I was wondering if anyone had done anything in the area of detecting shorts on power amp speaker outputs while the amp has its output muted by SSR.
What I mean is using some way to sense possible shorts on the speaker side while the speaker is disconnected from the amp by the SSR.
It's not difficult to do with electromechanical relays, using the other side of the switch that would normally not be used, to pickup the lead from the speaker while it's not physically linked to the amp's output, and when the relay is energized to unmute, the speaker is connected while the short sensing circuit no longer is...
The thing is, it's a bit different with SSRs, with no physical contacts being used, and there would be no way to disconnect the short sensing circuitry from the speaker once the amp is unmuted.
So far I haven't found any way to do this with SSRs, but perhaps some have given some thoughts and may have come up with something usable?
What I mean is using some way to sense possible shorts on the speaker side while the speaker is disconnected from the amp by the SSR.
It's not difficult to do with electromechanical relays, using the other side of the switch that would normally not be used, to pickup the lead from the speaker while it's not physically linked to the amp's output, and when the relay is energized to unmute, the speaker is connected while the short sensing circuit no longer is...
The thing is, it's a bit different with SSRs, with no physical contacts being used, and there would be no way to disconnect the short sensing circuitry from the speaker once the amp is unmuted.
So far I haven't found any way to do this with SSRs, but perhaps some have given some thoughts and may have come up with something usable?
I'll have to do some simulations.
The key is that whatever circuitry is used to sense this, it must be always connected, and so it must not disturb the amp when unmuted and not be disturbed by the strong signal when the amp is putting out.
And I'm aiming at the SSR being on the ground side, so that's a tiny bit different..
The key is that whatever circuitry is used to sense this, it must be always connected, and so it must not disturb the amp when unmuted and not be disturbed by the strong signal when the amp is putting out.
And I'm aiming at the SSR being on the ground side, so that's a tiny bit different..
Just took a quick peek at the p300 sch, and they have electromechanical relays on the outputs, so, different..
Do you mean that the detection circuit does not have access to the muting condition info?
The example I proposed easily meets that condition
That changes very little, and in fact it could help meet your first requirement more easily: when muted, the amplifier output will act as the low Z return.
If the amplifier continues to be active and delivers a signal, a differential acquisition might be required to avoid overwhelming the synchronous detector, but that's not even certain.
To make the stimulus signal unnoticeable in case it is required, it can be passed through a second order LP filter: a 15Hz, -100dB signal is not going to be audible
No, that can easily be arranged. Actually what I have in mind is using a PIC to control it all, so it would be actuating the muting and it obviously would know it's doing it.
I'm fuzzy on that point. Can you elaborate please?
Well, if it can be designed in a way that it can be somewhat deactivated, although it wouldn't be "unhooked" physically, it could be made into something like a tri-state connection, kind of like some ICs are in computers, with their "legs lifted" when in their tri-state condition, so it's kind of like it's not there.
Actually in a way, the speaker's leg taken off line by the SSR is kind of "lifted" that way in something like a tri-state condition.
Obviously it would have to be totally inaudible, especially in case some speaker is highly sensitive and could put out some noticeable sound from a tiny signal.
The sensing signal would only be sent out when the SSR has lifted the leg. So nothing would be issued when the SSR is in the unmuted state.
I was wondering something: I guess using caps provides some galvanic isolation, but perhaps optocouplers could also be used, for each leg, and that would be total galvanic isolation then.
Also one thing to keep in mind, is to make the system immune to whatever signal might be coming back from the speaker from ambient noises.
And I'm not sure about how this sync detect should be implemented..
Also one thing to keep in mind, is to make the system immune to whatever signal might be coming back from the speaker from ambient noises.
And I'm not sure about how this sync detect should be implemented..
A small (but high voltage capable) signal relay might be appropriate for connecting such sense circuitry to the speaker only at startup?
As for sensing a small DC current of a few mA might be appropriate, and measuring the voltage this creates with a comparator or two for open / short circuit checking. Say 5mA injected, 10mV and 100mV thresholds - enough to detect if the load's between 2 and 20 ohms.
In fact if there's a protection resistor and clamp diodes between speaker and comparator, and the current is injected from the +ve rail with a high voltage CCS, you might not need isolation during normal operation, but you might want to switch off the CCS to keep DC offset low.
As for sensing a small DC current of a few mA might be appropriate, and measuring the voltage this creates with a comparator or two for open / short circuit checking. Say 5mA injected, 10mV and 100mV thresholds - enough to detect if the load's between 2 and 20 ohms.
