Amplifier ouput relays

[d3imlay]

I'm looking for comments on output relays. My current plan is to use a DC sense circuit to open a relay. I'm curious if anyone has ever seen these circuits actually work. A few thoughts are:

1. The UPC1237 IC seems to be popular. I couldn't find it at the NEC site and am wondering if it's discontinued. Are there other IC's I should look into? I've used the Motorola MC3423 in non audio crowbar circuits and it works quite well but would need some additional driver circuitry.

2. The relay. Many of the posts I've seen make referenece to welded contacts. That's pretty simple. Nobody uses a relay intended for DC applications. To get the proper DC rating two or more contacts need to be wire in series with snubbers across them. Does anybody know of a 10 amp relay that has contacts rated for greater than 24VDC?

3. Time dealy. In order to ignore the audio signal, a time delay of about 100-200ms (10-5Hz) is in order. The delay time for the relay to drop out is another 50 ms also. It would seem to me that 150-250 ms time delay is to long to protect anything.

Any comments are much appreciated.

D3

[richie00boy]

1. Never heard of that chip and I try to avoid specialist chips as they always go obsolete. I just use good old op-amps.

2. A relay should be easy to find. I use FR-1 pinout types, which I believe are quite standard. The ones I use have 16A 250V AC rating. In the UK the RS order code is 198-6983.

3. I can't see why you need a time delay. That's time to fry your speakers. You feed a comparator signal fed from a 1Hz or so low-pass filter. Your music content should be well above this frequency so any signal present can be assumed DC and trip your comparator.

[Centauri]

The TA7317P has been around for quite a while for this application, and still seems to be readily available. Unless there is a suitable low voltage supply already available within the equipment, I tend to favour a discrete transistor circuit.

I have done a few tests watching how relay contacts arc when breaking DC current, and a 1uF capacitor of suitable voltage rating across the contacts quenches the arc quite well. I usually use relays with 30A contacts, and with the cap higher voltages should not be a problem.

richie00boy, a 1Hz low-pass filter would suggest a delay of 1 second. While delays are inevitable, there will also be a finite time required for speaker voice coils to heat up to the point where damage occurs - and hopefully the protection can beat that. I once played with the idea of using normally closed contacts and hitting the relay with a large spike to speed things up, but it would seem that most relays will actually drop out faster than they activate.

Cheers
Graeme

[mhennessy]

Originally posted by richie00boy
3. I can't see why you need a time delay. That's time to fry your speakers. You feed a comparator signal fed from a 1Hz or so low-pass filter. Your music content should be well above this frequency so any signal present can be assumed DC and trip your comparator.

I think he's thinking that a 10Hz filter would introduce a 100ms delay.

To the OP: I see your logic, but ask youself how long it takes a 1uF capacitor to charge to around 0.6-1 volt when fed from 50 volts via a 15K resistor?

50 volts represents a fault condition, and 1uF and 15K is a 10Hz filter. Answer: not long! Not sure I can calculate it with the calculator that comes with Windows, but I reckon it's probably less than a millisecond (I'll check that later when I have the calculator handy)

[d3imlay]

More importantly, how high is the cap going to charge when you have a 50hz signal at 10 volts. I'm guessing your sense circuit will trigger on that also.

Have a look at the following link. mc3423 app note

Figure 7 is the application. My concern is what will the speaker tollerate during the fault condition. Burn out is not the issue. However I once drove a JBL4311 woofer out of the gap with a dropped tonearm and a Phase Linear 700B. I want to protect against the jumped gap situation.

As far as relay speed goes, a diode across the coil of a relay will hold it in longer. In this case I think the diode needs to go around the transistor.

[Centauri]

Originally posted by d3imlay
Burn out is not the issue. However I once drove a JBL4311 woofer out of the gap with a dropped tonearm and a Phase Linear 700B. I want to protect against the jumped gap situation.

