Amps for (really) high voltage
Hi Bob,
Good to bump into you at the Montreal Audio show - was it in front of the MBL room???
On to my question - I'd like to try building up a SS amp to directly power some electrostatic panels. Let's say with rail voltages upwards of plus/minus 500 volts or more. Possible or not???
Charles
Hi Bob,
Good to bump into you at the Montreal Audio show - was it in front of the MBL room???
On to my question - I'd like to try building up a SS amp to directly power some electrostatic panels. Let's say with rail voltages upwards of plus/minus 500 volts or more. Possible or not???
Charles
Charles, 500v +/- voltage rails for solid state amplifier?
Even say 600 watt per channel have supply rails of 80v +/- dc...
Even say 600 watt per channel have supply rails of 80v +/- dc...
if you check the net for information regarding your needs you many come across what your looking for.
4 complementary 500Vce0 BJTs in cascode could possibly output maximum signals of around 1000Vpp
This would be just as dangerous as an 845 powered amp.
Maybe it could all be integrated into the speaker base/support for safer operation?
This would be just as dangerous as an 845 powered amp.
Maybe it could all be integrated into the speaker base/support for safer operation?
No problem. Just wear very thick rubber gloves whilst troubleshooting....
😀
My father used to work on high power transmitters. I remember on one occasion, a high voltage/frequency arc "escaped" the chassis and literally vaulted him out of his chair and across the radio shack. He lost consciousness briefly and was lucky to have survived. There's some good reasons for using transformers at the load. Having difficulty finding someone to work on faulty high voltage hardware might be one of them.
+/- 500 V is just electrostatic headphone amp territory - for loudspeakers you may need lots more
Mosfets and new SiC power fets can handle KV but no complementary devices exists at the highest V levels
Mosfets and new SiC power fets can handle KV but no complementary devices exists at the highest V levels
Hi Bob,
Looking at your web site I see mention of a C70 tube amplifier..........would you care to shed more light on it. It sounds that it might be a great amp but I can't
seem to find out much about it.
Regards,
Jam
Looking at your web site I see mention of a C70 tube amplifier..........would you care to shed more light on it. It sounds that it might be a great amp but I can't
seem to find out much about it.
Regards,
Jam
Hi Charles,
Its good to hear from you. The Montreal show was fun. Great City!
This is a tall order. I'd lean toward complementary MOSFETs, but the highest voltage P-chaanel MOSFET I know of is about 400V. If you stack two of them and pair them with stacked N-channels, you can have rails up to perhaps +/- 300V. Then if you make the amplifier bridged, you can get maybe 600V peak into the load.
What kind of maximum current levels do you anticipate? Safe area might be a real concern. Then there is the high-voltage IPS/VAS circuit...
Cheers,
Bob
Hi Bob,
Looking at your web site I see mention of a C70 tube amplifier..........would you care to shed more light on it. It sounds that it might be a great amp but I can't
seem to find out much about it.
Regards,
Jam
Hi Jam,
Funny you should ask about the C70. I'm right in the middle of writing an article on it that will include plenty of detail. When it is about to be published, I'll mention the publication.
Cheers,
Bob
Bob,
I had similar thoughts about a tube amp design and I am sure you will give some of us here much to think about............I hope to read it soon and fill in some missing pieces I have questions about. Keep up the good work.
Thanks.
Regards,
Jam
I had similar thoughts about a tube amp design and I am sure you will give some of us here much to think about............I hope to read it soon and fill in some missing pieces I have questions about. Keep up the good work.
Thanks.
Regards,
Jam
RC filter location
Bob, and all the others too
A little practical question : imagine the connection trace between the output of the IPS (with current mirrors) and the input of the VAS (cascoded darlingtoned). I know this trace is very important as it can be source of stability problems. Therefore I got convinced to insert a serial RC circuit connected between that node and the DC rail.
1) Is it of any matter to connect both components in a certain sequence? I mean capacitor connected to rail and resistor to node or vice versa? Or isn't it of any importance?
2) Since the trace is long and connects to many components (2Q's on IPS side, 1Q on VAS side, various bakerclamp diodes, miller cap 1Q as VAS Iq limiter, etc.) is there any importance to place the RC network close to the IPS side or rather close to the VAS side?
I don't want to clutter this thread so I placed circuit and pcb in my own thread :
http://www.diyaudio.com/forums/soli...ogy-construction-troubles-36.html#post2430925
Cheers,
(black? Seems like Guinness beer)
Olivier
Bob, and all the others too
A little practical question : imagine the connection trace between the output of the IPS (with current mirrors) and the input of the VAS (cascoded darlingtoned). I know this trace is very important as it can be source of stability problems. Therefore I got convinced to insert a serial RC circuit connected between that node and the DC rail.
1) Is it of any matter to connect both components in a certain sequence? I mean capacitor connected to rail and resistor to node or vice versa? Or isn't it of any importance?
2) Since the trace is long and connects to many components (2Q's on IPS side, 1Q on VAS side, various bakerclamp diodes, miller cap 1Q as VAS Iq limiter, etc.) is there any importance to place the RC network close to the IPS side or rather close to the VAS side?
I don't want to clutter this thread so I placed circuit and pcb in my own thread :
http://www.diyaudio.com/forums/soli...ogy-construction-troubles-36.html#post2430925
Cheers,
(black? Seems like Guinness beer)
Olivier
Bob, and all the others too
A little practical question : imagine the connection trace between the output of the IPS (with current mirrors) and the input of the VAS (cascoded darlingtoned). I know this trace is very important as it can be source of stability problems. Therefore I got convinced to insert a serial RC circuit connected between that node and the DC rail.
1) Is it of any matter to connect both components in a certain sequence? I mean capacitor connected to rail and resistor to node or vice versa? Or isn't it of any importance?
2) Since the trace is long and connects to many components (2Q's on IPS side, 1Q on VAS side, various bakerclamp diodes, miller cap 1Q as VAS Iq limiter, etc.) is there any importance to place the RC network close to the IPS side or rather close to the VAS side?
I don't want to clutter this thread so I placed circuit and pcb in my own thread :
http://www.diyaudio.com/forums/soli...ogy-construction-troubles-36.html#post2430925
Cheers,
(black? Seems like Guinness beer)
Olivier
http://www.diyaudio.com/forums/soli...ogy-construction-troubles-33.html#post2415670
http://www.diyaudio.com/forums/soli...ogy-construction-troubles-29.html#post2408546
Wahab,
I am aware of the difficulties and did not forget your previous comments. However, I am willing to give this another try. Even if I would use another circuit, the question of putting my compensation RC circuit near the IPS or near the VAS remains.
Forget the topology for a while.
Thanx
Olivier
I am aware of the difficulties and did not forget your previous comments. However, I am willing to give this another try. Even if I would use another circuit, the question of putting my compensation RC circuit near the IPS or near the VAS remains.
Forget the topology for a while.
Thanx
Olivier
Bob, and all the others too
A little practical question : imagine the connection trace between the output of the IPS (with current mirrors) and the input of the VAS (cascoded darlingtoned). I know this trace is very important as it can be source of stability problems. Therefore I got convinced to insert a serial RC circuit connected between that node and the DC rail.
1) Is it of any matter to connect both components in a certain sequence? I mean capacitor connected to rail and resistor to node or vice versa? Or isn't it of any importance?
2) Since the trace is long and connects to many components (2Q's on IPS side, 1Q on VAS side, various bakerclamp diodes, miller cap 1Q as VAS Iq limiter, etc.) is there any importance to place the RC network close to the IPS side or rather close to the VAS side?
I don't want to clutter this thread so I placed circuit and pcb in my own thread :
http://www.diyaudio.com/forums/soli...ogy-construction-troubles-36.html#post2430925
Cheers,
(black? Seems like Guinness beer)
Olivier
Hi Olivier,
The series R-C network you added from the VAS input node to the rail, consisting of 100pF and 100 ohms - what is its purpose and how does it affect the simulation results?
With small values like those, which one connects to the rail should not make any difference.
BTW, it looks like your use of C17 and R35 connecting the two VAS inputs may create problems and may hurt PSRR - what is their purpose?
Cheers,
Bob
Hi Bob,
The little circuit between VASes C17/R35 is in fact a remainder of an optional. Maybe I should be consequent and remove it from the routing since it will probably not be used.
As for the RC network between IPS and VAS it is needed for stability at HF. Later this evening I will show you the AC simulation with & w/o this network implemented. The impact is quite large... the origin of placing this network comes from another source but reading your book it seems something like you used in Miller Input Compensation. There you also placed them for stability reasons ... I hope the simulation I will send tonight makes it clear...
Cheers,
Olivier
The little circuit between VASes C17/R35 is in fact a remainder of an optional. Maybe I should be consequent and remove it from the routing since it will probably not be used.
As for the RC network between IPS and VAS it is needed for stability at HF. Later this evening I will show you the AC simulation with & w/o this network implemented. The impact is quite large... the origin of placing this network comes from another source but reading your book it seems something like you used in Miller Input Compensation. There you also placed them for stability reasons ... I hope the simulation I will send tonight makes it clear...
Cheers,
Olivier
Hi Olivier,
With TMC working properly from output to input of the VAS, I wouldn't think this network would be necessary, but maybe I am missing something. I don't think it is analogous to MIC.
Cheers,
Bob
RC
Hi Bob,
With a day of delay, here are the simulation results (AC) for the circuit with and without RC network C12/R25 - C13/R27.
The difference is a little smaller as I remembered but however, it eliminates the spike in the magnitude plot @ 50MHz.
With RC
PM : 89°
GM : 35dB
Without RC
PM : 90°
GM : 25dB
See graphs attached.
Is this worth installing the RC network? Or is the positive aspect of it negated by the added complexity and other issues accompagnied with it?
Cheers,
Olivier
Hi Bob,
With a day of delay, here are the simulation results (AC) for the circuit with and without RC network C12/R25 - C13/R27.
The difference is a little smaller as I remembered but however, it eliminates the spike in the magnitude plot @ 50MHz.
With RC
PM : 89°
GM : 35dB
Without RC
PM : 90°
GM : 25dB
See graphs attached.
Is this worth installing the RC network? Or is the positive aspect of it negated by the added complexity and other issues accompagnied with it?
Cheers,
Olivier
The additional RC network looks pointless to me because:
1. The only change is at 25 dB below unity gain
2. At 50 MHz your Spice model is not accurate
1. The only change is at 25 dB below unity gain
2. At 50 MHz your Spice model is not accurate
It only seems pointless if you are looking at the wrong loop. 😀
There are more loops, you know!
There are more loops, you know!