In trouble needing lots of amperes? If that is what you mean, I definitely agree.You are so in trouble.....
In post #1 I stated that I am loathe to use a transformer. Induced phase shift. Added inductance, Hysteresis distortion. Added engineering job of designing the transformer, figuring out saturation, etc etc. Added expense.
The ribbon is kind of an exotic driver. The transformer would be a pedestrian solution to the problem of how to get enough current to drive the ribbon.
The exotic ribbon wants an exotic amplifier as a partner.
Acoustat went this route with a direct drive tube amplifier for their X speaker system. I have auditioned one of these systems at a friend's house.
Hill Plasmatronics made an exotic direct drive tube amp for their plasma tweeter that included a helium tank. I have not heard one but would love to. I have a friend who owned a pair and I did listen to his Martin Logan Monolith speakers. He liked the Hill plasma tweeter better than the treble of the Martin Logans.
A good ribbon tweeter is probably the closest thing to the plasma tweeter as a low mass tweeter. It would be a shame to put a transformer between the exotic tweeter and an amplifier.
Last edited:
A configuration of series stacked drivers is not a general solution to a general problem. It is only a particular solution to a particular problem/usage case. My goal is to have an amplifier that solves a general problem for many particular users.
I wonder if 300 Amp peaks will rattle your door-knobs with the radiated field.
Ribbon microphones need careful design of the return path; using "enormous" wire. Also bringing a wire back down both sides of the ribbon to cancel external fields.
Steady current of 300A wants US#00 Copper wire 0.365" diameter. Which is not a common wire- we use Aluminum in such sizes. For this and other reasons your return path is more likely many wires. (Or bar-stock, but it has no conductivity rating and some minor impurities really hurt that.)
Since your average current is well under 30A, #12 wire won't burn, but has too much resistance.
I had thought the amplifier obviously must be ON the loudspeaker. Well, no. Take ten #6 Copper wires 10 feet long, 0.4r per 1,000 feet. It will be below a milliOhm. So the amp can be in the base.
With a complement P-type, and some clever copper-work, I would be confident of delivering 9V at 300A to a 0.030 Ohm load. I *really* would go massively parallel. Not just because your devices are rated 200A lead-limit, but because the connections and reservoir caps need to be massively parallel so the parasitic resistances do not add-up so fast.
If a 8-Ohm amp takes two 10,000uFd caps, then a 0.03r amp takes two 2,666,000uFd (3F!) caps. DigiKey lists 15 caps 1F+, and will sell you a few 1F@25V for $80, but only 15 in stock (24 week lead-time). A bipolar stereo set is 8 or 12. A kiloBuck. Not prohibitive. ESR 5mOhms each, OK. Ripple current 40A, so it won't do 300A as a welder, but fine for speech/music.
The Gate capacitances are on the order of several nanoFarads, even as a Source Follower. Penciling 10nF, that's around 1K Ohms at 16KHz, 100r at 160KHz (where you need controlled gain for NFB stability). OK, that's actually in the zone of some $1 chips.
One thing-- unlike the vacuum forum, here you have almost no fear of shock (after you button the line wiring). You do have a VERY dangerous machine nevertheless. Get your wedding ring across those caps and it will burn through your finger then vapor-deposit Gold all over the room. Arc-blindness is possible. The repulsion of 300A peaks can tear-apart large transformers, or your cap-buses. Every high-current conductor will shake with the beat.
I understand your goal. But this is new territory. I would *really* think to use a transformer to run the speaker. Monitor the actual current/voltage demand for good level, or estimate the additional juice needed to reach a good level. Just finding that 150A is ample saves ~~3/4 the stuff.
Transformer appears to be a large core with an 8 Ohm winding (any primary, N/C) and 3 to 5 turns of the fattest wire you can slip around the coil-build without scraping on the iron. If there is space left, wind bi-filar, and work parallel for less loss.
Ribbon microphones need careful design of the return path; using "enormous" wire. Also bringing a wire back down both sides of the ribbon to cancel external fields.
Steady current of 300A wants US#00 Copper wire 0.365" diameter. Which is not a common wire- we use Aluminum in such sizes. For this and other reasons your return path is more likely many wires. (Or bar-stock, but it has no conductivity rating and some minor impurities really hurt that.)
Since your average current is well under 30A, #12 wire won't burn, but has too much resistance.
I had thought the amplifier obviously must be ON the loudspeaker. Well, no. Take ten #6 Copper wires 10 feet long, 0.4r per 1,000 feet. It will be below a milliOhm. So the amp can be in the base.
With a complement P-type, and some clever copper-work, I would be confident of delivering 9V at 300A to a 0.030 Ohm load. I *really* would go massively parallel. Not just because your devices are rated 200A lead-limit, but because the connections and reservoir caps need to be massively parallel so the parasitic resistances do not add-up so fast.
If a 8-Ohm amp takes two 10,000uFd caps, then a 0.03r amp takes two 2,666,000uFd (3F!) caps. DigiKey lists 15 caps 1F+, and will sell you a few 1F@25V for $80, but only 15 in stock (24 week lead-time). A bipolar stereo set is 8 or 12. A kiloBuck. Not prohibitive. ESR 5mOhms each, OK. Ripple current 40A, so it won't do 300A as a welder, but fine for speech/music.
The Gate capacitances are on the order of several nanoFarads, even as a Source Follower. Penciling 10nF, that's around 1K Ohms at 16KHz, 100r at 160KHz (where you need controlled gain for NFB stability). OK, that's actually in the zone of some $1 chips.