In fact if there's a protection resistor and clamp diodes between speaker and comparator, and the current is injected from the +ve rail with a high voltage CCS, you might not need isolation during normal operation, but you might want to switch off the CCS to keep DC offset low.
I’d just inject a say 20 mV DC signal into the muted speaker through a 10 Ohm resistor, amplifier the resultant voltage divider circuit and make sure it was at the correct voltage +- say 10 or 20%.
No need to over complicate this guys!
In fact, if you’re using a micro, it’s a bit of code and 1 resistor.
No need to over complicate this guys!
In fact, if you’re using a micro, it’s a bit of code and 1 resistor.
The speaker will have one end tied to the output, which is a ~zero impedance point.
For the synchronous acquisition, it is all that matters: even if that point is polluted by a strong signal (the music program), it will only see its own stimulus, and if that stimulus isn't present, it simply means that the load is shorted.
To acquire a few mV against tens of volts of perturbation will require a hardware implementation of the sync-det: the ADC of the PIC would not be able to cope with such a dynamic range, but the detector can be as simple as a CD4053.
In fact, most of the functions required to implement the whole circuit could be carried out by a single CD4053
With the switch in the GND side, the signal will be totally inaudible during normal operation. Only when muted, a small 15Hz signal will be heard.
You do not even have to turn off the test signal during normal operation
Avoiding any electromechanical relays is a must, and so everything stays hooked up even when unmuted, so it must not be interfering while unmuted and be usable to sense shorts while muted.
I'm definitely all in for simple.
Aiming for the SSR on the ground side, and the short sensing should not cause any audible effects in speakers, including sensitive ones like tweeters, and of course make sure not to do anything that could be damaging to sensitive speakers.
A woofer can handle quite a bit, but sensitive tweeters could be more at risk if using overly large signals.
The sensing signal can easily be generated by the PIC, and there is no need for much electronics afterwards to get back the sense signal to be fed into the PIC.
Generating a signal that won't harm speakers is important. And all that stuff has to remain hooked up once the SSR goes unmuted.
I'm trying out a few sims, but nothing yet satisfactory..
I'm definitely all in for simple.
Aiming for the SSR on the ground side, and the short sensing should not cause any audible effects in speakers, including sensitive ones like tweeters, and of course make sure not to do anything that could be damaging to sensitive speakers.
A woofer can handle quite a bit, but sensitive tweeters could be more at risk if using overly large signals.
The sensing signal can easily be generated by the PIC, and there is no need for much electronics afterwards to get back the sense signal to be fed into the PIC.
Generating a signal that won't harm speakers is important. And all that stuff has to remain hooked up once the SSR goes unmuted.
I'm trying out a few sims, but nothing yet satisfactory..
Avoiding any electromechanical relays is a must, and so everything stays hooked up even when unmuted, so it must not be interfering while unmuted and be usable to sense shorts while muted.
I'm definitely all in for simple.
Aiming for the SSR on the ground side, and the short sensing should not cause any audible effects in speakers, including sensitive ones like tweeters, and of course make sure not to do anything that could be damaging to sensitive speakers.
A woofer can handle quite a bit, but sensitive tweeters could be more at risk if using overly large signals.
The sensing signal can easily be generated by the PIC, and there is no need for much electronics afterwards to get back the sense signal to be fed into the PIC.
Generating a signal that won't harm speakers is important. And all that stuff has to remain hooked up once the SSR goes unmuted.
I'm trying out a few sims, but nothing yet satisfactory..
I'm definitely all in for simple.
Aiming for the SSR on the ground side, and the short sensing should not cause any audible effects in speakers, including sensitive ones like tweeters, and of course make sure not to do anything that could be damaging to sensitive speakers.
A woofer can handle quite a bit, but sensitive tweeters could be more at risk if using overly large signals.
The sensing signal can easily be generated by the PIC, and there is no need for much electronics afterwards to get back the sense signal to be fed into the PIC.
Generating a signal that won't harm speakers is important. And all that stuff has to remain hooked up once the SSR goes unmuted.
I'm trying out a few sims, but nothing yet satisfactory..
I've been thinking about using a photovoltaic opto coupler, such as the PVI5050 or perhaps the APV1122, to generate the sensing signal from the PIC's output.
Should be simpler and it isolates galvanically.
It could be driven pretty much directly from a PIC's output, possibly without any additional parts..
The thing is, that PVI5050 is basically obsolete now, at least in its through hole version. So I suppose something like the APV1122 would do the job, unless there is an other I'm not aware of that could be more suitable.
