Standard Design

[analag]

http://www.twin-x.com/groupdiy/albums/userpics/OP.JPG I'm having problem with Q6 shorting out and causing R to burn when the input signal goes over a certain level, but by removing Q6 the opamp's reliability goes up tenfold. What gives.

analag

 

[JohnRoberts]

What is the CE breakdown voltage? can it handle near 50V?

JR

 

[Guest]

[quote author="JohnRoberts"]What is the CE breakdown voltage? can it handle near 50V?
[/quote]

It has already that 47V when idle, if you mean Q6...

 

[Samuel Groner]

Not sure of the exact breakdown mechanism but perhaps a diode across Q6/Q8 (see 990.pdf) helps--at least it makes clipping cleaner.

Samuel

 

[JohnRoberts]

The anti-saturation diode will not only provide faster recovery from clipping but also prevent Q6 from trying pull current from Q8's emitter resistor which might exceed Q6's dissipation capability across full supply.

It might be worth looking at the rest of Deane's design, while the input stage inductors may be over kill, the current limiting on the output stage is useful, and input anti-zener clamps will help keep low noise input transistors low noise.

JR

 

[analag]

Q6 is BC556 and all diodes are 1N4148. The rest are BC550 and BC560

analag

 

[clintrubber]

First thought to this is to make Q6 & Q8 a 2BJT-Darlington, but haven't checked for consequences...

 

[Guest]

Latch in the diffcascade maybe?

 

[JohnRoberts]

Look up the data sheet and see if the breakdown voltage is adequate. If yes, then add the Jensen 990 anti-sat diode from collector of Q8 to base of Q6 and see if that clears the failure mode.

It's a pretty simple circuit and shouldn't need massive tweaking to get it working.

JR

 

[Dan Kennedy]

The high gain causes excess current in Q6 via the base emitter junction of Q8. I use this same basic circuit in the EQ-2NV.

I wound up putting a 330 ohm resistor in the collector lead of Q6 to limit the current.

Q6 starts heating up as soon as the signal clips, and quickly goes into a runaway condition.

 

[JohnRoberts]

[quote author="Dan Kennedy"]The high gain causes excess current in Q6 via the base emitter junction of Q8. I use this same basic circuit in the EQ-2NV.

I wound up putting a 330 ohm resistor in the collector lead of Q6 to limit the current.

Q6 starts heating up as soon as the signal clips, and quickly goes into a runaway condition.[/quote]

The anti-sat diode will prevent the excessive current draw but with the added benefit of faster recovery from positive clipping.

JR

 

[burdij]

Or you could just get rid of it.

 

[Guest]

Consensus: 3 diodes in series from base to + rail.

 

[JohnRoberts]

[quote author="Wavebourn"]Consensus: 3 diodes in series from base to + rail.[/quote]

Consensus?

3 diodes from +V to base of Q6 could interfere with positive drive capability depending on actual value of Q8 emitter R relative to 100 ohm in Q7 and actual diode drops.

A single diode clamp seems better targeted to protect without otherwise limiting output capability. I won't presume to declare it the only possible solution. Apparently it was used more in 990 to prevent over dissipation than clip recovery or they would have used similar anti-sat circuit on negative pre-driver (2 more diodes) also.

Dissipation in q6 could be cut in half by connecting collector to ground instead of - supply, also adding a collector resistor would limit dissipation but allowing Q6 to saturate may add to the clip recovery issue.

JR

 

[Guest]

Ok.

Now, consensus: 3 diodes from base to +, appropriate resistor in emitter of VAS for a current limiting, resistor in collector of Q6 connected to ground, heatsink for Q6. :grin:
Analag puts a stool under a socket in a ceiling, stands on a stool, puts a bulb into a socket, 4 of us take stool and go around to rotate the stool with him with a bulb in hands, 4 more go around to prevent him from dizziness. :grin: