Having just burnt out a tweeter (thankfully a 'sacrificial' speaker) I was wondering if anyone has come up with a combined DC and HF protection crowbar-type circuit? It would be nice to have a small 'passive' box with 4mm sockets that you could stick across a precious speaker and know that it was safe whatever went wrong, without affecting the signal quality.
Beyond that, the ultimate 'amplifier testing station' might include a more flexible (CPU-controlled?) monitor circuit designed to detect misbehaviour (DC and/or oscillation) and to shut down MOSFETs in series with the power lines to the amp. Having experienced an amplifier with RF oscillation (~3MHz) it would be nice to detect this too, and indicate it with an LED perhaps.
Beyond that, the ultimate 'amplifier testing station' might include a more flexible (CPU-controlled?) monitor circuit designed to detect misbehaviour (DC and/or oscillation) and to shut down MOSFETs in series with the power lines to the amp. Having experienced an amplifier with RF oscillation (~3MHz) it would be nice to detect this too, and indicate it with an LED perhaps.
Hi Andrew.
Are we talking about Apexaudio? I couldn't find exactly the circuit you mention, but I'm beginning to think that what I want isn't possible anyway: the TRIAC will reset itself on every cycle and from what I can see in simulating with these things, they don't do a very good job of shorting an HF signal, even with a continuous gate trigger.
Also, from my recent unwanted experience, the tweeter can be killed with a relatively low PSU current so the PSU fuse(s) may not trip anyway.
Thinking about it, it looks to me as though the ultimate protection scheme must be that of turning off the PSU after the smoothing caps using a MOSFET in each rail - I think you have commented extensively on one thread which suggests this. Then the active fault detector can be as sophisticated as we like, and able to respond to DC offset and HF oscillation (- which I now worry about just as much as DC!). There was one thread I found (but have lost again!) which also limits PSU current to a set level and provides a soft start function.
I really would like to be able to connect my favourite speakers to a lashed-up amplifier and not worry about it!
Are we talking about Apexaudio? I couldn't find exactly the circuit you mention, but I'm beginning to think that what I want isn't possible anyway: the TRIAC will reset itself on every cycle and from what I can see in simulating with these things, they don't do a very good job of shorting an HF signal, even with a continuous gate trigger.
Also, from my recent unwanted experience, the tweeter can be killed with a relatively low PSU current so the PSU fuse(s) may not trip anyway.
Thinking about it, it looks to me as though the ultimate protection scheme must be that of turning off the PSU after the smoothing caps using a MOSFET in each rail - I think you have commented extensively on one thread which suggests this. Then the active fault detector can be as sophisticated as we like, and able to respond to DC offset and HF oscillation (- which I now worry about just as much as DC!). There was one thread I found (but have lost again!) which also limits PSU current to a set level and provides a soft start function.
I really would like to be able to connect my favourite speakers to a lashed-up amplifier and not worry about it!
I've only seen one loudspeaker protection circuit which appears to detect ultrasonics; it is the K2000L. I don't know who makes it (the kit) or where it came from... not Velleman but the protection could be added. Note IC1C,D are comparators for 1/2 wave rectified high-pass filtered audio with RV1 setting the trip point; they are wired-OR'd to the DC detect IC1A,B. That's my take on it.
I stumbled onto the schematic one day and thought it would be good idea, but I haven't tried it out.
I stumbled onto the schematic one day and thought it would be good idea, but I haven't tried it out.
Attachments
I fried a "sacrificial" tweeter last week. Not from oscillation , just a OEM standard underrated offering (5w tweeter in a 200w 3-way).
The week before I oscillated by having my amp insufficiently degenerated. In this instance the cheap tweeter survived the 2mhz signal because the zoble took the "heat". This is a rare occurrence (except on a test amp) , and would most likely never happen on a mature design. Seeing how the zoble did it's job , one could just rectify the voltage drop across the zoble resistor to trigger the protection. The only other HF mechanism the could take out a tweeter is massive clipping and the high order "garbage" that produces.Having a clean amp and a 50W tweeter / 50W mid should handle 300w total program power according to this :
An externally hosted image should be here but it was not working when we last tested it.
