How much magnetic field is there with a 1.5 Inch long filament wire that is carrying 1.2 Amps (like one of the filament wires of a 300B)?
The magnetic field rises and falls at the alternation rate, not at the frequency rate (there are 2 alternations per cycle).
And, there are several of those wires in a W or 2W filament.
Just sayin'
The magnetic field rises and falls at the alternation rate, not at the frequency rate (there are 2 alternations per cycle).
And, there are several of those wires in a W or 2W filament.
Just sayin'
Is that a homework assignment? If i remeber correctly its the Lorentz force.
Lorentz force - Wikipedia
Edit i just realized its a pointless endeavour because you need to know what the spring tentsion holding up the fillament on top is exactly, and the exact elasticity of the tungsten wire used. Furthermore you can only model this quite easily for a single wire in a cylindrical anode because the distance of the greatest mass (The plate) is uniform. Wheras in Oval plate structures you would need to get fancy with multiple integrals whos output is squared....
You can ofcourse model this, and get a number, but then you have a number. And what does this number tell you? very little unless your the ultimate expert of course. But by that time you have spend so much time working it out. It would have been cheaper to use DC.
Lorentz force - Wikipedia
Edit i just realized its a pointless endeavour because you need to know what the spring tentsion holding up the fillament on top is exactly, and the exact elasticity of the tungsten wire used. Furthermore you can only model this quite easily for a single wire in a cylindrical anode because the distance of the greatest mass (The plate) is uniform. Wheras in Oval plate structures you would need to get fancy with multiple integrals whos output is squared....
You can ofcourse model this, and get a number, but then you have a number. And what does this number tell you? very little unless your the ultimate expert of course. But by that time you have spend so much time working it out. It would have been cheaper to use DC.
Last edited:
Thermal effects were tested experimentally by Dmitry Nizhegorodov in 2003.
He showed that even by increasing the frequency to 600Hz, the ac-frequency components remain at the same amplitude, and as such thermal modulation can be dismissed as cause.
On Correlation Between Residual DHT Filament Hum and AC Frequency
The true original of hum and IMD from ac-heating can be traced by careful inspection of the triode curves.
I would have thought the most likely explanation is the that the heater generates its' own magnetic field and simply wobbles around changing its' distance to the grid in places and by a small fraction.
There's also an effect that one end of the heater is high and the other end low - the lower end will produce disproportionally more current due to the non-linear transfer function than when both ends are at the same potential. This will produce 100Hz.
I agree that that 50Hz IMD can be noticeable if it exceeds a certain level. Below a certain level its masked by the tone itself. This depends on the frequency of the tone and the frequency of the hum.
There's also an effect that one end of the heater is high and the other end low - the lower end will produce disproportionally more current due to the non-linear transfer function than when both ends are at the same potential. This will produce 100Hz.
I agree that that 50Hz IMD can be noticeable if it exceeds a certain level. Below a certain level its masked by the tone itself. This depends on the frequency of the tone and the frequency of the hum.
Last edited:
That could very well be, I read an article about AM interference caused by a light bulb: Exerp bellow. Full text here: Can incandescent light bulbs cause interference? | EMC and Regulatory Compliance
Recently the European market surveillance authorities (AdCos) on EMC have been investigating the potential of incandescent light bulbs to cause radio interference. The investigation began with reports that some incandescent lights may be the cause of reported interference to FM broadcast reception. The findings show that a specific type of incandescent bulb can interfere with radio reception. The type is referred to as a 'squirrel cage' Edison bulb in the report.
As an Electromagnetic Compatibility engineer, this came as a bit of a surprise to me. It's been accepted by EMC experts that an incandescent light bulb on its own wouldn't emit radio signals, in fact, the European standard covering interference from lighting states;
In the United States, the problem was known in the 50s to cause interference with T.V. reception. Although not in production in the U.S. at the time, there were still a number of the straight tungsten-filament Mazda bulbs in use. Tungsten-filament bulbs of the Mazda type were initially more costly than carbon filament bulbs but used less electricity. A direct comparison to the LED vs Incandescent light bulb today
This Popular Science article from 1953 titled "U.S. declares war on static", a campaign of hams and hobbyist sponsored by the FCC shows an example of how to identify T.V. interference from an obsolete light bulb. Today's filament light bulbs use an inert gas and relatively small coiled filament rather than a hard vacuum and a long zigzag filament structure.
