The average specified transconductance of a 45 tube is 2050 microMhos.
(2.05mA/Volt)
At 3.536V peak across the filament, that could never be more than 7.249mA, but it is Much less than that because the + and - voltage is so well balanced from end to end; when referred to the center of the filament wire, and is only 1.768V x 2.05mA = 3.624mA per filament end (again, fairly well balanced versus the filament center).
If instead, someone makes a mistake and connects one end of the filament directly to the top of the self bias resistor . . .
Then there is Lots of Hum!
Now, consider the same scenario with a DHT that has a 10Volt filament (14.14Vpeak), and a transconductance of perhaps 5,000 microMhos
(5mA/Volt). A more serious problem.
(2.05mA/Volt)
At 3.536V peak across the filament, that could never be more than 7.249mA, but it is Much less than that because the + and - voltage is so well balanced from end to end; when referred to the center of the filament wire, and is only 1.768V x 2.05mA = 3.624mA per filament end (again, fairly well balanced versus the filament center).
If instead, someone makes a mistake and connects one end of the filament directly to the top of the self bias resistor . . .
Then there is Lots of Hum!
Now, consider the same scenario with a DHT that has a 10Volt filament (14.14Vpeak), and a transconductance of perhaps 5,000 microMhos
(5mA/Volt). A more serious problem.
At the filament we have a very particular circuit , because of the current flow and simetry , as long as half filaments are equals and bottom two resistors are equals then the top and bottom voltages are equals too , it is not adding any voltage in series with the cathode ... For tubes with no physical center tap the middle can be imagined as a virtual center tap.
Of course it is why the tube is imune to hum when using AC filament voltage
It is used in Wheatstone bridge to measure resistors , impedances so on .
Of course it is why the tube is imune to hum when using AC filament voltage
It is used in Wheatstone bridge to measure resistors , impedances so on .
Last edited:
The question revolves around whether or not there is such a critter as a virtual filament center tap. Put another way, does the space charge have a "time constant" so much larger than signal's that all emission from the filament is unipotential (like a separate cathode)? Modulation of signal by AC heating voltage is of course related, but is it the same thing?
All good fortune,
Chris
All good fortune,
Chris
No. Even the DC current is skewed.
See for example:
H.Barkhausen: Elektronenröhren. Leipzig 1951:
THe diagram title means: "influence of the heating voltage across the Anode Voltage of the different [sections of the filament] on the triode curves"
- each curve in the diagram represents anode current in each section of a five-section filament. The bold line is the summed current.
In those days when only battery filament was used ( DC low voltage high current rectification was nearly impossible ) the circuit was simple , one filament terminal directly to ground with the tube grid biased from a negative voltage ( another battery of course ) . In such a circuit obviously one side of the direct heated cathode is ground and the other is the positive batery voltage . Obviously again the tube will produce those curves .
Last edited:
While on the subject of Hum and DHTs, consider this:
A 45 with AC powered filaments: 1.5A has 2.12A Peak current.
That creates a large magnetic field.
Magnetic fields attract magnetic steel.
If the filament is closer to one side of the plate than to the other side of the plate, we have an unbalanced magnetic field.
The filament moves toward the closer side of the plate, than to the further side of the plate.
The resultant hum is at 2X the power mains frequency; 100Hz or 120Hz.
On some 45 tubes, you can hear hum at 100Hz or 120Hz accordingly (even if the B+ has absolutely no ripple).
An adjustment pot can not cancel that 100Hz or 120Hz hum that you hear on some of the 45 tubes.
(The hum adjustment pot cancels 50Hz or 60Hz hum; it does not cancel 100Hz or 120Hz hum).
Those of us who have actually used 45, 2A3, and 300B tubes with AC powered filaments:
We know from experience, some DHT tubes have much much lower 100Hz and 120Hz hum, than other DHT tubes of the same type number.
(No matter where you turn the 50Hz 60Hz hum nulling pot).
Good luck finding the good DHT tubes in regards to the 100Hz 120Hz hum!
Or, power the filaments with DC.
A 45 with AC powered filaments: 1.5A has 2.12A Peak current.
That creates a large magnetic field.
Magnetic fields attract magnetic steel.
If the filament is closer to one side of the plate than to the other side of the plate, we have an unbalanced magnetic field.
The filament moves toward the closer side of the plate, than to the further side of the plate.
The resultant hum is at 2X the power mains frequency; 100Hz or 120Hz.
On some 45 tubes, you can hear hum at 100Hz or 120Hz accordingly (even if the B+ has absolutely no ripple).
An adjustment pot can not cancel that 100Hz or 120Hz hum that you hear on some of the 45 tubes.
(The hum adjustment pot cancels 50Hz or 60Hz hum; it does not cancel 100Hz or 120Hz hum).
Those of us who have actually used 45, 2A3, and 300B tubes with AC powered filaments:
We know from experience, some DHT tubes have much much lower 100Hz and 120Hz hum, than other DHT tubes of the same type number.
(No matter where you turn the 50Hz 60Hz hum nulling pot).
Good luck finding the good DHT tubes in regards to the 100Hz 120Hz hum!
Or, power the filaments with DC.
Last edited:
Add to findings:
I had a 300B PP, AC fed. Both tubes on the very same AC winding; with the appropriate common Rs of something like 470ohms.
Very slight hum (94 dB speakers).
Then I reversed the filament connection on one of the tubes. Hum gone!
I had a 300B PP, AC fed. Both tubes on the very same AC winding; with the appropriate common Rs of something like 470ohms.