As long as the amp is not clipping or is VERY poorly designed , the tweeter should be no where near it's limit with even 200 Watt's of clean power.
As far as DC , this comes down to whether the rail fuses or speaker voice coil are the first to go. I have done both , Big amp- small speaker ,voice coil is toast. Small amp- large speaker , I have blown the amp fuse after hearing a short power supply "hum". I agree , neither scenerio is optimal and a DC detection circuit is a good thing. The " omni" protection circuit (by Dr. Bora) is very good. Some don't like relays between their amps and speakers ... but if they have $200 bass drivers , that is their issue..
"Omni" is in my" mongrel" amp linked below...
OS
By then the oscillation would of fried something , especially with a large amp. My zobel R took over 2 seconds to "smell"... phewweee! A fast diode would rectify the drop across the resistor instantly , we want mS response.
os
I'm not doubting that the sensing method wouldn't be faster, but I would surmise that your zobel R isn't optimised for thermal detection.
At oscillation bands where the zobel becomes effective, typical network design will mean that the zobel will be taking the brunt of the power in the output circuit. A lower than typical R (not too low though - otherwise you'll kill the ops) coupled with a small power dissipation rating (1W for example) would heat up like insanity. Wrap it nicely with a small thermo with a lowish rating (e.g 60 degrees) and I'd say that you'd cut off well before the VC gets a chance to burn.
Of course an in band oscillation would still burn the tweeter - but then even electronically sensing the zobel won't help you, you would need to use an integrating power detector.
The situation differs from DC where you both have instantaneous motional stress and a low impedance path through the tweeter, HF oscillation situations are subject to output L and voice coil L limitations.
Anyway, I'm still yet to encounter an oscillation related burn - hopefully never
There is a part of the Crimson circuit, around the VAS that looks like an HF detector. It triggers a two transistor thyristor and that in turn starves the LTP of tail current. This is effective at shutting down the amp if the HF arrives at the input. Because it is a thyristor, it latches. I have not tested it, but I suspect it will latch whether the HF is applied to the +IN or to the speaker terminals (-IN via the NFB).
The output Thiele Network acts as an HF load on the output of the amp irrespective of what is hung on the speaker terminals. It also acts as a filter attenuating HF sent to the speaker.
Further it can be arranged to also attenuate RF coming in from the speaker cables.
DC blocking in the speaker goes a bit further and can be used as part of the treble filter.
Adding DC detect and isolation would prevent that cause.
Adding a turn on delay to let the blocking cap to charge gets over noisy start up. That same relay with instant drop out on loss of power also prevents noise at shut down.
We have 4 circuits, two of which are very common and all 4 are cheap to implement to protect your tweeter from abuse.
The biggest saviour is a stable amplifier that can drive adequate voltage to the tweeter and never overload, i.e. never clip the treble signal.
The output Thiele Network acts as an HF load on the output of the amp irrespective of what is hung on the speaker terminals. It also acts as a filter attenuating HF sent to the speaker.
Further it can be arranged to also attenuate RF coming in from the speaker cables.
DC blocking in the speaker goes a bit further and can be used as part of the treble filter.
Adding DC detect and isolation would prevent that cause.
Adding a turn on delay to let the blocking cap to charge gets over noisy start up. That same relay with instant drop out on loss of power also prevents noise at shut down.
We have 4 circuits, two of which are very common and all 4 are cheap to implement to protect your tweeter from abuse.
The biggest saviour is a stable amplifier that can drive adequate voltage to the tweeter and never overload, i.e. never clip the treble signal.
Many thanks for the ideas. I do have an aversion to relays in the signal and ideally I would have liked a passive 'black box' that could be hung on the speaker terminals, but I can also see the advantage of things that act further back in the chain. An "integrating power detector" seems to be the key requirement. Time to get designing and simulating I think! Testing it could be a bit scary, though.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.