An investigation of the science and physics behind the phenomenon indicate this is caused by the filament oscillations coupled with thermionic emissions from the heated filament in a hard vacuum, similar to properties found in vacuum tubes. This post "Rustika lightbulb FM measurements " provides data and analysis and links to other works.
Recently the European market surveillance authorities (AdCos) on EMC have been investigating the potential of incandescent light bulbs to cause radio interference. The investigation began with reports that some incandescent lights may be the cause of reported interference to FM broadcast reception. The findings show that a specific type of incandescent bulb can interfere with radio reception. The type is referred to as a 'squirrel cage' Edison bulb in the report.
As an Electromagnetic Compatibility engineer, this came as a bit of a surprise to me. It's been accepted by EMC experts that an incandescent light bulb on its own wouldn't emit radio signals, in fact, the European standard covering interference from lighting states;
"incandescent lights "are not expected to produce electromagnetic disturbances"
yet this new data proves otherwise. So what's changed? After a bit of research I found the answer, nothing, we've just stumbled on an old problem inherent to the once obsolete but now trendy retro vintage style vacuum Edison bulb.In the United States, the problem was known in the 50s to cause interference with T.V. reception. Although not in production in the U.S. at the time, there were still a number of the straight tungsten-filament Mazda bulbs in use. Tungsten-filament bulbs of the Mazda type were initially more costly than carbon filament bulbs but used less electricity. A direct comparison to the LED vs Incandescent light bulb today
This Popular Science article from 1953 titled "U.S. declares war on static", a campaign of hams and hobbyist sponsored by the FCC shows an example of how to identify T.V. interference from an obsolete light bulb. Today's filament light bulbs use an inert gas and relatively small coiled filament rather than a hard vacuum and a long zigzag filament structure.
An investigation of the science and physics behind the phenomenon indicate this is caused by the filament oscillations coupled with thermionic emissions from the heated filament in a hard vacuum, similar to properties found in vacuum tubes. This post "Rustika lightbulb FM measurements " provides data and analysis and links to other works.
Yes, that's what I mean by inspecting the curves.
If the centre-point of the filament is the the effective return-point of the anode-current, the two ends swing (in opposition) in time to the line waveform. If everything were prefectly symmetrical, there would be no hum, and no line-derived IMD. But symmetry is not perfect...
It is not simply Vgk that varies with the line waveform; Vak does as well - and since the amplitude is non-negligible, even for 2.5V filaments, the effect is more complex than at first sight.
I love this. You havent yet seen a problem, but have already budgeted for solution.
Here is one for you to solve. In a typical supply, where a full wave rectifier is charging a bank of capacitors, the current waveform is so bad that transformer has to be derated about 30% because HF components heat core and windings. 30%! This garbage corrupts the mains waveform, so, dealing, for example with AC heaters or compensating ground loops, one has to take care not only of the fundamental, but the plethora of harmonics.
I know this is unavoidable in modern environment where virtually everything uses this kind of power supply. I am lucky to live in countryside, with my house powered from its own transformer. I have a dedicated filtered line for my audio. I dont want to corrupt my reasonably clean power. All my power supplies are choke input, tube rectified, and I want to keep it that way.
Solid state rectification is a much smaller problem; it is masked by the elephant-sized capacitor input one. But it is still a problem.
At April 2019 I built a 845 PP amplifier and measured IMD with AC and DC heaters:
We discussed about that here: 845 PP Prototype
We discussed about that here: 845 PP Prototype
Last edited:
Im not so sure of this. Its not in any textbooks that im aware of. Just the textbook that covers the basics of electrical engineering. I think if you took the first week of my classes you would have realized this.
