Gary, I can see your concern. The first question, though, is what was going on with the radio? Did it just put out a continuous DC level (and if so, WHY!!), or was it just a turn-on transient? Even if the radio puts out a transient pulse at turn-on, it should be brief and not be so long as to allow you time to notice what was happening and then react. There's something wrong with that radio.
The LM3886 is inside the feedback loop of the overall composite amp. Pulling the mute pin on the LM3886 opens up that loop and thus the controlling op-amp is now in an open-loop situation. I expect that pulling the mute will kill the speaker feed without any adverse consequent, but I'm not sure what will happen when you un-mute -- you may cause a pop or other undesirable output.
The LM3886 is inside the feedback loop of the overall composite amp. Pulling the mute pin on the LM3886 opens up that loop and thus the controlling op-amp is now in an open-loop situation. I expect that pulling the mute will kill the speaker feed without any adverse consequent, but I'm not sure what will happen when you un-mute -- you may cause a pop or other undesirable output.
Last edited:
The only reason I used the radio was cos it was handy given the location of the other kit. When I first set it all up the radio had already been running for awhile before I switched on the amp. The power cut revealed the power on surge. I tested it afterwards and it was outputting 1.2vdc which tailed off after a few seconds. Nasty. It has gone back to being a standalone unit and will never be part of a test set, however basic.
As for the mute option if I go that route it will bench tested and scoped before it goes anywhere near speakers.
Wow. Bad source! Bad! Sorry to hear about your misfortune there.
The DC servo in the MOD86/286 will correct for some DC on the input. I forget the exact number but a few hundred mV come to mind. DC on the input beyond that will cause DC on the output. That's the drawback of a DC coupled amp.
The LM3886 is a pretty rugged chip. While it can fail (any component can) it is rather unlikely to be the first to fail. I'm not overly worried about speaker protection. I don't use any on my amps. There is enough demand for a protection circuit that I'll entertain that at some point. Hopefully over the summer.
As far as preventing DC out from a preamp driving the MOD86/286, I suggest using a muting circuit. The easiest way to accomplish this is by a relay on the preamp output. Just make sure to pick one that's designed for the switching of signals rather than power.
Tom
The DC servo in the MOD86/286 will correct for some DC on the input. I forget the exact number but a few hundred mV come to mind. DC on the input beyond that will cause DC on the output. That's the drawback of a DC coupled amp.
The LM3886 is a pretty rugged chip. While it can fail (any component can) it is rather unlikely to be the first to fail. I'm not overly worried about speaker protection. I don't use any on my amps. There is enough demand for a protection circuit that I'll entertain that at some point. Hopefully over the summer.
As far as preventing DC out from a preamp driving the MOD86/286, I suggest using a muting circuit. The easiest way to accomplish this is by a relay on the preamp output. Just make sure to pick one that's designed for the switching of signals rather than power.
Tom
Yep. I'd expect a turn-on plop from the radio. DC on the output is a no-no.
Messing with the LM3886 mute pin in the MOD86/286 is not recommended. I should try that in the lab sometime to see what would happen. I'd expect the DC servo to rail as it can't control the output DC voltage when the LM3886 is muted. The driving opamp will likely saturate one way or the other as well. The loop is fairly fast so I would expect it to recover relatively quickly on un-mute. The only sticking point would be the DC servo. With a settling time of seconds, it would probably run your speaker cones through quite the exercise regime before settling completely.
A mute on the input of the MOD86/286 is better implemented by a relay which opens the inputs during mute and short the two signal wires (input connector pins 2, 3) together.
Tom
Hi Tom,
Whats wrong with a protection relay for the speaker? You can also mute the input at the same time to save the relay from signal when it first turns on. Just delay the signal release by 100 msec or so. That will save speakers and provide a more polished power state behavior.
What happens if the relay contacts get burned? Smile, it just took a bullet for your speakers. Replace the relay and be happy. When replacement time rolls around, then get another common relay of the type that you used in the first place. You did choose a more common industrial relay, didn't you?
I don't think there is any excuse not to use a relay protection setup, and no, they do not cause distortion unless they need to be replaced.
-Chris
Whats wrong with a protection relay for the speaker? You can also mute the input at the same time to save the relay from signal when it first turns on. Just delay the signal release by 100 msec or so. That will save speakers and provide a more polished power state behavior.
