Mentor Pass,
I stumbled on an old (July 2002) DIYaudio thread regarding use of MOSFETS to direct drive Electrostatic Loudspeaker panels:
http://www.diyaudio.com/forums/solid-state/4762-esl-amp-switching-amp.html
In the thread, you noted:
"If you can get those MOSFETs to switch at the voltages
and speeds needed to direct drive ESLs, then you will
become an audio divinity, and temples will be erected in
your honor. And I will visit them."
That was 13 years ago.
Are we there yet? With the current generation of SIT and SIC MOSFETS (or even IGBTs) can a direct drive, tube-less, transformer-less quality amp be built with existing technology? I kinda envision a "stack" of high-voltage MOSFETS to achieve ESL voltages (5-6 kv or so).....
Is it time for the divinity to appear? (And do you have a schematic?) 🙂
Greedy boyz want to know.......... 🙄
I stumbled on an old (July 2002) DIYaudio thread regarding use of MOSFETS to direct drive Electrostatic Loudspeaker panels:
http://www.diyaudio.com/forums/solid-state/4762-esl-amp-switching-amp.html
In the thread, you noted:
"If you can get those MOSFETs to switch at the voltages
and speeds needed to direct drive ESLs, then you will
become an audio divinity, and temples will be erected in
your honor. And I will visit them."
That was 13 years ago.
Are we there yet? With the current generation of SIT and SIC MOSFETS (or even IGBTs) can a direct drive, tube-less, transformer-less quality amp be built with existing technology? I kinda envision a "stack" of high-voltage MOSFETS to achieve ESL voltages (5-6 kv or so).....
Is it time for the divinity to appear? (And do you have a schematic?) 🙂
Greedy boyz want to know.......... 🙄
Last edited:
We have Mosfets which will withstand 1700V (Silicon Carbide) which is a
definite improvement in our prospects.
So I can imagine balanced systems which would deliver DC coupled
+/- 1500 V or so peaks across the stators.
Do you think that would do it?
😎
definite improvement in our prospects.
So I can imagine balanced systems which would deliver DC coupled
+/- 1500 V or so peaks across the stators.
Do you think that would do it?
😎
This could be close. A common step-up ratio for the input transformers in ESLs is 75:1 to 100:1. With the Vsupply of an amplifier at +\- 30 to 50V this means we are about half way there at full power, and all the way there at reasonable listening levels.
This one can take single 3200V rail and output +/-3000V at its differential outputs :
http://www.diyaudio.com/forums/pass-labs/153030-semisouth-jfets-what-they-good-3.html#post2178227
And here is a lower voltage version for ESL headphones using currently available devices :
http://www.diyaudio.com/forums/pass-labs/248112-pass-stat-headphone-amplifier-concept.html
And if you are willing to use MOSFETs, there are many devices that can go to 2500V and even 4500V these days, i.e. 5000V or 9000V rail.
IXTH02N250 IXYS | IXTH02N250-ND | DigiKey
IXTF02N450 IXYS | IXTF02N450-ND | DigiKey
With TO247 packages, you can dissipate 30W per device.
That would be 12mA at 2500V Vds. You can alway parallel devices to get more bias.
Patrick
http://www.diyaudio.com/forums/pass-labs/153030-semisouth-jfets-what-they-good-3.html#post2178227
And here is a lower voltage version for ESL headphones using currently available devices :
http://www.diyaudio.com/forums/pass-labs/248112-pass-stat-headphone-amplifier-concept.html
And if you are willing to use MOSFETs, there are many devices that can go to 2500V and even 4500V these days, i.e. 5000V or 9000V rail.
IXTH02N250 IXYS | IXTH02N250-ND | DigiKey
IXTF02N450 IXYS | IXTF02N450-ND | DigiKey
With TO247 packages, you can dissipate 30W per device.
That would be 12mA at 2500V Vds. You can alway parallel devices to get more bias.
Patrick
The load that an ESL panel would provide to this amplifier is right in line with these devices.
There are two main types of ESL panels and the loads differ. In an unsectioned panel, the front and back stators create a capacitor with air as a dilectric. A 12"x48" panel would have a capacitance of about 500pF. The intrinsic SPL response of this type of panel has a 10dB/decade upward slope.