You are accurate in some of your assertions, but not all and some correct information is mixed in with poor engineering understanding.
You are conflating the fact that because if you go from AC to DC with an ideal rectifier your DC voltage increases by the square root of two. And in an ideal transformer, no power is wasted so the current goes down proportionally. This formula is 1/2 * SQRT(2) which is approximately .715 or thereabouts (from the top of my head) This means that a transformer can indeed deliver 0.7A of DC for 1A of AC rating (This is tied in to the VA rating of transformers)
In short:
I think you are conflating the Derating of going from AC to DC. With Skin effect losses only seen at 100's of KHz transformer frequency. Solid state diodes produce SOME switching junk, but at 50/60Hz this is hard to measure unless you use some VERY slow diodes.
Mains harmonic distortion (Flattening of the mains waveform due to capacitor charging) is indeed a problem, If you rectify the mains directly so you can feed it into a SMPS /Flyback or what have you. That is why most good SMPS of over say 200W for the sake of argument use something called a POWER FACTOR CORRECTION.
A flyback takes a fixed amount of current from the supply every cycle, so if you chop up the rectified mains voltage at a fixed frequency and modulate the PWM with the mains frequency, you can have the pulse with follow the mains frequency. So the Switchmode power supply behaves like an ideal resistive load. So the mains AC doesnt get loaded during the charging peaks.
You can see the same thing happening if you have ever seen a single phase AC motor, there is usually a big capacitor on the side. That is to correct the power factor so the motor can get the energy to start up. If this capacitor goes open circuit, the motor will have trouble starting. If you look on the type shield on the motor there is someting called COS (ϕ) which is the angle current expressed in a cosine.
Phi - Wikipedia
Waveform distortion in transformer insulated supplies.
When you rectify from the secondary of a transformer, a few things happen, first of all you can estimate the peak capacitor charging current from the DC resistance of the wiring added to the DC resistance of the transformer and add the ESR of the capacitor for good measure. Divide the AC voltage by the sum of these resistances and you have the Absolute maximum current the capacitor can draw. In reality this current is lower because the secondary of the transformer is also an inductor, so the current does not automatically track the output voltage.
Protip: If you want clean power: Look for a medical grade insulation transformer with a shield between primary and secondary, this will give better noise suppression and solve most common mode issues.
You can also try a line stabilizer, which is essentialy a saturation transformer that produces a very clean and stable mains, regardless of load. This means that if you switch on a heavy load in your house and the mains on the input of the saturation transformer drops, it will compensate the voltage drop at the output. There are models on the marked that can keep the mains at +-1% instead of the normal +-10%* This is a great product, because it shields your precious tube fillaments from the effects of fluctuation in the mains voltage..
I know first hand from the owner of one of the Czech DHT manufacturers, that this tolerance is important, he actually suggested to me a 5.0V 300B should be run at 5.1V because underheating damages the tube, He told me over the Phone "You dont want to run a 2A3 at 2.3 Volts, its better to run it at 2.6V, it will last a LOT longer"
*This depends on what utility or state you live in, but this is just a wild guess at the allowed mains voltage tolerances state or federal regulators allow.
artosalo,
Thanks for posting your earlier thread of 2019.
The obvious conclusions of anyone who carefully studies that thread, are priceless.
A set of spectral measurements on a push pull DHT amplifier.
Measurements come through again.
The studies a co-worker and I conducted were very similar, only we tested a single ended DHT amplifier.
An old quote: "Seeing is Believing".
An LP album I have of Earl Garner is titled "Feeling is Believing".
Some people both see and feel, and they still do not believe.
(in my opinion, they can not be helped).
Thanks for posting your earlier thread of 2019.
The obvious conclusions of anyone who carefully studies that thread, are priceless.
A set of spectral measurements on a push pull DHT amplifier.
Measurements come through again.
The studies a co-worker and I conducted were very similar, only we tested a single ended DHT amplifier.