On the sensing side of it, a regular LED/Transistor type of optocoupler could be used, and that would make it all completely isolated. Some protection from large signals from the amp would be needed to prevent damage, since it all stays hooked up when unmuted, but that shouldn't take much.
Should be simpler and it isolates galvanically.
It could be driven pretty much directly from a PIC's output, possibly without any additional parts..
The thing is, that PVI5050 is basically obsolete now, at least in its through hole version. So I suppose something like the APV1122 would do the job, unless there is an other I'm not aware of that could be more suitable.
On the sensing side of it, a regular LED/Transistor type of optocoupler could be used, and that would make it all completely isolated. Some protection from large signals from the amp would be needed to prevent damage, since it all stays hooked up when unmuted, but that shouldn't take much.
One thing that I think would be worthwhile to integrate with the short detector is the DC detector.
Since the sensing is done while muted, it seems sensible to also checks for DC and prevent unmuting in case either a short or DC were detected.
Now this brings up the point that in case of amp failure and DC is present on the amp's output, it's not just the amp's output signal that needs to be taken into account, but also that potential presence of DC, and so the sensor needs to be immune to this, while being able to sense for it when needed and remain inert when unmuted.
Since the sensing is done while muted, it seems sensible to also checks for DC and prevent unmuting in case either a short or DC were detected.
Now this brings up the point that in case of amp failure and DC is present on the amp's output, it's not just the amp's output signal that needs to be taken into account, but also that potential presence of DC, and so the sensor needs to be immune to this, while being able to sense for it when needed and remain inert when unmuted.
If the SSR is on the ground side and you accidentally short the output to the chassis you can have a problem.
If the output plug is right on the PCB, how can such a short to chassis happen?
The kind of shorts that are much more likely to encounter are the ones beyond the output plug, on the wires somewhere, and probably not even at the end on the speakers, just between the amp's plug and the speaker's plug.
I'd be curious to find out how a short to chassis can happen physically when there is nothing between the amp's actual output on the PCB and the the output plug. Not including whatever user probing happening during initial setup.
And by the way, so far, unless I missed it, I have yet to see any kind of practical implementation of a short detection with SSRs instead of relays...
I think a simple way to do the sensing is using a comparator, with 2 dividers, one having the speaker as part of itself, the other a fixed resistor that fixes the limit below which anything is to be considered a short. No need for fancy stuff there, and it's all mostly done and triggered by the micro-controller, and it can be galvanically isolated with optocouplers.
The kind of shorts that are much more likely to encounter are the ones beyond the output plug, on the wires somewhere, and probably not even at the end on the speakers, just between the amp's plug and the speaker's plug.
I'd be curious to find out how a short to chassis can happen physically when there is nothing between the amp's actual output on the PCB and the the output plug. Not including whatever user probing happening during initial setup.
And by the way, so far, unless I missed it, I have yet to see any kind of practical implementation of a short detection with SSRs instead of relays...
I think a simple way to do the sensing is using a comparator, with 2 dividers, one having the speaker as part of itself, the other a fixed resistor that fixes the limit below which anything is to be considered a short. No need for fancy stuff there, and it's all mostly done and triggered by the micro-controller, and it can be galvanically isolated with optocouplers.
If your amp is on and you touch the speaker positive accidentally to the chassis - which should be grounded - you have a short to 0V. Believe me because it’s the techniques I used in the nx-Amp and I had to warn builders to be very careful on that point.
You could put a ground lifter in and then detect when the diodes are conducting and use that to shut the PSU down, but it’s a bit messy.
You could put a ground lifter in and then detect when the diodes are conducting and use that to shut the PSU down, but it’s a bit messy.
Yes, ground lifters are a good thing to use, and not much to add really.
But what I'm really curious about is how effectively a contact from that positive can actually be possible, if the output plug is right there on the PCB. What could be done to put that positive output signal to chassis, by what path?
Having no wiring at all between the positive and the plug, I don't see what could come in to make any contact.
I use the Neutrik SpeakOn plugs, PCB mounted. So I'm wondering how this situation could be even possible.
But what I'm really curious about is how effectively a contact from that positive can actually be possible, if the output plug is right there on the PCB. What could be done to put that positive output signal to chassis, by what path?
Having no wiring at all between the positive and the plug, I don't see what could come in to make any contact.
I use the Neutrik SpeakOn plugs, PCB mounted. So I'm wondering how this situation could be even possible.