On this forum, "SOA" is Safe Operating Area, the combinations of Power, Voltage, and Current that a transistor is safe to work at.
Asking an average power transistor to hold 90 Watts is like asking you to hold 90 kiloGrams all day long.
____________
I would think an air-core could be fine for the upper half of the audio band, especially for a compression driver which, in home use, would hardly-ever get over 1 Watt. (If you did peak to 20 Watts, ear-distortion would exceed amplifier distortion.) (Large PA systems are different; I have run way over 100W in large venues and outdoor concerts.)
As a bare minimum, the choke value could be similar to the value you would use in a simple passive crossover. 1mH at 1KHz is 6.3 Ohms, so you can work out 800Hz, 1.6KHz, 8, or 16 Ohms from there.
1mH air-core is a couple pounds of copper wire. Not an economy part, but cheap compared to any good compression driver.
If you use this value the load-line is very oval. Since the amplifier has no reserve signal current on the down-swing, power will be down. You want a larger value than bare-minimum.
PartsExpress lists several heavy air-cores. 7 mH 0.98 ohms DCR is 52 bucks, comparable to the iron-core for bass. For DIY without math, "multilayer air core inductor calculator" search brings up the tools. (Try to go near ID and length about the same, OD about 3X larger, but 2X either way is near optimum.)
"Infinite" seems good, but is not in stock. Any practical very-very-large choke will have high copper resistance. (Silver not much better, and that much Silver in private hands invites legal/banking investigation.)
DCR is not a vital parameter here as long as it is small compared to load. If much less than 1/4 of the load impedance, we can approximate the effect as a need for a slightly higher supply voltage, which is easily done. (If a perfect coil with 16V supply and 8 Ohm load gives 16 Watts, then a 2 Ohm DCR needs another 4V of supply or 20V, or essentially a common 19V laptop supply.)
Air-cores throw some field around. In the MoFo, I would not be too worried about left-right coupling because the MOSFET's impedance is so very low that it will absorb induced energy. Crossover builders do not have this luxury but do fine putting coils cross-ways to each other.
Asking an average power transistor to hold 90 Watts is like asking you to hold 90 kiloGrams all day long.
____________
I would think an air-core could be fine for the upper half of the audio band, especially for a compression driver which, in home use, would hardly-ever get over 1 Watt. (If you did peak to 20 Watts, ear-distortion would exceed amplifier distortion.) (Large PA systems are different; I have run way over 100W in large venues and outdoor concerts.)
As a bare minimum, the choke value could be similar to the value you would use in a simple passive crossover. 1mH at 1KHz is 6.3 Ohms, so you can work out 800Hz, 1.6KHz, 8, or 16 Ohms from there.
1mH air-core is a couple pounds of copper wire. Not an economy part, but cheap compared to any good compression driver.
If you use this value the load-line is very oval. Since the amplifier has no reserve signal current on the down-swing, power will be down. You want a larger value than bare-minimum.
PartsExpress lists several heavy air-cores. 7 mH 0.98 ohms DCR is 52 bucks, comparable to the iron-core for bass. For DIY without math, "multilayer air core inductor calculator" search brings up the tools. (Try to go near ID and length about the same, OD about 3X larger, but 2X either way is near optimum.)
"Infinite" seems good, but is not in stock. Any practical very-very-large choke will have high copper resistance. (Silver not much better, and that much Silver in private hands invites legal/banking investigation.)
DCR is not a vital parameter here as long as it is small compared to load. If much less than 1/4 of the load impedance, we can approximate the effect as a need for a slightly higher supply voltage, which is easily done. (If a perfect coil with 16V supply and 8 Ohm load gives 16 Watts, then a 2 Ohm DCR needs another 4V of supply or 20V, or essentially a common 19V laptop supply.)
Air-cores throw some field around. In the MoFo, I would not be too worried about left-right coupling because the MOSFET's impedance is so very low that it will absorb induced energy. Crossover builders do not have this luxury but do fine putting coils cross-ways to each other.
@PRR: what about the suitability of a powdered core type inductor like I liked to earlier? See:
Jantzen Audio 3.3mH 15 AWG P-Core Inductor Crossover Coil
Jantzen Audio 3.3mH 15 AWG P-Core Inductor Crossover Coil
Interpretation
Common belief, due to large application of ferrite core transformer in sub MHz’s SMPS, that ferrite is inappropriate for audio spectrum.
In fact it is not, as long as the core is unsaturated :
Air core transformer L=2.8mH
Ferrite transfo L=2.73mH
1) Bias Iq for best THD+N at 1KHz
Air core 2.5A
Ferrite 1A
2) THD+N for 0.3Vrms 1KHz input
Air core 0.054% @ -11.1dB output/8 Ohm dummy.
Ferrite 0.30% @ -11.2dB
Clearly those numbers indicated that my FX2243 ferrite core 2.73mH is inappropriate to build a MoFo.
