Could someone please answer this Transistor question

[OMNIFEX]

If I can trouble the amplifier developers on what is considered an easy answer from your perspective.

When an amplifier offers 12 Transistors per channel, how would one determine the lowest impedance it can handle?

I've read using multiples in parallel enables a more stable amplifier when low impedance is the goal. However, quantity is not discussed.

The Transistors in question are MJ15024.



Cheers!

[AndrewT]

download one of the SOAR spreadsheets.
Bensen, Janneman etc.

It's not a stability issue, it's overheating of the output device junctions.

[OMNIFEX]

Cheers AndrewT!

[djk]

What's the rail voltage?

How good is the heatsinking?

Ultimately these all go into the SOA calculations.

12 MJ15024 per channel suggests a quasi-comp PA amplifier with 5 pair as outputs, one pair as drivers. With real good heatsinking and under ±80V rails it could probably run a 2R load. This is essentially how the old BGW750, Peavey CS800, etc., were.

The Phase Linear 700B was also as described, but running on ±100V it was a bit marginal at 4R with the MJ15024, but would be fine at 4R with the MJ21196 (if you could keep it cool).

[fotios]

You must don't forget as well the VI limitter of output included as defacto in such type amplifiers.
Under this option, you can connect enough speakers in output, down to 0,5Ù and when the V X I exceeds the SOA as said Andrew (the sum of all output devices predicted by the constructor of amplifier) then the output power is reduced accordingly.
In reality, you can drive enough speakers without problem but only from their low to mid power level.

Regs
Fotios

[djk]

"You must don't forget as well the VI limitter of output included as defacto in such type amplifers"

By all means, remeber that it is there, because when it kicks it it will sound like h3ll, and if using full-range speakers it will probably blow the tweeters.

SOA limiting done with VI limiters is not designed to be engaged in the normal use of the amplifier. It is there to protect the amplifier in the case of a short.

[fotios]

Originally posted by djk
SOA limiting done with VI limiters is not designed to be engaged in the normal use of the amplifier. It is there to protect the amplifier in the case of a short.

I think we discuss about this issue exactly. Your question in your first post, was:
"When an amplifier offers 12 transistors per channel, how would one determine the lowest impedance it can handle?"
The VI limiter has relation with this, i.e. to protect the output from heavy loading - very low impedance load - and not to protect tweeters or woofers.

Regs
Fotios

[djk]

A member called me on the carpet for using h3ll.

Thank you, I was just too lazy to describe how a VI limiter sounds driving a reactive load-line.

It can be best described as a farting and popping sound, very nasty. The leading edge of the rail-to-rail voltage spikes are passed through the crossover to the HF units, many of which cannot handle this without damage (and they blow up).

Crest employes a circuit they call IGM, on low impedance loads it instantaniously modulates the gain of the amplifier (IGM) to prevent excurions outside of the SOA limits. This sounds very anemic, and changes the tonal balance in multi-amped systems.

Crown uses ODEP. On the vz5000 and vz3600 it cuts the rail voltage in half, effectively cutting the available power by 6dB.

In practice, amplifiers rated at 2R may geneally drive three 8R drivers, but seldom can handle four (without the protection becoming active, and the audible consequences of same).

VI limiting is like the air-bag in your car, do you really want to see if it works?

[fotios]

Hi djk

By the chance given from your report for "Output Devices Emulation Protection" of Crown and "Instantaneous Gain Modulation" of Crest, i have to reffer that tose are types of error computers comparing input and output. The named VI limiters are other thing. These included between the drivers and output devices. Their operation is by some way continuous, as they track the current flow thru emiter resistors, and when the current exceeds some limit, then start to thieve current from the bases of drivers.
Instead, the error computers are of different operation, and can produce significant amount of distortion when activated. I have a lot of experience from Peavey amplifiers. In these included a simillar computer, the known DDT (Distortion Detection Technic) which i can confirm you that when activated, the signal bellow 120Hz becomes very distorted, particularly in its positive going edge .

Regs
Fotios

[djk]

"Their operation is by some way continuous, as they track the current flow thru emiter resistors, and when the current exceeds some limit, then start to thieve current from the bases of drivers. "

Those are the ones that sound horrible.

"Instead, the error computers are of different operation, and can produce significant amount of distortion when activated. "

Are you sure? The IGM in the Crest, the McIntosh power guard, the Crown PIP-Clip, these kind of circuits are virtually distortion free. The QSC and Carver optical limiters use exactly the same part as the McIntosh, and are also free of distortion.

The Peavey DDT is a very different circuit. I have never observed the behavior you described. I will say that it is not instantaneous, it takes quite a while after you drive it into clipping before it pulls the signal back, perhaps this is what you see/hear?