Very slight hum (94 dB speakers).
Then I reversed the filament connection on one of the tubes. Hum gone!
The 50Hz/60Hz hum was gone.
Wonderful!
Were the 2X power mains frequency intermodulation upper and lower sidebands also gone (no longer there at -60dBc, or -80 dBc on each musical note, and on each musical harmonic?
Did you measure that?
Push Pull, and Parallel Single Ended can sometimes partially cancel that effect.
Wonderful!
Were the 2X power mains frequency intermodulation upper and lower sidebands also gone (no longer there at -60dBc, or -80 dBc on each musical note, and on each musical harmonic?
Did you measure that?
Push Pull, and Parallel Single Ended can sometimes partially cancel that effect.
Last edited:
Yes, this is indeed the theory.
Realistically speaking, the effect could be considered as marginal for DHT valves where the voltage is only 2.5V though. Maybe it is somewhat more significant for DHT's that have 5V or 7.5V heaters.
Ian
6A3sUMMER
I have built cheap 2a3 amps with AC heaters, and the most deciding factor on Hum with them was that AC heater transformer itself. Some "budget" heater transformers I tried out were noisy all on their own! Beyone that, indeed some 2a3's or 45's are just going to make some kinds of distortion noises anyway too. We neglect that a LOT too. Making music using valves/tubes is kind of crazy for this reason alone! 😏
I threw away half a dozen (or more) Chinese 2a3's at one point because they were just too noisy and I could not in good concience use them. Of course, these were all made in the late 90's and were of the "inexpensive" variety. The good ones that I selected out (which were not noisy) sounded great and as far as I gather, have lasted a LONG time too.
I did regulated DC for 2a3 single ended amps a few times, and guess what? Nobody could "hear" the difference vs. AC with decent quality filament transformers. You may very well measure intermodulation distortion if you were specifically looking for it. I don't doubt that. But if you managed to wring out all the significant sources or distortion in your design and build, intermodulation distortion due to AC heaters is not going to be very significant. At least that was I concluded so far.
That said, for 6A3's or 300b or 801a - no question - DC is the way to go. But this wear on the filament due to DC theoretically will be there at some point or another with these valves. The 801a thoriated tungsten filiments are fragile, so handle them really carefully. I had a few arrive damaged in transit which is heartbreaking.
Ian
Last edited:
Just had a thought. For those who worry about uneven emission with DC filaments. Use some latching relays so that on every other power up the filament continuity is reversed.
That was done for the really big (and expensive - they were even rebuilt when aged out) final valves in analog TV transmitter days.
All good fortune,
Chris
All good fortune,
Chris
Provided the DC filament voltage is held at the correct value, the Barkhausen curves will remain valid - with a higher current density of emission at the Filament + side, and a lower current density at Filament -.
The emissive current density depends on the pattern of the E-field, which is set only by the relative electrode voltages.
Adding a more complex cathode return path may alter the current distribution in the metallic substrate of the filament (the nickel or tungsten wire), but this does not substantially alter the emissive current distribution.
Yes, I believe that this is all true, but would always take the precaution of allowing a good margin for the power dissipation of the DHT, compared to the maximum permitted value, to allow for any skew of heating.
In any reasonable design, this presents no problem at all. For instance, DHTs like the 300B sound worse with excessive anode voltage, IME.
My DHT filament regulator is sufficiently quiet that the cathode return can be made on one side, and constructors have told me that their DHTs sound best with this connexion.
Correct,
When a single ended DHT amplifier is only putting out low or medium power (versus its rated power output):
Then, the 2X power mains frequency sidebands are very far down (-dBc) from the music notes and music harmonics that they intermodulate with.
Nothing to worry about; nothing to 'hear'.
But at fairly high power output, when the grid voltage is at the contact potential, then the amplitude of the 2X power mains frequency sidebands can be significant (even though the grid is still negative versus the DHT filament voltage).
Just like many other artifacts and characteristics of amplifiers, they start to struggle when they get near to their maximum power output.
As i always say, if it sounds worse at high power, either turn down the volume control,
or go purchase another amplifier with at least 3dB more maximum power.
Just my opinions
When a single ended DHT amplifier is only putting out low or medium power (versus its rated power output):
Then, the 2X power mains frequency sidebands are very far down (-dBc) from the music notes and music harmonics that they intermodulate with.
Nothing to worry about; nothing to 'hear'.
But at fairly high power output, when the grid voltage is at the contact potential, then the amplitude of the 2X power mains frequency sidebands can be significant (even though the grid is still negative versus the DHT filament voltage).
Just like many other artifacts and characteristics of amplifiers, they start to struggle when they get near to their maximum power output.
As i always say, if it sounds worse at high power, either turn down the volume control,
or go purchase another amplifier with at least 3dB more maximum power.
Just my opinions
With my 300BC I never got so far. When I (on the scope that is, not the room) push the 300B SE, then I see it will sort of max out at a fairly higher voltage than expected, say 40-50V. It won't go zero. The bottom of the sine blunts out really fast, from about 60V. It puzzled me.
- So is that also the moment that the signal starts to have a significant higher current from one sde of the filament compared to the other probably? Could be due to the concept of the tube. I remember some old diagrams of the 300B tube that showed the grids are NOT eavenly wound because of some unequal transmission of the cathode at their ends.
- Anyway because of that I concluded the 300B is a <5W SE tube.
This is big contrast to my earlier thoriated tungsten VT25/10Y that I can drive much better.