What happens if the relay contacts get burned? Smile, it just took a bullet for your speakers. Replace the relay and be happy. When replacement time rolls around, then get another common relay of the type that you used in the first place. You did choose a more common industrial relay, didn't you?
I don't think there is any excuse not to use a relay protection setup, and no, they do not cause distortion unless they need to be replaced.
-Chris
Nothing. But like any other precision circuit, it requires proper design to work well. Distortion in power relays is well documented by Self, Cordell, et al. At the precision levels involved in a regular amp, it's "just" a matter of choosing the right relay. So you buy a bunch of them and measure them until you find a low THD one. Whether this approach is adequate for the precision levels of the Modulus-86 is a question I intend to answer at a later date. If the relay causes more THD than the amp, one would have to apply post-relay feedback to maintain the high performance of the Modulus-86.
In addition to not causing additional THD, the protection circuit will also need to actually protect the speaker if something goes wrong. It seems given, but there are circuits out there that don't provide that much protection or will protect in some scenarios while neglecting others.
I came across a mention in some of Hypex's documentation that stated NOT to use output relays as the contacts could weld shut during a failure condition. It may be a result of the ~30 A output current their amps can provide and is likely not an issue for protecting the Modulus-86. But it did make me wonder if there's more to relay protection than initially meets the eye.
Your notion about replacing the relay "when replacement time rolls around" is nice ... if you can find a relay that fits the footprint at replacement time. The relay I used in the power supply for the DG300B went obsolete and there was no replacement that would fit the footprint. I had to throw the last supply boards out and make new ones with the new footprint. It was a pretty generic Omron relay. The replacement relay looked identical and performed identically but had the pins in different locations, thus, necessitating a board spin.
The measurements performed by Douglas Self and Bob Cordell beg to differ.
Tom
Last edited:
A DC coupled Power Amplifier should have a series of protections for the speaker.
The DC servo is there to correct output offset drifts as temperatures vary inside the Power Amplifier.
A DC servo is NOT there to correct source offsets.
The list of protections could include:
Output offset detection.
Output isolation.
Input offset detection.
Input muting.
DC servo nearing full correction capability.
DC servo reaching full correction capability.
Quiet alarm (flashing LED) for minor offset.
Loud alarm for major offset.
Shut down and latch for major fault.
Some of the above can be omitted from an AC coupled Power Amplifier. But I still consider offset detection and isolation as well as shut down and latch for major fault to be mandatory.
The DC servo is there to correct output offset drifts as temperatures vary inside the Power Amplifier.
A DC servo is NOT there to correct source offsets.
The list of protections could include:
Output offset detection.
Output isolation.
Input offset detection.
Input muting.
DC servo nearing full correction capability.
DC servo reaching full correction capability.
Quiet alarm (flashing LED) for minor offset.
Loud alarm for major offset.
Shut down and latch for major fault.
Some of the above can be omitted from an AC coupled Power Amplifier. But I still consider offset detection and isolation as well as shut down and latch for major fault to be mandatory.
I was aware that a direct coupled amp could put something nasty on the output given a silly connection issue, but initially I thought I would be careful when wiring and turning on etc. This issue just proved me wrong. Also a simple loose connection could cause hum etc. So I'm now considering some kind of protection.
I'm not technical enough to know what needs to be monitored and how it is best done. My thought about using the mute pin just proved that. However if there were protection along the lines Andrew is suggesting then it is a short hop to make them available externally.
I'm thinking, remote ON/OFF & remote mute for control by the pre-amp. The power LED could be changed to an RGB one and used to indicated the different conditions.
I'm not technical enough to know what needs to be monitored and how it is best done. My thought about using the mute pin just proved that. However if there were protection along the lines Andrew is suggesting then it is a short hop to make them available externally.
I'm thinking, remote ON/OFF & remote mute for control by the pre-amp. The power LED could be changed to an RGB one and used to indicated the different conditions.
I was considering using the SoftStart and Speaker Protection products from here Zesilova?e in the power amp and the ON-OFF Module from here ON-OFF modul when I build Tom's preamp.
(You'll need to know Czech or use your browser's translation feature to read those pages)
Colin
Hi Tom,
Actually, Doug Self referred to a specific relay that channeled current through the frame. For real purposes, any distortion generated by relays I would use for a speaker protection relay is nowhere even close to the distortion delivered by real power amplifiers or the speakers themselves. Don't get caught up in absolute values, view those figures in context with the system performance.
