• 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.

Constant Current DC Heater Supply

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
In a post the other day, ErikdeBest posted a diagram form the LM338 datasheet for making a constant current source:

ErikdeBest Post

I am looking to build a phono preamp, and have a circuit but no power supply, so I went to look at the datasheet, but there is no math for this circuit. To the right, and below it are current limiter/regulator circuits, with math, so I thought maybe I could figure it out.

I need 6.3V for 4 tube heaters at 0.18A each, or 0.72A total. So first, the LM338 goes to 8A, and in need <1A, so I ended up at the LM317 with a 2.2A limit.

Using the datasheet circuits I 'glued' the current regulator after the voltage regulator, and then put a transformer/diode bridge/capacitor in front of it for a DC source. Put the extra caps and diodes in the circuit per the notes to improve ripple rejection, etc.

Since I'm still learning here, and I don't trust myself with this stuff yet, the result is attached, so maybe I can get some help from the ether.

So, questions:

Is the circuit right?
Is the math to get the voltage/current values correct?
PSUD makes the DC before the first LM look real sloppy. Is that a problem? More filtering required?
Could I do the same thing from a 6.3V transformer? I think the LM317 wants at least a 5V difference in to out, so you'd have to double the 6.3V first, in place of the bridge?

Thanks in advance for helping.

Bill
 

Attachments

  • lm317 heater dc.pdf
    71.9 KB · Views: 381
I hate to point this out but those 1n4002 diodes are going to
be kind of noisy. It might not be a bad idea to start with a better
diode to begin with. Runing those 4 tubes off one current source
may not be a good idea. If one of the tube sockets doesn't make
good contact or you have a tube burn out the other tubes will
then be overloaded.

happy soldering,

Woody
 
Jorge -

You are of course, correct. I went back to the datasheet, but I can find a clear reference about how much extra the second one needs. The minimum Vin-Vout difference is stated as 3V, so maybe that's it?

Woody -

Is there a less noisy diode to use? I searched here, and found lots of discussion, but not many conclusions. Higher rated value so it's less taxed? I also see your point about the extra current. I guess you could wire the heaters serially, if that doesn't foul something else up, or replicate the current regulator for each tube? Other options? Is one way better than another?

Thank you both.

Bill
 
bwarden said:
Jorge -

You are of course, correct. I went back to the datasheet, but I can find a clear reference about how much extra the second one needs. The minimum Vin-Vout difference is stated as 3V, so maybe that's it?


Bill

Yes, 3 volts more the voltage across R4.

But at the end of the day , I will put the same voltage across IC1 and IC2 , for even power dissipation in both regulators... ;)

Say 17-6= 11

11:2 = 5,5

So at the output of IC1 around 12 volts, will be fine..
 
Woody is correct. If you want to run constant current, you have to wire the heaters series and they must all have the same current specification. Keep in mind that the voltage required is the sum of the series tubes. If you want to parallel the heaters, run constant voltage (current sums, in this case).

For indirectly tubes, constant voltage should be fine. Some prefer constant current for directly heated tubes, as the filament is part of the audio circuit and high impedance is claimed to be an advantage in that case. I don't know why constant current would have any sonic benefit for IDHT tubes.

Sheldon
 
Jorge-

Fiddling with the math here, I think the voltage drop across R4 is 1.25v (0.7138A x 1.74 ohms = 1.249 volts). Now that I do the math, any combination of allowable resistors from 0.8R to 120R results in the exact same 1.25 volt drop, which is then made up by the 1.25v from the adjustment leg, right? So the net result of R4 is no change in volts.

If the minimum difference is 3v, then I need to present IC2 with 9.3 volts. But I cannot, apparently, regulate volts and amps at the same time in the same IC, so I can't push 12v out of IC1.

So, if I want to dissipate the same voltage across IC1 and IC2, I need to lower in incoming volts with a resistor in front of IC1, or drop the transformer down to reach 12v, I think.

The LM317 datasheet sees to use 240R for R1 a lot, but I have yet to figure out why. When I revise the values of R2 to up the output of IC1, I may change that too.

An probably run IC1 a few tenths high, to stay way from the 3v minimum. But that will make the dissipation being somewhat higher in IC1, but closer.
 
Sheldon-

I started this out for a phono preamp with indirectly heated tubes - 6GK5's. So maybe the whole current regulator can go for this application.

But if I had the need, I would raise the voltage to 6.3 x 4 = 25.2 volts, and wire them in series, so if one dies the whole string dies. Is there anything bad about series wiring otherwise?
 
Bypass the 1N4002's with film caps to deal with ringing. ERO MKP (or even MKT) 0.01uf works for me.

I do this with generic cheap bridges for DC filament with sonically significant results.

This is cheaper than schotty sb340 (soft recovery) diodes and really not difficult. It also sounds better IMHO.

As for wiring heaters in series, the ear 834p has this and it works well enough... but it makes tube rolling a little more difficult. If you can adjust filament voltage on the fly you will be fine.

You will want to tube roll in order to get the quietest tubes in the first and second stages.

Tip: try running those filaments a little lower than 6.3VDC. not too low though. Some heaters run a little quieter at ~6VDC...
 
bwarden said:
But if I had the need, I would raise the voltage to 6.3 x 4 = 25.2 volts, and wire them in series, so if one dies the whole string dies. Is there anything bad about series wiring otherwise?

That will work and has been done. Sometimes it's done because there is a readily available supply. Take what I say here with the caveat that I have fairly limited experience, especially as compared to others here. With a phono stage, noise is key. I don't think that the heater string itself should be too much of a problem, but I think it's a good idea to tie the heaters to a reference lifted about 25V or so above the cathodes. You have to incorporate the 25V supply into your calculations to make sure that you don't exceed cathode to heater ratings.

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