One thing-- unlike the vacuum forum, here you have almost no fear of shock (after you button the line wiring). You do have a VERY dangerous machine nevertheless. Get your wedding ring across those caps and it will burn through your finger then vapor-deposit Gold all over the room. Arc-blindness is possible. The repulsion of 300A peaks can tear-apart large transformers, or your cap-buses. Every high-current conductor will shake with the beat.
I understand your goal. But this is new territory. I would *really* think to use a transformer to run the speaker. Monitor the actual current/voltage demand for good level, or estimate the additional juice needed to reach a good level. Just finding that 150A is ample saves ~~3/4 the stuff.
Transformer appears to be a large core with an 8 Ohm winding (any primary, N/C) and 3 to 5 turns of the fattest wire you can slip around the coil-build without scraping on the iron. If there is space left, wind bi-filar, and work parallel for less loss.
20Vpp @ 300A peak output will only need perhaps a small ~13lbs transformer, a tiny hard to see smd sized ~3lbs for 10Vpp @ 150A peak, mundane stuff. As you said, an itty bit more interesting when something like 20 parallel Mofos are used instead. But at the moment, I don't think woofertester is ready to consider
Pass Forum class A solution. If he is, beware BA2018 organizer.
Thank you for making technical points about the original topic. This sort of reply is what I have been waiting for since my original post. Many of your points are exactly what I have been thinking.
The high current stuff should happen after the 1F capacitors. The transformer should not see the hundreds of amps unless it is kickback/flyback.
Copper bar stock for connecting drain and source of output FETs. Probably have to learn how to gold plate over nickel plate over the copper bar stock or send out the bar stock pieces to a professional gold plate house.
1F +/- of capacitance connected across the above bar stock across every power FET. The super capacitors may not have low enough ESR. Will have to experiment for selection of capacitors to live on top of the power FETs.
Understood about the safety concerns. It might be fun to hang a conductive mobile nearby to see it dance with the music. Or a rack of wind chimes.
Massively parallel is how the commercial amp solves the problem. I believe they use 10 pairs of 50A transistors.
For the proof of concept, one pair should suffice to take measurements to help calculate full-scale implementation.
I would like to have a workable transformer design to satisfy those people who are willing to live with the transformer artifacts.
Nice inputs!
Hello folks !
All yours assumptions are based on the dB/m/W estimation . You compare an honorable 1/2 inch Audax with a half space spherical wave with a 1/2 inch and 10inchs long ribbon tweeter with a more directive cylindrical wave , that's not the same distance attenuation . You can at normal listening distance add a good 10dB . Musical signal is poor over 3kHz and power requirement are 6 to 10dB Lower . Real world listening is in the 70/90 dBA range and that's rarely more that 10W peak with classical loudspeaker .
All in all you don't need to flash your tweeter with 500 amps . Think , most of the power will be wasted in thermal energy , can your aluminum foil ... yes it is not a toaster .
Your first assumption with 20V 40A is quite good looking , the foil dissipate at the max 800W quite hot enough . Power requirement stay in the high side but are more realist .
You possibly need less , but the solutions are on the acoustic side .
Limit the acoustic shortcut , the use of a wave guide or horn , a longer ribbon ...all that was explored .
Simplex2
All yours assumptions are based on the dB/m/W estimation . You compare an honorable 1/2 inch Audax with a half space spherical wave with a 1/2 inch and 10inchs long ribbon tweeter with a more directive cylindrical wave , that's not the same distance attenuation . You can at normal listening distance add a good 10dB . Musical signal is poor over 3kHz and power requirement are 6 to 10dB Lower . Real world listening is in the 70/90 dBA range and that's rarely more that 10W peak with classical loudspeaker .
All in all you don't need to flash your tweeter with 500 amps . Think , most of the power will be wasted in thermal energy , can your aluminum foil ... yes it is not a toaster .
Your first assumption with 20V 40A is quite good looking , the foil dissipate at the max 800W quite hot enough . Power requirement stay in the high side but are more realist .
You possibly need less , but the solutions are on the acoustic side .
Limit the acoustic shortcut , the use of a wave guide or horn , a longer ribbon ...all that was explored .
Simplex2
Understood. If maximizing ribbon sensitivity were my goal, I would have aired this in the exotic speakers forum. I am happy with the ribbon as-is. It is the result of multiple generations of design. The ribbon is simplified for construction, cost and repair.
I posted here because this is the forum of amplifier excess and I was looking for an excessive amplifier solution.
I agree about the capacitance. Measuring the capacitance curves is on the list of things to do sooner than later.
Somewhere I read that Elektor Output amplifier for ribbon loudspeakers was transformed from originaly circuit lowering from +/-45v to +/-15v rails,parallel buy double source resistors and as remember correctly connect one speaker out to ground and the other out to transformer center tap , making it V to I amplifier with 30a current to drive 0,4 ohm ribbons.
Is that correct ?
If it is, maybe Jan Didden can give us advise is it possible to transform his Pax amp on the same manner for use with ribbon wide range.
references;
Elektor's "Ulti Amp" post#3
Output amplifier for ribbon loudspeakers (1): | Elektor Magazine
Is that correct ?
If it is, maybe Jan Didden can give us advise is it possible to transform his Pax amp on the same manner for use with ribbon wide range.
references;
Elektor's "Ulti Amp" post#3
Output amplifier for ribbon loudspeakers (1): | Elektor Magazine
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.