Common belief, due to large application of ferrite core transformer in sub MHz’s SMPS, that ferrite is inappropriate for audio spectrum.
In fact it is not, as long as the core is unsaturated :
Air core transformer L=2.8mH
Ferrite transfo L=2.73mH
1) Bias Iq for best THD+N at 1KHz
Air core 2.5A
Ferrite 1A
2) THD+N for 0.3Vrms 1KHz input
Air core 0.054% @ -11.1dB output/8 Ohm dummy.
Ferrite 0.30% @ -11.2dB
Clearly those numbers indicated that my FX2243 ferrite core 2.73mH is inappropriate to build a MoFo.
Sorry to jump in
If Jantzen releases the “permite ” B-H spec, one can based on BH loop to work out the Le or “effective length” of an A/m or Ampere per meter to calculate core saturation point.
With a saturated core, the 3.3mH Jantzen is a pure DCR, within non saturated region it is a 3.3mH choke like any body else.
My case, the FX2243 is saturated at 1A @ 1kHz.
I connected this amp (3A limited power supply) to an Android music player:
Coil inductance went down with frequency, creating a pure DCR load to mosfet’s source.
Vicious circle: low frequency -> low inductance -> high Ids.
At any drum beat from my music, the PS will drop down to keep a safe preset current limit, hah hah a free built in Wah Wah pedal.
If Jantzen can guarantee a 2A flow @ 1KHz, i would buy it with a fine print “don’t less any drum beats in” (you are the
in DSP)Hi PRR,
Hope you had a nice week-end, mine was bounded indoor.
The 90 Watt was selected for best THD+N to that specified iron core and IRFP240, no thought about similarity in M.K.S.A.. (abréviation de mètre, kilogramme, seconde, ampère)
Average power dissipation in my chook shed
Attachments
Last edited:
After playing with this calculator a moment it looks like a decent inductor won't take all that much wire after all.
For a 1mh coil using # 16 wire on a 1" form .5" long would only use .46 lb of wire and have a resistance of 24ohm.
Or for a 3.1 mh coil wound on a 1" form .5" long using # 18 wire would only use .5 lb of wire and have a resistance of .67 ohms
Air Core Inductor Designer / Calculator
For a 1mh coil using # 16 wire on a 1" form .5" long would only use .46 lb of wire and have a resistance of 24ohm.
Or for a 3.1 mh coil wound on a 1" form .5" long using # 18 wire would only use .5 lb of wire and have a resistance of .67 ohms
Air Core Inductor Designer / Calculator
So a 4mH air core will get me to -1dB at 20Hz? That’s not bad at all. In fact, darn good for most speakers that I have.
Jantzen Audio 4.0mH 15 AWG Air Core Inductor Crossover Coil
Sound of air core should be better right?
Jantzen Audio 4.0mH 15 AWG Air Core Inductor Crossover Coil
Sound of air core should be better right?
Here is my air cored bifilar chokes I use in my split load phase splitter amplifier.
Split load phasesplitter amplifier with bifilar choke load
It is a variation of the MoFo. It has 2 times the gain and a second harmonic cancellation effect to lower THD. If one wants to get all fancy you could call it a feed forward error correction quite similar to Nelson Pass super-symmetry. If you want to drive a high frequency unit then you would probably only need two 30 meter cables to each choke.
With two times 100 meter I get a nice response down to 10 Hz from this amp.
Cheers,
Johannes
Last edited:
Do it if you like it.
I think iron (or ferrite) is the most problematic material in audio. It just won't be linear, and not even simple non-linear. It does make a major difference to the amount of copper. Also the side-spray. So it can be good, maybe better.
But "nothing" sounds good to me, and air is about 99.999% as good as nothing.
1mH is awful reactive for 1KHz and 8r load. You want the inductive reactance significantly higher than the load impedance. More like 3 to 7mH.
24 Ohms of DC resistance is far too high for 8 Ohm loading. We may as well use incandescent bulbs (while we can still get some) (didn't someone else try that?
) Or is it 0.24, typoed?
SMALL signal. The MOSFET has a very-very low small-signal output impedance, maybe a quarter-Ohm. This is indeed 0.004H at 10Hz.
But in a *power* amp, the choke impedance must be greater than the *load* impedance at the lowest frequency where we want big power. For 1KHz 8r this means more than 1.3mH, so really 3 to 7mH. Yet with DCR much less than load impedance, certainly under 2 Ohms.
Well, don't be fooled by the FR measurement alone. This example isn't a full range amp, for all the astute reasons which PRR provides, and the distortion goes quite high below 600 Hz, which is where the posted graphs start. The OP was asking about 1 kHz and up, and I think it wise to get well below whatever XO frequency is intended.
I would like to bark ..... but how long is the charge .... and how many volts that battery delivers?And, if you really want to bark at the moon you can power your compression driver amp with a battery.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.