An old quote: "Seeing is Believing".
An LP album I have of Earl Garner is titled "Feeling is Believing".
Some people both see and feel, and they still do not believe.
(in my opinion, they can not be helped).
Last edited:
V4lve lover, I think I did not explain clearly what I meant. In capacitor input supply, current does not flow through transformer secondary during large portions of the sine wave. Current only flows when instantaneous voltage exceeds that of partially discharged capacitor, refills the capacitor, and is then cut off as instantaneous voltage drops below that of the capacitor. As a result, secondary current is not a sine wave, but a train of pulses. You can see it in PSUD2 if you look at secondary current waveform.
By contrast, secondary current of choke input supply IS sine wave.
These are basic of the basics of power supply design, you should have heard about it in your classes.
By contrast, secondary current of choke input supply IS sine wave.
These are basic of the basics of power supply design, you should have heard about it in your classes.
You are correct, If you run a spectral analysis on the current spikes you get some nasty harmonics into the Khz range. But this is not what causes the core to heat. Those are the quadratic current losses themselves. The Lion share of the loss contribution is under 1Khz
Im eyeballing this, i dont want to have to install Matlab to do a FFT on something like this.
I had trouble believing you would get actual HF as HF is defined by convention as 3-30Mhz. And those frequencies can only exist if for some reason the supply starts ringing, or you use some rather slow rectifiers.. In any case you should google the skin depth for OCF copper at those frequencies.
However these pulses and their possible interference are very manageable. For low current stuff you can block most if not all of the interference with two SMD 1206 parts. Furthermore, you can get some insanely good low ESR Low ESL caps nowadays, that take care of this junk right away.
Look up AN101F by Jim Williams for more.
One way to get a cleaner supply is to use a centre tapped transformer, that means more copper resistance that keeps the charging peaks in check somewhat.
I think your concerns are vastly overblown. But i also think a saturation transformer is not such a bad investment.
Im eyeballing this, i dont want to have to install Matlab to do a FFT on something like this.
I had trouble believing you would get actual HF as HF is defined by convention as 3-30Mhz. And those frequencies can only exist if for some reason the supply starts ringing, or you use some rather slow rectifiers.. In any case you should google the skin depth for OCF copper at those frequencies.
However these pulses and their possible interference are very manageable. For low current stuff you can block most if not all of the interference with two SMD 1206 parts. Furthermore, you can get some insanely good low ESR Low ESL caps nowadays, that take care of this junk right away.
Look up AN101F by Jim Williams for more.
One way to get a cleaner supply is to use a centre tapped transformer, that means more copper resistance that keeps the charging peaks in check somewhat.
I think your concerns are vastly overblown. But i also think a saturation transformer is not such a bad investment.
Last edited:
By HF I didn't mean radio frequencies, just everything above 2nd harmonic, plus transformer self-resonances. Eliminating them is not as easy as it seems because of multiple frequencies involved. Snubbers are calculated for certain frequencies.
For typical lamination thickness of power transformers, iron losses are very substantial at frequencies above the fundamental. A well-engineered transformer has close to 50:50 iron:copper losses at its nominal frequency.
Returning back to AC vs. DC filaments, I see a dilemma like this. I can keep my AC supply clean, feed filaments AC and mitigate residual hum, which will be largely the fundamental (an easy fix). Or, I can disregard PSU noise and stuff my chassis with regs, one for each filament. There is something awkward, not elegant, brute force in the second solution.
For typical lamination thickness of power transformers, iron losses are very substantial at frequencies above the fundamental. A well-engineered transformer has close to 50:50 iron:copper losses at its nominal frequency.
Returning back to AC vs. DC filaments, I see a dilemma like this. I can keep my AC supply clean, feed filaments AC and mitigate residual hum, which will be largely the fundamental (an easy fix). Or, I can disregard PSU noise and stuff my chassis with regs, one for each filament. There is something awkward, not elegant, brute force in the second solution.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.