any theoretical limit on class B power?

[lt cdr data]

ie could you go on for ever bridging and paralleling output devices?

[richie00boy]

I suppose you could. Remember that to series amp outputs the amps themselves need to be floating.

[pinkmouse]

Assuming you had an output device that could handle infinite voltage, otherwise you are limited by your device specs and supply rails. Remember Ohms law!

[richie00boy]

No Al you can series amp outputs to get around the device voltage limits. It's not a great way of doing it though, and you need to ensure matched and balanced voltage sharing.

[pinkmouse]

Got any examples? I can't see it being that stable personally, but I'm always willing to learn.

[richie00boy]

It's not something I'd do in practice.

Basically you just series connect amp outputs, i.e. red terminal of bottom amp connects to black terminal of top amp. However, both amps must be fully floating with no grounding otherwise you short things out. This usually entails each amp being driven by it's own signal isolating transformer to provide galvanic isolation.

[djk]

Early SAE and Threshold had models with three outputs in series per rail. This made sense as some of the outputs back then had SOA curves starting to fold as low as 30V.

I also have a Bozak like that, but it makes no sense here as the outputs they used had good SOA up to 100V.

A good series connected output stage:

http://users.ece.gatech.edu/~mleach/...mp/circuit.pdf

The Heathkit AA1800 was similar to this and ran on ±100V

A Crown VZ5000 will do 4KW (over 500V peak-to-peak) into 8 ohms bridged, and used MJ15024/25 in earlier versions.

Much larger amps exist than the VZ5000 of course, but they may have as many as four rail voltages plus and four more minus.

The Crown is a bridge amp to start, and the two channels are bridged again for 4KW at 8 ohms. How much voltage do you really need?

[AndrewT]

Hi,
another way to bypass the inherent device voltage limits is to cascode the devices.

Leach invented his "double barrelled" amplifier to overcome the voltage problem when devices were much worse than are available now.

Eg. two pair of MJL21193/4 can manage good power from +-100Vdc supply rails (350W to 400W into 8ohm).

A double barrelled version could, I suspect, manage good power from +-200Vdc supply rails (1kW to 1.5kW into 8ohm from 3 or 4 quad sets).
The resistive ladder arrangement could easily be extended to a three level cascode.

Another alternative is the switching rails (is it ClassG?) when the high voltage rails only get connected when high output power is called for.

[AndrewT]

Hi,

The Crown is a bridge amp to start, and the two channels are bridged again for 4KW at 8 ohms.

One cannot bridge an amp that is already bridged.

Were they paralleled to enable lower impedance to be driven?

[Workhorse]

Originally posted by AndrewT
Hi,
Another alternative is the switching rails (is it ClassG?) when the high voltage rails only get connected when high output power is called for.

Linear Switching of rails is referred to be Class-G
Step PWM Switching of Rails is known as Class-H

Originally posted by AndrewT
Hi,
One cannot bridge an amp that is already bridged.

Were they paralleled to enable lower impedance to be driven?

No, already Bridged amps could be Bridged, Crown does this by adopting the Grounded Bridge Topology and implementing the "Quad Bridging Technique"
CROWN Macro-Tech 5000VZ , QSC PL9.0, Electrovoice/Dynacord P3000 are all Grounded Bridge amps and bridging could be done to already bridged amps....

see the diagram

http://www.diyaudio.com/forums/attac...amp=1150172273

K a n w a r