A speaker disconnect protection circuit is only there to protect against burnout in the event of amp failure - no more. Most people also utilise the relay for turn on/off dethump. This type of circuit CANNOT protect against your scenario - for that you want a STEEP (min 24dB/oct) subsonic filter, and peak limiter, operating BEFORE the power amp.

Cheers
Graeme

[cunningham]

An inductive reactive load will draw current out of phase with the voltage. This means that when Vout is 0V, current will still flow depending on the phase angle. Since all the voltage is droped on the transistor, too much current may exceed the SOA of the device. In this case, I don't see how a relay would protect the outputs, just the speaker if one 'melts'. Any time delay in this case would make in the speaker...unless maybe if it is way over rated for the amp. Depending on the relay, contact resistance could be a problem if it ever arcs.

[Leolabs]

To me,protection is for PA(or professional) amplifier,not for home project.Why?Unless u r a first time DIYer,otherwise u will know when and how things go wrong.As long as the circuitry itself is stable,components are reliable,and the load is not something" ",everything will be fine.U may not agree with me,but think about it.Accident???HA!!!U must haven't pray yet!

[anatech]

Well, protection should be included in home gear. Why? Because electronic parts have a failure rate regardless of how good the design may be. This failure may cause high DC offsets to occur. Electronic repairs are normally less expensive than speaker repairs. Not to mention that you need the electronic repair regardless.

We haven't even touched on the stupid things that happen to gear in the home. More than one person has left the owner's manual on top of an amp. One sheet of paper is all it takes to block air flow. Boom (the cust wants warranty too!)
More power by hooking amps together - Boom - you guessed it, warranty? Let's bridge a normal home amp (not designed to bridge) - Boom - this guy is screaming for warranty - yeah right.

My feeling is that unless you build it and design it, for your home only, there should be some kind of effective protection for the load. I really don't care what happens to the amp 'cause they don't support combustion.

-Chris

[mhennessy]

Originally posted by d3imlay
More importantly, how high is the cap going to charge when you have a 50hz signal at 10 volts. I'm guessing your sense circuit will trigger on that also.

Yes indeed - my point was that the reaction time isn't just a simple function of 1/f - in other words building a 1Hz filter doesn't result in a 1 second delay. Like all things, you have to engineer it - experiment to determine the filter TC that gives the best reaction time while not triggering falsely on typical signals.

Purely out of interest, where are you going to find a 10V, 10Hz tone from? That's only 12dB down on full output (assuming 100W amp, and peak voltages) - I've not come across any music CD's that have such high levels of infra-sonics. Is the amp for PA use? If so, I seriously wonder if a high-pass filter is required, because amplifying these sorts of signals is just a waste of energy.

Have a look at the following link. mc3423 app note

Figure 7 is the application.

This only appears to deal with one polarity of DC - useless for audio... Forget this IC

The normal solution is to use a pair of small-signal transistors to detect bipolar fault conditions - have a look here http://sound.westhost.com/project33.htm for a reasonable example...

My concern is what will the speaker tollerate during the fault condition. Burn out is not the issue. However I once drove a JBL4311 woofer out of the gap with a dropped tonearm and a Phase Linear 700B. I want to protect against the jumped gap situation.

Ouch - there's a lot to be said for that IEC ammendment

To be fair, your example isn't really a DC-offset fault - just a very large signal. On the occasions when I use records, I generally turn the volume down 20dB (IE press "mute") when handling the pickup arm..

As far as relay speed goes, a diode across the coil of a relay will hold it in longer. In this case I think the diode needs to go around the transistor.

This is true - Douglas Self did an investigation into this issue - can't remember offhand all the findings, but they might have involved using Zener diodes to limit the back-emf while speeding up the turn-off. I can look up the articles at the weekend if you remind me (PM).

IIRC, the additional delays incurred were relatively small. Also IIRC the same article had good examples of complete circuits, which would form a good starting point. They were all discrete however - personally, I'd prefer this to using special chips that go obsolete, as richie00boy said. Also, I'll echo what anatech has said - DC offset protection should be mandatory on any amplifier. I've also lost woofers in the past