90+90 volt rails

[sakis]

come across this schematic that looks a fairly simple and nice construction ....

the truth is that sudenlly i have the complete list of the original semiconductors available ( yeS !!!!) but what really troubles me is that from expirence it seems that rail voltage looks way higher than needed

also looking at this circuit notice that there is no VI limiters or any oher form of limiting and also heatsinks are notreally something huge ....

so my calculation say that this doesnt look safe ....

your opinion please ????

[h_a]

90 V rails with only 2 pairs of output transistors?

15 V rails for the frontend? Difference shamelessly burned with a resistor?

Does not charm me

Have fun, Hannes

[wg_ski]

Just because whoever built the amp the last time used only two sets of output transistors doesn't mean you have to. Seems silly to have the VAS doubled up and skimp on outputs. I'd beef the drivers up to 15032/15033, but you could parallel the Jap types here, too.

[sakis]

that is correct and thank you very much

but this is a factory made design called armstrong .... here is a link
http://www.audiomisc.co.uk/Armstrong/700/700page1.html

[wg_ski]

Originally posted by h_a
15 V rails for the frontend? Difference shamelessly burned with a resistor?

Let's see you find (affordable) JFETs that can handle the 90 volt rails. With two differential stages, you can get away with running the first one off lower voltage - its no different than an op-amp front end.


A word of caution - there isn't any frequency compensation in this amp schematic. It will need the usual collector-base caps on both sides of the VAS and because of the extra front end stage, a lead compensation cap as well. An experienced builder will know how to calculate this (or at least tune it in experimentally).

[wg_ski]

Originally posted by sakis
that is correct and thank you very much

but this is a factory made design called armstrong .... here is a link
http://www.audiomisc.co.uk/Armstrong/700/700page1.html

The schematic given may be incomplete (ie, no compensation values, no SOA limiters, etc). It also looks like 'consumer grade' electronics to me. Two output pairs will do 8 ohms just fine if that's all you ever drive. Sony receiver/amps can't handle anything but 8 ohm loads, either.

[AndrewT]

the Armstrong claims ONLY 200W for a BJT output stage from +-90Vdc.
The literature confirms 85Vpk output on music transients and also states that no current limiting nor protection is installed.

I suspect that a small transformer and low values of smoothing capacitance are being used to protect the output stage during abuse conditions.

Normally we'd expect 200W from +-60Vdc to +-65Vdc.

[wg_ski]

Originally posted by AndrewT
the Armstrong claims ONLY 200W for a BJT output stage from +-90Vdc.

Gotta wonder if it will make point five past lightspeed (or the Kessel Run in sub-12 parsecs).

[h_a]

Let's see you find (affordable) JFETs that can handle the 90 volt rails.

Naah!

Much more elegant possibilities available than to roast a resistor. Like use a cascode and employ a nice series regulator for the frontend. Or use the resistor at the very least as part of a CRC filter.

Not that it's strictly needed, but hey, using a power resistor feels just primitive IMHO. Does not speak in favour of the design.

Missing compensation does not either

If I have to rework that circuit before building it, I can build something entirely different instead as well

Have fun, Hannes

[Michael Chua]

90V rails, 200W rating = Dynamic Headroom.

Would be wise to double the outputs.