Circlotron said:
Circlotron,
my mistake, about your name. As far as the view program not working I don't get it.
Once again this russian article makes me wonder how many proprietary schematics are around on russian and chinese sites where copyrights laws are more shall we say 'relaxed', and that we'll never know about?
grataku said:
grataku,
if you can read russian, you would notice that this particular schematics a) published by the original idea author and b) with full knowledge and approval from Creek Audio Ltd. Mind you, I did develop this circuit before I've started to work for Creek and 4240 schematics is just an example of a product using the idea. There were some later developments of this circuit which I will not disclose as these used in a current line of products and do belong to the company. However the original idea belongs to me and I am quite happy to share it.
Al
Actually I missed out something on the schematic. There is a small bifilar wound transformer winding from gate to source on each fet opposite phase to each other that makes sure that each fet gets exactly the opposite drive to the other one, no matter what. Blocking cap inseries with each winding too.
Circlotron,
ah, fantasies, dreams, nightmares
. First circuit will work of sorts, but the symmetry is gone completely, as you've probably noticed. Second idea actually was implemented ages ago both on fets and bipolars. It works fine, but because of the transformers you'll have all kinds of small problems (some people treat these as advantages ). Did you have a chance to work out how my circuit works and why it is very symmetrical even in class B? If not, just ask me and I will explain in details .
Al
ah, fantasies, dreams, nightmares
. First circuit will work of sorts, but the symmetry is gone completely, as you've probably noticed. Second idea actually was implemented ages ago both on fets and bipolars. It works fine, but because of the transformers you'll have all kinds of small problems (some people treat these as advantages ). Did you have a chance to work out how my circuit works and why it is very symmetrical even in class B? If not, just ask me and I will explain in details . Al
Hi x-pro. When I ran the first cct the symmetry seemed really good. For example, under a particular load the gate drive to one fet was 700mV p/p and the other was 704mV p/p. I could live with that!
Or do you mean something more sinister? For years I looked at these kind of cct's and because I didnt understand then very well I always thought they relied on NFB to straighen out otherwise woeful asymmetry. Now I know that if you do it properly it can be as symmetrical as a comp symm cct. Yes, I do think I understand your cct but from memory in the hand drawn diagram the the two diodes hanging off the upper ccs are pointing opposite directions whereas on the pic from the Russian site they are pointing the same way.
So I must have missed something. Therefore, please fill me in on the class B and symmetry thing. I am most interested!! 
Back in post #7 of this thread http://diyaudio.com/forums/showthread.php?postid=161958#post161958 I gave a TL783 voltage reg based cct that would swing 120v p/p output. I knew we could do a little better so I went back to an LM317 (my favourite 
) and cascoded it with a high voltage fet. I had previously used this cct as a high compliance current sink - it has a slope resistance of > 400 megohms at 1 amp - and as a result I reckoned it would be fun as an amplifier too. The pic says it all really. The output positive swing is limited by the supply voltage and the negative swing by how far up the "cathode/source/emitter" is pulled. As you can see, it cranked out 400 volts p/p sinewave and looked as clean as you like. No visible distortion till about 50kHz. At 20kHz and 400v p/p the slew rate is just a little over 50 volts/uS. Now just what am I going to do with it? I dunno. Maybe add a cascoded fet/'317 follower with a constant current sink load and use that to drive the grid of a power tube into the positive (grid current region)? Maybe I'll try a class A2 single tube power amp just for a bit of a giggle. All I need is an output tranny...
Anyway, something just happened. "click".
/Circlotron - left-clicks the Submit Reply button and then leans back in his chair and lets out a long sigh.
) and cascoded it with a high voltage fet. I had previously used this cct as a high compliance current sink - it has a slope resistance of > 400 megohms at 1 amp - and as a result I reckoned it would be fun as an amplifier too. The pic says it all really. The output positive swing is limited by the supply voltage and the negative swing by how far up the "cathode/source/emitter" is pulled. As you can see, it cranked out 400 volts p/p sinewave and looked as clean as you like. No visible distortion till about 50kHz. At 20kHz and 400v p/p the slew rate is just a little over 50 volts/uS. Now just what am I going to do with it? I dunno. Maybe add a cascoded fet/'317 follower with a constant current sink load and use that to drive the grid of a power tube into the positive (grid current region)? Maybe I'll try a class A2 single tube power amp just for a bit of a giggle. All I need is an output tranny...Anyway, something just happened. "click".
/Circlotron - left-clicks the Submit Reply button and then leans back in his chair and lets out a long sigh.
