isolated high voltage amp

[unclejed613]

i have a subscription to EDN (electronic design news), and i get some of the tidbits from the magazine emailed to me before i actually get the paper one in the mail. they have audio circuits in their "design ideas" column often enough to make it worthwhile getting the magazine.

i just got this one in my email, and it looks interesting, even though it is a quasi comp.......

http://www.edn.com/article/CA6643367...36&rid=1250817


the wording about overcurrent protection isn't really clear, but i'm guessing he meant 125ma fuses on the rails. at 125ma output current and driven to 1kv, that's an 8kohm load, so as-is this circuit is "not yet ready for prime time", but it has possibilities....

[jackinnj]

Hmmm, I did quite a bit of work 4 years ago w analog opto-isolators --not the 6N36 device shown in the EDN article -- but Hewlett Packard (AVAGO) and Siemens devices -- very linear, but not particularly fast.

[sakis]

very primary circuit but it looks like the best pa amplifier ever....this is probably the lowest part count for such a big amp ....

so whats the idea here unclejed ????

regardless the protection do you have anything else that troubles you ???? cause constructing a toy like that is probably 1 hour of work for testing perpus ( and i am really willing to !!!!! )

join in !!! tell us more !!!

small detail : i think that forum policy is against online circuits but at the age of 43 and after a univercity degree i consider safety a must but on the other hand rules are rules ....so in case that moderators decide to stop this i would like to continue this in private ...i try to e mail you through the forum but it doesnt work ...so you may email me any time you like .... thank you and stay in touch !!!!

[wg_ski]

Originally posted by sakis
very primary circuit but it looks like the best pa amplifier ever....this is probably the lowest part count for such a big amp ....

Not necessarily a *big* amp - 8k ohm load is kind of high. But it could be scaled up. This does, however, look like an interesting way to drive a pair of hexfets/bipolars/etc from 15 volt op amp supplies without having to resort to using separate power for each channel.

[sakis]

Originally posted by wg_ski



Not necessarily a *big* amp - 8k ohm load is kind of high. But it could be scaled up. This does, however, look like an interesting way to drive a pair of hexfets/bipolars/etc from 15 volt op amp supplies without having to resort to using separate power for each channel.

that is correct but you know what i mean..... this could be adopted or scalled down to a PA application ....what do you think ????

[jcx]

you will quickly run into slew rate problems with large input C devices - the high speed optos used can't reliably control more than 10-20 mA - and that requires device selection for higher CTR

also rail efficency is too low at the lower Vsupply of 4-8 Ohm loudspeaker power amps

[unclejed613]

it would take 1000 of these (paralleled) as-is to drive an 8 ohm load, but i thought it was an interesting drive method.... an optical VAS. it looks a bit tricky to bias, since that's handled by the pots in series with the LEDs.

[wg_ski]

Originally posted by jcx
you will quickly run into slew rate problems with large input C devices - the high speed optos used can't reliably control more than 10-20 mA - and that requires device selection for higher CTR

You could add a nice little complementarty EF pair, biased at a few tens of mA in between the opto's output transistor and the gates of the compound FETs. That will drive an amp or more of peak gate current if you need it at high frequency.

[rtarbell]

Great find unclejed!

I have seen this kind of amp used in some high voltage industrial amplifiers. A conceptual question:

If the output voltage is going positive, then the top half of the circuit will be supplying the current, through the source of Q2 (a common drain-type configuration). If the output voltage is going negative, then the bottom half of the circuit will be supplying the current through the drain of Q3 (a common source type configuration). Won't this create some kind of "mismatch" to the load in terms of output impedance of the amplifier?

I guess the same question could be posed about any kind of amplifier that uses a quasi-complementary output stage.

[wg_ski]

Both sides behave as common-drain and have equal voltage gain. The signal for both FETs is injected at the gate-source junction as a voltage, and the output is a current in the drain circuit. The DC power supply changes "position", but that's zero ohms as far as signal is concerned.