Well, almost, feeding heaters with constant voltage is inherently more accurate and more safe for the heater itself.
I am not a supporter of the Schrödinger's heater.
What I intended to say is that if heaters designed for constant voltage are feeding with constant current, on the long term you will ended up with overheated heaters.
The OP has the same issue IMHO, even common resistors thermally over-stressed end up with a larger value or fried.
On the other hand, for some directly heated triodes, for the aim of sound, there is no chance but constant current feeding as I am planning for my 300B's, something like "plug and pray"
Last edited:
There's an odd jiggle at 12.6V, which makes no physical sense.
Ignoring that, it is as we always knew: a heater is a resistor that *tends* somewhat toward constant current rather than simple resistance.
Nobody here remembers series-string lamp? Holiday festoons? Runway lights? A string of same-design filaments is stable. It would be stable even if the same filament wire were cut 4V for tuner and 40V for H-sweep tube. The "problem" in series-string TVs is that various heater designs were used, some slower-heating than others. The late 1950s agreement improved the situation. (Lowest-bidder sourcing ensured continued troubles.)
Series-string lamps were also fed with Constant Current supplies (electromechanical!). If in a long string, one lamp burned, the whole street or runway was dark. Instead each lamp had a shunt gap with paper insulator. (I have one on my porch, though I've bypassed the cutout.) If lamp voltage rose from 100V to 2,000V, the paper punched-through, that one lamp shorted, the rest of the string ran at nominal current with slightly less total voltage on the string.
Attachments
Last edited:
Funny you mention that, PRR. Once upon a time I decided to run two MR16 50W halogen lamps in series from a 24V transformer. Slowly, but surely, one lamp got brighter as the other one got dimmer. Eventually the brighter one burnt out. Thankfully, tube heaters are designed for heat instead of light, eh?
Running a series string from a voltage supply is not a good idea. It may be (I would have to think about this) that two in series is the worst combination, as the individual loop gains effectively add together whereas a longer string would dissipate a small change in voltage across a larger number of filaments.kodabmx said:
I just knew those buck converters would come out!
LOL.
My own dumb-as-a-box-of-hammers approach really simple: for that one tube, look up its nominal current and voltage requirement, and figure out the size resister to put in series so that 12.6 volts drives it OK. The ceramic 5 watters are plenty cheap. And last forever.
$1.06 qty 1, Xicom, 5 watt. Two wires. Life is simple.
Just saying,
GoatGuy
LOL.
My own dumb-as-a-box-of-hammers approach really simple: for that one tube, look up its nominal current and voltage requirement, and figure out the size resister to put in series so that 12.6 volts drives it OK. The ceramic 5 watters are plenty cheap. And last forever.
$1.06 qty 1, Xicom, 5 watt. Two wires. Life is simple.
Just saying,
GoatGuy
99 cents, and it's 80+% efficient 3 wires...
XL6009 *1Pcs DC-DC Adjustable Step-up Power Converter Module Bette r than LM2577 | eBay
FWIW, a "dumb as a box of hammers" resistor is quite reliable and foolproof...
3 wires...XL6009 *1Pcs DC-DC Adjustable Step-up Power Converter Module Bette r than LM2577 | eBay
FWIW, a "dumb as a box of hammers" resistor is quite reliable and foolproof...
It seems so for the purpose of correct operation, however for some reason they sound different.
Other than possible intermod distortion from different implementation (poor filtering on the constant voltage source, or poor/nonlinear frequency response of the current source) I I fail to see how either will sound different. Done right the net result is_the_same- quiet DC across the filaments.
Any other claims are expectation bias unless some measurements are presented.
Any other claims are expectation bias unless some measurements are presented.
What you believe you heard is anecdote, not evidence.
As I said, the thermal time constant of the heater-cathode system means that you would need a heater supply with serious 1/f noise in order to hear any difference between voltage and current drive. Normal AC drive will have much worse noise etc. yet that is fine for almost all audio purposes.
As I said, the thermal time constant of the heater-cathode system means that you would need a heater supply with serious 1/f noise in order to hear any difference between voltage and current drive. Normal AC drive will have much worse noise etc. yet that is fine for almost all audio purposes.
Anecdotal to you, evidence to me (my ears to please only and YMMV etc).
Ah ok, I was meaning AC vs DC heating. I've not compared current regulation with voltage regulation for use with IDHT.
I have completed replacing the current regulation for the twelve heaters of my RTP3 with voltage regulation, and checked the whole thing is functioning properly. I have obviously been listening to an under-par preamp for twelve years! Now the heaters have the correct voltage, the odd noises are gone, and everything is clearer and more "present".
The thing that still puzzles me is that I was using Allen Wright's recommended brand of valves from his recommended supplier, and feeding the heaters in exactly the manner he recommends in print, but this highly sophisticated circuit was not working as intended. The moral of the story, I suppose, is to question everything, and to check everything.
Once the festive season is over, I'll do some measurements of the EH6922 and the Sylvania branded valves - my suspicion is that the NOS parts are working according to spec, but that the EH ones are taking a higher current. Anyway, I'll report back in due course.
Alex
The thing that still puzzles me is that I was using Allen Wright's recommended brand of valves from his recommended supplier, and feeding the heaters in exactly the manner he recommends in print, but this highly sophisticated circuit was not working as intended. The moral of the story, I suppose, is to question everything, and to check everything.
Once the festive season is over, I'll do some measurements of the EH6922 and the Sylvania branded valves - my suspicion is that the NOS parts are working according to spec, but that the EH ones are taking a higher current. Anyway, I'll report back in due course.
Alex
Last edited:
I should have made it clearer, appreciate your response and wish you a great festive period. Extended of course to all.
Last edited:
OK, the measurements are in! I put together a simple circuit to apply an adjustable 6.30V across the filaments of two of my EH 6922s from different batches, as well as a NOS Sylvania 6922. After fifteen or twenty seconds, I measured the voltage across a 1.0R (tested to be within 1%) 1W resistor in series with the valve in each case, and the results are:
EH 6922 (date stamp 0806): 362 mA
EH 6922 (date stamp 1804): 358 mA
NOS 6922: 304 mA
So I conclude that the EH valves are way out of spec for a 6922. The data sheet for the EH6922 from the BTB webpage specifies 335mA as the maximum current draw, so they are not even consistent with the manufacturer's own specifications.
Alex
EH 6922 (date stamp 0806): 362 mA
EH 6922 (date stamp 1804): 358 mA
NOS 6922: 304 mA
So I conclude that the EH valves are way out of spec for a 6922. The data sheet for the EH6922 from the BTB webpage specifies 335mA as the maximum current draw, so they are not even consistent with the manufacturer's own specifications.
Alex
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.