this is going into a class A amp with constant current draw. If you were building a big push-pull amp or some other circuit that had large current swing, I would think it best to use a "swinging choke" for a choke-input PS.

Good luck! Attention to layout and execution will pay off big in the end for a quiet amp.

 

[Kennyg2209]

You've got some extra current needed to run the frontend tubes, so I'd plan on a 3 Henry 200 mA choke minimum. Get one that is potted to minimize mechanical buzz. Don't forget to include the DC resistance of the choke in your voltage calculations. For a perfect choke with no loss, the choke will pass 0.9 of the RMS input voltage from the secondary. But in reality, at full load, you will see more like 0.85.

So reasonable estimate of output voltage of a choke input PS, with a 40 ohm DCR choke, a 380V RMS PT secondary, and 200 mA current draw total, would be:

(380 * 0.85) - (200 mA * 40 ohms) = 323 - 8 = 315V DC. (this assumes silicon rectification, if you use tube rectification you need to subtract the voltage drop across the tube rectifier).

You can add as many L/C or R/C sections after that that are called for in your design.

Also the choke will throw off a magnetic field and could interact with the magnetic field from the power transformer and introduce hum. Good construction techniques are essential--place the choke as far away as possible from the power transformer and output transformers and rotate the choke's axis 90 degrees from the axis of the power transformer. It would be a good idea to look at the layouts of other's work to get some ideas here. The place I'd start though as an example, is on a rectangular chassis, place the power transformer in the back left corner, the choke in the front left corner, and the output transformers on the far right side. Put the audio circuit and tubes front right on the chassis. If you want to build mono blocks, it becomes a bit easier since iron can be a tad smaller.

You can (probably) get away with using a new construction choke (like Hammond or Triad) for this purpose because this is going into a class A amp with constant current draw. If you were building a big push-pull amp or some other circuit that had large current swing, I would think it best to use a "swinging choke" for a choke-input PS.

Good luck! Attention to layout and execution will pay off big in the end for a quiet amp.

Really appreciate all the information. I went ahead and ordered a 5h Hammond choke. It is impregnated with polyurethane. I assume it's more or less potted. I'm also going to build the power supply in a separate chassis. I'll still keep the direction of the choke in mind. Again I really appreciate your help. Going this route using the psu2, it looks like my ripple will be less than 6mv at a very low frequency.


Kennyg

 

[gadget73]

Potted basically means the choke is in a can, surrounded by some sort of material. Poly impregnated is fairly common, its your standard choke that is dipped to help keep the lamination plates from buzzing around. If its noisy to some extent you can probably mount it on rubber isolators to keep it from transmitting mechanical noise to the chassis.

 

[primosounds]

Really appreciate all the information. I went ahead and ordered a 5h Hammond choke. It is impregnated with polyurethane. I assume it's more or less potted. I'm also going to build the power supply in a separate chassis. I'll still keep the direction of the choke in mind. Again I really appreciate your help. Going this route using the psu2, it looks like my ripple will be less than 6mv at a very low frequency.


Kennyg

Potted chokes are quite expensive if new. Usually at least hundred and up. Impregnated with polyurethane is a type of paint that is very tough like epoxy. It does not mean it is potted.

 

[Kennyg2209]

Potted basically means the choke is in a can, surrounded by some sort of material. Poly impregnated is fairly common, its your standard choke that is dipped to help keep the lamination plates from buzzing around. If its noisy to some extent you can probably mount it on rubber isolators to keep it from transmitting mechanical noise to the chassis.

When I had the 2- 500mh chokes connected the first one did have a slight buzz. I sat it on a piece of rubber and it went away. Maybe the 5h won't do it. Hopefully I'll have it tomorrow. If not Monday. Thanks

Kennyg