Protection Circuit - Crowbar - Zener vs SBS

[Byrd]

Hi

I have bee looking at crowbar circuits and came across this thread. In it is decribed the dc protection circuit used on the quad 405. It uses a Silicon Bilateral Switch to activate a triac which then shorts the amp out. It also contain a filter network to ensure the thing doesn't trigger all the time.

My question here is simple. What advantage in this instance does the SBS hold over a zener? The only thing that I can see is that it may be a bit more accurate with the "trigger voltage", but surely this voltage does not realy matter.

The reason I ask is that these SBS's are horrendously expensive but at the same time I am convinced that quad must have made this decision for a reason.

[mzzj]

Originally posted by Byrd
Hi

I have bee looking at crowbar circuits and came across this thread. In it is decribed the dc protection circuit used on the quad 405. It uses a Silicon Bilateral Switch to activate a triac which then shorts the amp out. It also contain a filter network to ensure the thing doesn't trigger all the time.

My question here is simple. What advantage in this instance does the SBS hold over a zener? The only thing that I can see is that it may be a bit more accurate with the "trigger voltage", but surely this voltage does not realy matter.

The reason I ask is that these SBS's are horrendously expensive but at the same time I am convinced that quad must have made this decision for a reason.

I think if you replace SBS with pair of zeners in quad style circuit the Triac will never fire. If your triac needs 100mA on gate to fire its never going to happen in quad style circuit. Of course you can scale down some of the resistors to get enough current but you end up with some losses on those resistors and you need way more bigger cap for same filter response. And even then you have the problem of slowly rising triac gate current. I am not sure if it is an issue with triacs but at least BIG SCR's switch faster and more reliably with sharp trigger pulse.

[djk]

The NTE6403 and the 2N4992 are bringing close to $8 these days. A couple of years back I was paying $0.50 for them. Fortunately, they are just four $0.10 transistors and a couple of zeners. Carver uses this discrete version in most of their magnetic field amplifiers. You can also use the reverse breakdown of a BE junction of an inexpensive transistor instead of the zener, they usually have better match of the breakover voltage that way too.

http://www.onsemi.com/pub/Collateral/MBS4991-D.PDF

[mzzj]

Seems that around 30V diacs are easy to find and cost only cents, but anything less than 30volts is ceased 10 years ago.

[Byrd]

Exatly my experience - What about using a "protection IC" to trigger these triacs. It has relay drive capability - 80mA

Prehaps in conjunction with a BTB08-600C triac - these only require 25·00 25·00 25·00 50·00 mA in the respective trigger quadrants.

[KISS]

When I built the Leach amp in the early 80's, I added a protection circuit which does a few things:

1) Thump supression
2) Logarithmic ramping of the audio.
3) Catastrophic amp protection.
4) Slow turn-on because of 40,000 uf of capacitance.


These were prompted by one power supply design where an electrolytic cap on the amp board shorted. It blew one of the rail fuses, but would not shut down. Bad things happen when one rail goes down. Smoke and sizzle. I thought that this wasn't a good idea.

I was working on the amp with the protection circuit in place and accidently switched the output transistors. The component(s) that were supposed to fry did (1 resistor and 2 fuses) and no other damage.

You get the consumer rated DC detect circuits to outsmart the consumers, I just wanted to protect the amp and the speakers, providing I use the proper fuses.

The AGX fuse on the output will protect the speakers. The fuses on the rails will protect each amp. If I loose a single rail, I lose a metal oxide resistor. If the amp won't power up while the AC is in current limit, and the speakers are off and the input is disconnected, then I loose the resistor.

The speaker AGX fuse usually blows first with the volume turned up and that's it.

I wanted to add thermal protection and a clipping indicator, but that never happened. I did design a clipping circuit for another application which would light a bi-color LED red or green depending on the polarity. It also extended the time on to 1 second. References were +-10V so clip would occur if +-10 was exceeded.

I made a mirror image of the circuit board becasue the ones published in Audio Magazine wern't labeled nicely and could not for the life of me find the problem for a long time. Ended up switching the npn and the pnp's and moving a couple of PC traces and had a working AMP.

Initial amp had single bipolar supply rated at 20A with a constant voltage transformer. Bass was superb, but the transfomer itself was too noisy. Had a custom torroidal xformer made with 4for 35V, 3A secondaries which in my opinion are too small. Put it in a 2U custom made rack case with a screen top.

[Workhorse]

How about using back to Back SCR's in Crowbar, yet another inexpensive solution....

[Byrd]

It is odd that the leach reacted in this way. In my experience if one of the rails go there is no dc at output.

Can you give us some more details on your protection circuit. Diagram?

[d3imlay]

SCRs offer an advantage over triacs in that they are available in much higher currents. None of the crowbar circuits I've seen limit the current to a safe value of the DIDT rating of the triac/SCR. It seems as though the assumption is that if it triggers, the amp has already failed and it is reasonable to sacrifice the triac. I think that a .10 or .20 ohm series resistor would probably limit the current enough to protect the triac.

The MC3423 chip is specifically designed for crowbar applications and the data sheet gives info for proper SCR ratings. It could probably be adapted for audio signals. Attached is the data sheet.

[AndrewT]

Hi Kiss,
do you have a schematic you could post?