Attachments
Circlotron said:
Hi Circlotron,
diodes on the hand drawn pic are the same + one more added in series with the p-ch transistor's source, to protect the output fets from a direct current if overvoltaged on the input.
On your circuit you may have a symmetry problem on low frequencies because of a capacitive bootstrap. This circuit looks more like a circuit by a Swedish engineer (I'll look up his name later at home), published in Electronics World in 1994. Also it won't work with logic level fets as you will have not enough voltage on the regulator. Even with a standard IRF540's you will have very little headroom, so if you'll try to drive it on load (did you simulate it on load?), I think you may find that the bottom transistor will not get enough voltage to produce some decent current. That s my impression, anyway
.Al
Understood. Actually, I never simulate anything outside of my head. I can't figure out those simulators. All numbers are from real hardware. Actually, I did measure 24 v p/p into 6 ohms load with +/- 17 volt rails. I do know the problem you mean. Of course, the cct ain't finished yet.x-pro said:
If your 317/High Voltage FET circuit could stand a capacitive load, it might be a really simple driver for an electrostatic panel.
I've been following this thread for a while with interest... I can't wait to see where you plan to use a 500 volt audio driver. (You might use it as a solid state replacement for a triode final in a tube amp...)
Joe L.
I've been following this thread for a while with interest... I can't wait to see where you plan to use a 500 volt audio driver. (You might use it as a solid state replacement for a triode final in a tube amp...)
Joe L.
Sure! Use two of them as a long tailed pair for a balanced output. You may want to lower the output impedance by lowering the resistors and upping the current depending on the size of the fet.
Actually it is *extremely* pentode-ish in that the "Rp" is greater than 400 megohms. The transconductance is set by (and equal to) the bottom resistor. For a given bias voltage, when you raise the "anode" voltage the "anode" current goes up up up to a certain point and then just levels off dead flat. It approaches the ideal pentode. Again, tie the two "cathodes" together with a common resistor to earth to make a LTP, bias them to class A, drive only one "grid", hang them off each end of an output tranny and... woohoo! If I had a suitable tranny I would definitely try this one. I have several 900v fets from an old SMPS that would do just fine. They would be good for ES panel drive to actually.
Circlotron said:
Woo. The things you can do with LM317's. Is the next step to use the big TO-3 ones as output devices? ;-O
I'm a bit worried that when the output swings to above the rails, your + (or -) 15V regs will have its input lower (or higher) than its output - are you happy that's OK, and doesn't e.g. burst into oscillation?
It's also not entirely clear where the stability margin at HF comes from. Presumably one of the devices in the feedback loop provides a dominant pole?
Cheers
IH
Very interesting!
I was not aware of those threads when I created these:
http://www.diyaudio.com/forums/chip-amps/175457-just-fun-regulator-chip-amplifier.html
http://www.diyaudio.com/forums/chip...oss-tringlotron-regulator-chip-amplifier.html
http://www.diyaudio.com/forums/chip-amps/176052-now-regulator-chip-jlh-amp.html
http://www.diyaudio.com/forums/chip...ator-chip-amp-family-welcomes-new-member.html
I was not aware of those threads when I created these:
http://www.diyaudio.com/forums/chip-amps/175457-just-fun-regulator-chip-amplifier.html
http://www.diyaudio.com/forums/chip...oss-tringlotron-regulator-chip-amplifier.html
http://www.diyaudio.com/forums/chip-amps/176052-now-regulator-chip-jlh-amp.html
http://www.diyaudio.com/forums/chip...ator-chip-amp-family-welcomes-new-member.html
Elvee is Circlotron's new best friend
I think we may have lived parallel lives.
I found ST LM317 was very good.
Have a look at this, also 7-1/2 years ago
http://www.diyaudio.com/forums/soli...att-class-amp-using-voltage-regulator-ic.html
I think we may have lived parallel lives.
I found ST LM317 was very good.
Have a look at this, also 7-1/2 years ago
http://www.diyaudio.com/forums/soli...att-class-amp-using-voltage-regulator-ic.html
Wow!
Quite impressive!
And I thought I had come up with something new
Yeah, with a seven year time-shift.
They are the ones I happened to use. Not a deliberate choice, that just what I had available when I built my prototypes.
They are fine for class A, but their slew-rate is too low for a proper class B operation. What a pity.
Another approach is to use a smaller emitter resistor and put a diode across it (Baxandall diode). Look up Baxandall diode and you will get an explanation of how it works.
However then you will need to reduce the gain of that stage using a C-B capacitor of around 100pf.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.