To counteract this, a feed resistor is used to create a low-pass RC filter. Typically this resistor is placed on the primary of the transformer, but in this case there would be no transformer so the feed resistor would be in the range of a few hundred K to a couple of M Ohms.
For a segmented ESL, the panel is divided into many vertical sections, with these sections separated by resistors to form a large RC ladder (see diagram below, from user bolserst). By having the highest frequencies on just the thinnest center section, horizontal dispersion is much improved. Again, there is a feed resistor, but is is of a much lower value. In my case, the R is about 350K and the C is about 70pF, making the feed resistor about 150K.
With a +/- 3000V supply and 150K of feed resistance, we are looking at about 20mA drive current required. Based on EVULs post, this is in line with what these devices can handle if paralleled.
There are two main types of ESL panels and the loads differ. In an unsectioned panel, the front and back stators create a capacitor with air as a dilectric. A 12"x48" panel would have a capacitance of about 500pF. The intrinsic SPL response of this type of panel has a 10dB/decade upward slope.
To counteract this, a feed resistor is used to create a low-pass RC filter. Typically this resistor is placed on the primary of the transformer, but in this case there would be no transformer so the feed resistor would be in the range of a few hundred K to a couple of M Ohms.
For a segmented ESL, the panel is divided into many vertical sections, with these sections separated by resistors to form a large RC ladder (see diagram below, from user bolserst). By having the highest frequencies on just the thinnest center section, horizontal dispersion is much improved. Again, there is a feed resistor, but is is of a much lower value. In my case, the R is about 350K and the C is about 70pF, making the feed resistor about 150K.
With a +/- 3000V supply and 150K of feed resistance, we are looking at about 20mA drive current required. Based on EVULs post, this is in line with what these devices can handle if paralleled.
Attachments
Thanks, gents.
I agree with you. I've done some cursory research on semi advances, and I think we are at the "starting point" for a direct-drive semiconductor ESC amp. Pretty exciting stuff.
Time for me to look at some specific semi data sheets, and press on from there. If you'all have any suggestions for circuits, I'd be interested. I think I'll need to move further threads to the Solid State or Exotic Speaker forums, unless you can see a Class A ESL amp design that would be practical.
We live in exciting times. I have some "retired" Martin Logan Sequel panels, and would love to use them, along with a good tuned port woofer or sub design. I doubt if I have the "smarts" to develop a quality amp design, but I still have a dream to try.
I agree with you. I've done some cursory research on semi advances, and I think we are at the "starting point" for a direct-drive semiconductor ESC amp. Pretty exciting stuff.
Time for me to look at some specific semi data sheets, and press on from there. If you'all have any suggestions for circuits, I'd be interested. I think I'll need to move further threads to the Solid State or Exotic Speaker forums, unless you can see a Class A ESL amp design that would be practical.
We live in exciting times. I have some "retired" Martin Logan Sequel panels, and would love to use them, along with a good tuned port woofer or sub design. I doubt if I have the "smarts" to develop a quality amp design, but I still have a dream to try.
Hi,
Against an amplifier for dynamic speakers, where the loading speaker is rather irrelevant, the requirements for a DD-ESL amp vary widely.
Basically it depends on the ESL panels efficiency and type of construction.
Fullrangers are inherently low efficiency and as such can be ruled out, as voltage and power requirements are way too much.
Leaves us with hybrid panels, which could be divided into electrically segmented and non-segmented. Both of which may be driven directly to decent SPLs with amps of >2kV supply lines.
The non-segmented panel though confronts the amp with such high capacitances that the current and power demand towards the upper audio bandwidth become rather non-practical.
The electrically segmented hybrid panel seems the only one left where a direct drive amp could actually make some sense.
Depending on the segmentation it may require just 1/10 of the current and power to drive a same-sized non-segmented panel.
It remains questionable though if there is any benefit against a decent transformer driven panel.
jauu
Calvin
Certainly not ... that amp will still be a devilish machine 😉
Against an amplifier for dynamic speakers, where the loading speaker is rather irrelevant, the requirements for a DD-ESL amp vary widely.
