Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site fluke.UUCP
Path: utzoo!linus!decvax!microsoft!fluke!rzdz
From: rzdz@fluke.UUCP (Richard Chinn)
Newsgroups: net.audio
Subject: son of ampzilla bridge
Message-ID: <30@tpvax.fluke.UUCP>
Date: Tue, 27-Sep-83 12:26:35 EDT
Article-I.D.: tpvax.30
Posted: Tue Sep 27 12:26:35 1983
Date-Received: Thu, 29-Sep-83 01:27:04 EDT
Organization: John Fluke Mfg. Co., Everett, Wash
Lines: 98

I'm posting this reply to the net since it may be of general interest.

Bridging an amplifier is no big deal. I don't like to do it for a couple of
reasons. 

    1. If the load impedance stays the same (say 8 ohms) then each amp
       channel must drive 4 ohms. If the output stage has enough SOA (safe
       operating area), then fine. Otherwise, expect smoke.

    2. There's twice the opportunity for problems if one of your amp's goes
       south.

    3. Neither side of the speaker line can touch ground, since both are hot.

I must admit that most of my prejudice comes from having been in the touring
PA bizzness, where reliability is paramount. 

On the positive side, you get twice the power into twice the impedance (nice
for driving 16 ohm speakers) and 4 times the power into the same impedance
(nice for subwoofers).

If you can live with the above mentioned shortcomings, heres what you need
to do.

I don't know if the ampzillas phase invert. If they do and if GAS sold a
*passive* bridging box then they *must* phase invert (unless the box had a
transformer in it), then build the following:

    A voltage divider whose division ratio is equal to the amplifier's
    voltage gain.  That is, if the amplifier has a gain of 28 then the
    divider must be a 28:1 divider.  

    The output impedance of this divider must be *at least* 10 times less
    than the input impedance of the amplifier. (If the amplifier has a 22 k
    input impedance, then the shunt arm of the divider should be 2.2k ohms
    maxixmum.) 

    The division ratio shoud be accurate (use 1% resistors). 

    Connect the input of the divider to the output of the first amp channel,
    connect the output of the divider to the input of the second amp channel. 

    Connect the load across the *hot* terminals of each output. Don't use
    the ground terminals for anything. Neither terminal of the load can come
    in contact with anything else. 

Example:

    The amplifier has a voltage gain of 33. The input impedance is 22k.
    I assume you know what a voltage divider looks like.

    The divider must have a division ratio of 33:1, thus the resistor values
    must be in the ratio 32:1 (value ratio = 32:1, then division ratio =
    33:1)

    If the shunt leg must be 2.2k, then the series leg is:

	  2.2k * 32 = 70.4k. The nearest 1% value is: 69.8k.

    this results in a division ratio of: 31.73, probably close enough.


If your amplifier doesn't phase invert, then there is no way to do what you 
want passively (except to use a center tapped transformer). I don't 
consider transformers truly *passive*, especially for a high performance 
stereo system. 

You can do what you want actively. It will take one op-amp. I suggest 1/2 
of a ne5532 (can't use the 5534 as it is not unity gain stable) as an 
inverting amplifier. this time it can have unity gain. Make the gain 
setting resistors the same value, say 22k ohms, 1%. The input of this goes 
to the output of the preamp, and one amp channel. The output of the 
inverter goes to the remaining amp input. 

One reply to the net recommended using two op-amps, with one running as a
voltage follower. While this will work, the voltage follower has been
condemmed as a source of distortion, both slew induced and common mode,
since it is running at 100% feedback. If your preamp will drive the combined
loads of the inverter and power amp input in parallel, this is far superior.
If your preamp won't, then I suggest running the non-inverting stage with
some gain, preceeding it with a voltage divider to exactly offset that gain.

In either case, the speaker (load) goes between the two hot terminals of 
the amplifier. The ground terminals are *not* used.

I hope this helps. If you want more details, reply by mail or usmail or 
phone:

	 Rick Chinn
	 John Fluke Manufacturing
    	 MS-232E
	 POB C9090
	 Everett WA  98206
	 (206) 356 5232

or fluke!tpvax!rzdz