I'm sorry your choice in relay wasn't the market's choice. Go for a common telecom relay, or a common industrial relay. I've been replacing the same type of realy for over 30 years now.
They make high current relays that have contacts suitable for lower current levels. I've only ever seen welded contacts when an amplifier failed catastrophically. The events leading up to that failure caused the load (speakers) to fail before the amplifier did.
Don't pick out exceptions to the rule, we can argue those all day long if you want to. However, there are two other ways you can protect a speaker. One, short the output with a triac in a fault condition. Two, limit the stored energy in the power supply and control the AC input as Carver does. That method works well, but the filter capacitance would be lower than what would be considered acceptable for these simpler amplifier designs. Personally, I like the relay route, and my favorite amp (that measures the best also) uses speaker relays. I've had to replace that once so far, not due to any failure. Just age.
-Chris
Actually, Doug Self referred to a specific relay that channeled current through the frame. For real purposes, any distortion generated by relays I would use for a speaker protection relay is nowhere even close to the distortion delivered by real power amplifiers or the speakers themselves. Don't get caught up in absolute values, view those figures in context with the system performance.
I'm sorry your choice in relay wasn't the market's choice. Go for a common telecom relay, or a common industrial relay. I've been replacing the same type of realy for over 30 years now.
They make high current relays that have contacts suitable for lower current levels. I've only ever seen welded contacts when an amplifier failed catastrophically. The events leading up to that failure caused the load (speakers) to fail before the amplifier did.
Don't pick out exceptions to the rule, we can argue those all day long if you want to. However, there are two other ways you can protect a speaker. One, short the output with a triac in a fault condition. Two, limit the stored energy in the power supply and control the AC input as Carver does. That method works well, but the filter capacitance would be lower than what would be considered acceptable for these simpler amplifier designs. Personally, I like the relay route, and my favorite amp (that measures the best also) uses speaker relays. I've had to replace that once so far, not due to any failure. Just age.
-Chris
Chris,
I'm sorry my words were seen as fighting words. That wasn't my intent.
All I'm saying is:
- While many amps work just fine without a protection circuit, I agree that some may want a protection circuit.
- Other experts in the field have reported that some relays cause distortion.
- Other experts have expressed concern about welding of contacts.
- Any solution implemented should protect the speaker.
- Any solution implemented should NOT degrade performance.
- I have previously committed to looking at this and to design a circuit.
- I would like to do the design well as my business hinges on good performance.
- There is only one of me and only 24 hours in my day.
I am not here to argue or to start fights. I apologize that I have not solved all the world's engineering challenges by now. There. Are we good?
I'll very likely include a protection circuit in the 150-200 W design that I'm working on. It would seem reasonable to spin the protection circuit off as a separate board. I have mentioned this several times in this thread so I'm a bit confused as to how you've concluded that I'm somehow incapable or unwilling to design a protection circuit.
I thought Gary wanted a solution that didn't involve a speaker relay, so I responded accordingly. It is possible that I misread his request. I am only human.
I'm doing the best I can with the resources I have available.
Tom
I'm sorry my words were seen as fighting words. That wasn't my intent.
All I'm saying is:
- While many amps work just fine without a protection circuit, I agree that some may want a protection circuit.
- Other experts in the field have reported that some relays cause distortion.
- Other experts have expressed concern about welding of contacts.
- Any solution implemented should protect the speaker.
- Any solution implemented should NOT degrade performance.
- I have previously committed to looking at this and to design a circuit.
- I would like to do the design well as my business hinges on good performance.
- There is only one of me and only 24 hours in my day.
I am not here to argue or to start fights. I apologize that I have not solved all the world's engineering challenges by now. There. Are we good?
I'll very likely include a protection circuit in the 150-200 W design that I'm working on. It would seem reasonable to spin the protection circuit off as a separate board. I have mentioned this several times in this thread so I'm a bit confused as to how you've concluded that I'm somehow incapable or unwilling to design a protection circuit.
I thought Gary wanted a solution that didn't involve a speaker relay, so I responded accordingly. It is possible that I misread his request. I am only human.
I'm doing the best I can with the resources I have available.
Tom
I think Doug is measuring at the microscopic level on new perfect devices when he talks of distortion in relays.
The problems of relays causing (audible) distortion as the contacts age and tarnish seems to be a recurring theme on the forums. It crops up a lot, always the same symptoms.
