• 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.

New DHT heater

Neat and simple package!

If I understand it right the voltage over the gyrator is set by the base current through R1, R5 and R9? If not it must float with enough voltage over it for 2*Vbe + the voltage over 0,47ohm? To be fixed if even necessary couldn´t a resistor parallelled with C2 to ground be used?

Many questions:)!

Lars - Yes, you can increase the volt-drop across the Gyrator using a resistor in parallel with C2.

C2 may need to increase if you do that.
 
Hey Rod,
Deleted the last sentence as I thoght the base-voltage shouldn´t be related to ground.

Lars, We must be careful about the word "Ground". You are right, we don't like the filament voltage to be connected (by any circuit) to the B+ return: B- or 0V(B) or HT- etc.

But it's OK to connect a resistor to the filament supply 0V return, if you like, call it 0V(fil)
 
I have had requests for more information on the filament circuit across the five years it has been up here.

Mostly, component questions - It seems that all of the transistors listed are no longer obtainable.

And the circuit published changed a little, in order to use a regular darlington configuration for the gyrator section.

So here's a suggested circuit and parts data for anyone wanting to build today. The ST 2FT1480 is a new part, but has the desirable high ft of 100MHz. I would like to specify a nice TO-220 isolated package NPN with 100M+ for ft, and when I find one I'll try it. Meantime the MJE15032 should work OK, but in the current sink section Q3/Q4, be sure the pair don't oscillate. Hold an AM radio close to it if you don't have a scope. Using the MJ for both Q3 and Q4 may fix the problem if it occurs. In any case, keep the wiring for the current sink as SHORT as possible, especially keeping the base resistors close to the transistors.

The bootstrap circuit C1/R5/R2 is optional. It raises the impedance seen at the gyrator's base, and aims to reduce ripple. Easy to try, but build without it at first!

Be sure to use a heatsink for the power transistors!

Post to correct wrongly connected Darlington - Thanks Stig!
 

Attachments

  • dhtheat2010.jpeg
    dhtheat2010.jpeg
    86.4 KB · Views: 3,092
Regal, that's helpful.

The 2FT1480 is a new part with promising spec, but I haven't had any here for testing yet. If anyone tries one, please report the amount of ripple reduction you get, and the Vce it was measured at.

In the meantime the TOSHIBA 2SC4881 looks even better for applications up to 3A filaments, with only 2V across Vce, and specified base current of only 10mA for 2A at 1V Vce. This one has been around a while, but I don't know how easy it is to obtain.
 
OK,I'll try harder!

All the NPNs below are at least as good as 2ST1480, and all surpassing MJE15032 by a big margin. The Fuji part is a real Boster, as you may hear said in Birmingham:

All Japanese. Given with Cob (pF); Hfe at 1A, and base current needed for 1V Vce at 1A

Toshiba 2SC4688; 105pF; 55..160; 20mA
Sanyo 2SC3748; na; 70..280; 20..40mA
Fuji Electric 2SC3852; 50pF; 400..500; 3mA

These are certainly in the realm of Obtanium! Try a TV component shop rather than a catalogue distributor, since these try harder with Japanese parts.

In the Europe/UK you can get all these at Cricklewood, and I recommend their web shop as fast and reliable.

Cricklewood Electronics CCTV and Electronic Components
 
For those that might be interested, I made a variation on Rod's design which is more easily adjustable, by using a TL431 as a reference in the CCS. R2 is a pot which allows the output to be adjusted.

It seems to be very quiet and sounds good, too.
 

Attachments

  • CCS_heater_431.jpg
    CCS_heater_431.jpg
    100.9 KB · Views: 2,501
Lars, ZTX449 NPN & ZTX549 really the best. BDxx a bit too heavy for 220mA, capacitance too high, Hfe too low. Use these for the small signal positions too, for least chance of oscillation.

If you can't get these, Japanese transistors with high Hfe and high Ft are worth looking for.

Run the ZTX with 2V across each, and watch this doesn't get too high & overheat them. Some copper EMC foil around them helps cooling.
 
This was actually a 300B (1.2A, 5V) version, You can simply vary the value of the sense resistor, R7 to change the current. For 1A, just increase the sense resistor, R7. The idea is that the T-431 sets the base voltage to 0.7V + the voltage across the sense resistor. I actually used an LMV431 because that has a minimum output of 1.25V, otherwise, it gets a bit inefficient.
 
Last edited:
This was actually a 300B (1.2A, 5V) version, You can simply vary the value of the sense resistor, R7 to change the current. For 1A, just increase the sense resistor, R7. The idea is that the T-431 sets the base voltage to 0.7V + the voltage across the sense resistor. I actually used an LMV431 because that has a minimum output of 1.25V, otherwise, it gets a bit inefficient.

Cool it sounds like this version would maintain voltage as the tube ages which has been a problem with LT1085 CCS's?
 
Cool it sounds like this version would maintain voltage as the tube ages which has been a problem with LT1085 CCS's?

Actually, it's no different in that respect to Rod's design, it's still a CCS, rather than a voltage reg. The reason I did it was that it's easier (cheaper) to find low-wattage adjustable resistors than the high wattage ones required if you want to vary the sense resistor. (Or alternatively, trim with fixed resistors, but that makes tube changes a pain).
 
Rod,
Just came across this thread and noted above your concerns with oscillation in the Q3 Q4 "Ring of Two" Current Source.
I use "Ring of Two" Sources reasonably often and always build them with a 100 Ohm resistor in the base of the Pass Transistor, in your circuit above this is Q3. So add a 100 Ohm between base of Q3 and collector of Q4 and I think you can stop worrying about any stability problems.
Cheers,
Ian