Basically it depends on the ESL panels efficiency and type of construction.
Fullrangers are inherently low efficiency and as such can be ruled out, as voltage and power requirements are way too much.
Leaves us with hybrid panels, which could be divided into electrically segmented and non-segmented. Both of which may be driven directly to decent SPLs with amps of >2kV supply lines.
The non-segmented panel though confronts the amp with such high capacitances that the current and power demand towards the upper audio bandwidth become rather non-practical.
The electrically segmented hybrid panel seems the only one left where a direct drive amp could actually make some sense.
Depending on the segmentation it may require just 1/10 of the current and power to drive a same-sized non-segmented panel.
It remains questionable though if there is any benefit against a decent transformer driven panel.
jauu
Calvin
Well, I have a mile roll of mylar and a gallon of coating and a 10KV supply for
bias. Maybe one of these days Jam will help me build some panels.
😎
bias. Maybe one of these days Jam will help me build some panels.
😎
Gee..... I can see a whole new Nelson Pass product line emerging..... Joining "First Watt" will be "First Kilovolt".......!
When can I DIY one???? Can I build a pair for less than a case of Petrus??
My mind is boggled.....
......hopefully soon Mr.Pass.
Here is something that might be adapted for even higher voltages using the silicon carbide mosfets?
Jam
Here is something that might be adapted for even higher voltages using the silicon carbide mosfets?
Jam
1. you two are nutz (Pa & Jam ; CanAm isn't far ... )
2. be careful , Jam, someone will certainly chime in , saying that HI_V_AMP is nothing else than HiV F5
2. be careful , Jam, someone will certainly chime in , saying that HI_V_AMP is nothing else than HiV F5
Jam,
Thanks for the schematic! The "stack" of MOSFETS was pretty much what I had been pondering. Interesting, too, that this article came out in 2004, using the technologies of a decade ago(!). If more robust hi-voltage devices can be adapted to this topology, I can see where swings of +/- 5 KV may be feasible......
ZM.... when you say I'm borderline crazy, I take that as a compliment!!!
Thanks for the schematic! The "stack" of MOSFETS was pretty much what I had been pondering. Interesting, too, that this article came out in 2004, using the technologies of a decade ago(!). If more robust hi-voltage devices can be adapted to this topology, I can see where swings of +/- 5 KV may be feasible......
ZM.... when you say I'm borderline crazy, I take that as a compliment!!!
Last edited:
Hhhhmmmm.......
I just looked at the tech data sheet for the MTP2P50E (schematic attached to JAM's post)
Datasheet here: http://www.onsemi.com/pub_link/Collateral/MTP2P50E-D.PDF
Characteristics look kinda ugly for an amp, though..... More research (and searching) to follow, I guess.
I just looked at the tech data sheet for the MTP2P50E (schematic attached to JAM's post)
Datasheet here: http://www.onsemi.com/pub_link/Collateral/MTP2P50E-D.PDF
Characteristics look kinda ugly for an amp, though..... More research (and searching) to follow, I guess.
Zen Mod,
I am just but a humble student of Master Pass ......
I am sure the schcamatic I posted can be updated with better devices higher rails and a discreet front end..............any volunteers?😉
Jam
I am just but a humble student of Master Pass ......
I am sure the schcamatic I posted can be updated with better devices higher rails and a discreet front end..............any volunteers?😉
Jam
On a different note, a patent of some interest. Note the bias supply.
CanAm Man, we might have to volunteer you for this ESL amplifier project.😀
Jam
CanAm Man, we might have to volunteer you for this ESL amplifier project.😀
Jam
Uncanny, Jam. I stumbled upon the same patent info about two hours ago. (Truly, the Internet is our friend!)
Having built ESL panels and amps (to include Pass DIY designs) in the past, I'd love to give it a shot...... unfortunately, I'm in the midst of (re)building a 1970's McLaren Can-Am race car (hence and ergo, my screen name on this forum). Having said that, if we can develop a good (contemporary MOSFET) design that would be compatible with my existing (Martin Logan Sequel) panels, this would give me something to do in my free time.
(I'm retired. Whoever said "retirement" gives you time to slow down, must not be retired!)
- Status
- Not open for further replies.