To me, this was always one of those issues that 'always affected someone else', until it happened to me. When it did, I couldn't believe what I was hearing... which was gross distortion on a low power sine of around 2kHz. I honestly thought the speaker was caput it was that bad. Cone voice coil rubbing, it was just like that. On music it was fine, not noticeable as such. I turned the volume up briefly.
It was 'fixed' and then it clicked (with me, not the relay
) as to what was happening. I was a convert to solid state relays from that moment on and would now not entertain a mechanical one in any shape or form, whether it be speaker duty, line level or dare I say even mains switching.
Relays are a weak link. An antiquated anachronism that we can move beyond in this day and age.
The problems of relays causing (audible) distortion as the contacts age and tarnish seems to be a recurring theme on the forums. It crops up a lot, always the same symptoms.
To me, this was always one of those issues that 'always affected someone else', until it happened to me. When it did, I couldn't believe what I was hearing... which was gross distortion on a low power sine of around 2kHz. I honestly thought the speaker was caput it was that bad. Cone voice coil rubbing, it was just like that. On music it was fine, not noticeable as such. I turned the volume up briefly.
It was 'fixed' and then it clicked (with me, not the relay
) as to what was happening. I was a convert to solid state relays from that moment on and would now not entertain a mechanical one in any shape or form, whether it be speaker duty, line level or dare I say even mains switching.Relays are a weak link. An antiquated anachronism that we can move beyond in this day and age.
Yes and it seemed there was a simple fix in using the mute pin but I thought I'd check with the expert and got my answer.
A few years ago, someone in Europe was selling a relay designed for power amp outputs. It had two sets of contacts. The first was a normal beefy set of power contacts. The second were gold-plated contacts. Mechanically it would close the 'grunt' contact first and open it last. The good contacts would close second and open first. That way they were protected from arcing and pitting, yielding, according to the advertising, an 'ideal' and long-lived speaker relay.
One could do essentially the same thing with two properly-chosen relays and appropriate circuitry to drive them.
edit: Found the relay: it was (is?) made by Amplimo; don't know if it still is around. And the contacts were Tungsten (grunt) and Silver (good)
One could do essentially the same thing with two properly-chosen relays and appropriate circuitry to drive them.
edit: Found the relay: it was (is?) made by Amplimo; don't know if it still is around. And the contacts were Tungsten (grunt) and Silver (good)
Last edited:
How much is the common-mode voltage of the Legato?
As long as Vcm + Vsignal(peak) < 13 V, you will be fine.
If he has data to indicate that the relays cause no additional distortion even at the <-125 dBc level, I'm all ears. Talk is cheap. Data rule. Sorry if that rubs anyone the wrong way. It's how I roll.
Do you recall which solid-state relay you used for speaker switching duty?
They're a known trouble spot, yes. I recall conversations with the test engineers at National/TI about that, actually.
Tom
Interesting idea. The drive circuit wouldn't be too bad. It's basically your standard non-overlapping clock circuit with a much longer non-overlap than the 100s of ps I used to aim for.
Relays with silver contacts are pretty common. It was my impression that they were designed such that the contacts rubbed a bit to clean off the tarnish.
Looks like tungsten + gold according to the writing on the package: https://www.toroidal-transformer.com/shop/high-end-audio/loudspeaker-relay.html
Tom
Because the reservoir caps generally hold enough energy to destroy the speaker. One could, as Chris pointed out a few posts back, short the output with a crowbar and turn off the power supply. Let the crowbar handle the energy in the supply caps. For the LM3886/MOD86/286 that would probably work as the output is current limited. In a discrete design without current limit, you'll probably destroy the output stage devices with the crowbar.
Tom
Hmm, I don't recall any commercial amp I have used over the past 30 years had output protection. The only recent built was the Myref which was used mainly to solve a startup output surge problem. For the Mods, I think rather than think about performance, perhaps "idiot proof" is more important to some people as an add on. People that look for performance first are generally less likely to stick more stuff in, you can always hear a difference when you add stuff. The effects just add up. Since the distortion figures of the Mods depend heavily on layout, any add on will reduce performance, then you get into the argument "is this audible?", which further brings up the question, are there any audible differences between the different Mod models when used with speaker loading fine for all the models?
Sent from my iPhone using Tapatalk
Sent from my iPhone using Tapatalk
Last edited: