triac's used as crowbar??

triac's etc

Hi hugobross

I haven't seen it, but I have played with it a long time ago. Please note that triac crowbars are like switches that close, and can only be reset when you interrupt the current through it. Using it on an amplifier output would surely protect the speaker, but solidly short the amp output until you switch it off. I'm not sure that would be good for the amp.

They way I did it was to detect DC offset (or long-term large output signal, which comes out to the same) and then fire a couple of triacs across the supply lines to either break the fuses or interrupt a resettable breaker.
The problem then becomes to balance the sensitivity between not sensitive enough and fry your speaker, or too sensitive and having the nuisance to reset the breaker after a false alarm. Which is always the trade-off for protection circuitry, of course. In the end I did away with it.

Cheers, Jan Didden
 

Asen

Member
2001-01-14 1:19 am
Sofia
Take a took at the Quad 405 schematic. They use such a protection circuit. The parts are hard to be found though. The circuit shorts the output and the fuses blow. In this case you should use a fuse between the PSU and the amp. I'm not shure it will be a good solution for a class A design.

Regards

Asen
 
My old QRO amp had a DC-protection with 7 Hz 18 dB/octave active filter and a simple comparator. The opamps were cheap LM324.

The filter was designed so the amp could deliver 300 W @ 20Hz with no problem and detect 1.4 V DC(I think) and be fast! Advanced solution but cheap to implement. The protection circuits were feed from a separate power supply.

The triac solution sounds dangerous. Careful designing is needed so you really get protection and not disaster.
 
peranders said:
[snip]
The filter was designed so the amp could deliver 300 W @ 20Hz with no problem and detect 1.4 V DC(I think) and be fast! Advanced solution but cheap to implement. The protection circuits were feed from a separate power supply.
[snip]

Peranders,

The circuit may be fast to do whatever it is supposed to do, but it will still need time to decide whether it is looking at a full-size 20Hz signal or DC. That takes time, many 10's of millisecs. The actual recation time of the circuit then becomes largely irrelevant.

Jan Didden
 
janneman said:


Peranders,

The circuit may be fast to do whatever it is supposed to do, but it will still need time to decide whether it is looking at a full-size 20Hz signal or DC. That takes time, many 10's of millisecs. The actual recation time of the circuit then becomes largely irrelevant.

Jan Didden

DC-protection doesn't need to react in milliseconds. The step response for a 7 Hz, 3rd order filter is fast enough for avoiding disaster. In my case "DC" is from 0-1,5 Hz (not sure of the exact value). The reaction time has nothing to do with the (normal) audio signal since it is filtered out.
 
peranders said:


DC-protection doesn't need to react in milliseconds. The step response for a 7 Hz, 3rd order filter is fast enough for avoiding disaster. In my case "DC" is from 0-1,5 Hz (not sure of the exact value). The reaction time has nothing to do with the (normal) audio signal since it is filtered out.

Check the delay time for a signal going into your filter. You will see it takes some time for the signal to propagate through the filter. It takes time for the filter to decide whether a signal is LF or DC. You cannot escape that; the response time to DC is intimitely connected to the filter corner frequency. Suppose your amp starts a temp runoff, the DC offset at the output may rise to the rail in the course of a several seconds. The damage may be done before the filter "decides" there is DC at the output. That is what I mean by reaction time. Once the decision is made that there is DC, the filter output activates the protection circuitry, and this part of the event chain is usually much faster. Because of the long lead time to the decision, it is irrelavant whether this last part takes 10uS or 1mSec. You can speed it up: in this example, if you change the corner of your filter to say 10Hz, it will react much more quickly, because anything below 10Hz (roughly) is considered DC. But then again, a strong 10Hz audio could activate the protection. So, as always, there is a trade-off.

Cheers, Jan Didden
 
Voice coils can withstand high DC for quite a period without damage.
In fact Peter Daniel spoke of his Focals withstanding 40V DC for several minutes without complaining.
Millisecond response time is not required.

BTW, old Accuphase amplifiers had a simple short circuit load sensing circuit that would disable operation of the output relay in the case of an abnormal load.

Also in my experimenting I have jammed the relay armature in the operated position and connected or disconnected the coil current, and found it to cause an audible effect.
A cowbar circuit to crash the supplies (fuse or breaker protected) is a better approach in my opinion.

Eric.
 
Hey thanks!!

Asen, you're right: the Quad 405 uses such a circuit, you can see it below. Thanks, that's what I was looking for!

Indeed, when using this circuit you've got to be carefull otherwise the whole amp can burn out. I would use it in combination with fuses between the PSU and the pos. and neg. rail of the amp, AND with a short circuit protection of the output devices.

Eric, indeed very good speakers can handle high DC voltages; but I'm not going to take that risk when using somebody elses speakers connected to my amps ( ;) ).

I started this thread because I don't like relays in amps, so I hope I can use the triac solution in the future. Right now, I use the velleman K4700 kit to protect my speakers (but it also uses relays).


thanks,

HB.
 

Attachments

  • triac protection.jpg
    triac protection.jpg
    10.5 KB · Views: 1,726
You have two different parts in the problem:

1 DC-detection

2 Breaking device

Relays breaks in 5-20 ms, Thyristors, triacs in us.

The detection time is dependent of

* Max output power at the lowest frequency
* Acceptable DC-level
* Filter sharpness

"DC" is always "AC" in some respect but where do you draw the line?

"DC" is lower freqeuncy than 0.01Hz or so.
 
I have also thought about RF detection but in those days it were harder to design a good and rather simple (did you hear that?;) ) wideband amp with sharp HP-filter and accurate RF rectifying. It did never leave my desk and my mind. The RF-protection is more or less the same thing as the DC-protection but the real problem here is RF. I admit though that nowadays is much easier to make a good wideband amp, use a modern CFB opamp.
 
hugobross said:
I would use it in combination with fuses between the PSU and the pos. and neg. rail of the amp, <snip> HB.

I think so too. Short out the rails on the amplifier side of the fuses. Use a pair of SCR's here, not triacs because SCR's have a much higher surge rating. (Itsm) Make sure you drive the gate on good and quick and strong so the SCR can safely stand the rapid rise of anode current.

Probably a good idea to have bleed resistors across your main filter caps so they will safely discharge after the power is turned off because they now have no load.

GP.
 

djk

R.I.P
2001-02-04 4:23 am
USA
"Use a pair of SCR's here, not triacs because SCR's have a much higher surge rating. (Itsm) Make sure you drive the gate on good and quick and strong so the SCR can safely stand the rapid rise of anode current."
Why not use a triac? Motorola TO220 triacs can do an ITSM of 450A. The 350A spec is for 8.3mS, not the exponential decay of a capacitor. Peavey uses the MAC224A5 in their 1100W amp. The only time the triac blows is when some idiot has put tin foil behind the mains fuse. By that time every emitter resistor, every output, and the bridge rectifier are gone. Do you care about another $1.50 for the triac at that point? If you have the correct fuses it is not a problem.