Anti-thump

[suzyj]

Hi guys,

My new amps make an impressive thump as the supply voltage collapses, putting about 1V across the speakers.

I was thinking of putting some sort of anti-thump circuit in with a relay in series with the speaker, that has a turn-on delay of a couple of seconds, and shuts off fairly quickly when the voltage on the main power transformer dies. That way the speakers are disconnected during turn-on and turn-off transients.

Anyway, here's a simple circuit I've been simulating:



Operation is fairly simple. It's designed to go on a 40V AC transformer (56V peak). The transistors base threshhold is raised to 7V or so by the zener. On power-up, D2, D3, and C3 create a +56V supply. When AC is applied, the voltage at the transistor base slowly charges. When it hits 7V, the transistor switches on, energising the relays, with ~48V. When AC is removed, the transistor base voltage is discharged via R2, and the relays are shut off within a hundred milliseconds or so.

So what do people think? Is this the way to do it? My main concerns are arcing across the relay contacts. Given that the poweramp power is switched by my preamp, I could program the preamp micro to ensure the volume is minimum before the power is switched.

Regards,

Suzy

[darkfenriz]

The way you've drawn it, to me it seems incomplete. AC is floating or what?

[johndiy]

looks good to me
btw nice circuit

regards

john

[clem_o]

- Not sure what voltage your relay coils are rated at, but you can investigate the idea of puting the coils in series instead and remove the 6.2V zener, which will probably get pretty warm - I think the 6.2V zener serves the purpose of increasing the bias voltage requirement to give a more predictable delay? Replace with forwad-biased 1N4001 and increase base capacitance.


- The turn-off time will be symmetrical to the turn-on time (on transistor side). For your circuit, the turn-off is made quick by the small capacitance on the 'hot' side of the relays. BUT - if you have a situation where the amp is turned off briefly then turned back on - it could potentially thump, as the relay control tranny would not have turned off yet. Create a quick discharge path for the bias capacitor...

You may want to consider using a MOSFET instead - Gate voltage requirement higher, and no gate current for turn-on - much easier to figure the RC requirement, and if you keep the R high and the C small, much easier to discharge quickly too...

Cheers

[AndrewT]

Hi,
after the transistor has switched on, what voltage is on the collector? (about 7V?).

What voltage is across the 12V relays?

I think you need very fast off, a few tens of mS (one or two mains cycles).

The speaker returns do not need to be relay switched.
The two poles of the relay could be paralleled and control just the "hot" (live) speaker feed.
Borbely uses the relay contacts in inverse parallel to reduce some problem (I wish I could remember what he called it).

[dutch diy]

Hi Andrew

a suggestion i would make is to change the way thw relais switches the speaker.

I've added a small diagram. Needs per channel one resistor R 15Ohm 10W will do.

My circuit always has a load to the amplifier (either 15Ohm or (during crossover of relais) 15Ohm in series with the speaker or the speaker).

btw Amplimo provides special hifi relais which contain two diffent kind of contacts. One massive for taking the main part of on/of current, and a silver one which closes slighty later, and opens slightly earlier to satisfy a very low resistance.

Good luck

[suzyj]

Originally posted by clem_o
- Not sure what voltage your relay coils are rated at, but you can investigate the idea of puting the coils in series instead and remove the 6.2V zener, which will probably get pretty warm - I think the 6.2V zener serves the purpose of increasing the bias voltage requirement to give a more predictable delay? Replace with forwad-biased 1N4001 and increase base capacitance.

I've got a pair of 48V coils in parallel. They each have a 4K coil resistance, so the current requirement is quite modest. I can also get the relays with 24V coils (1K resistance) and put them in series.

The zener does two things. It ensures that ~48V (56-6) is across the relays when the transistor is switched on, and provides a much higher (~7V) switch voltage.

I could replace it with a diode string, but the current is only 20mA, so dissipation in the zener should be fairly low (150mW).

Originally posted by clem_o
- The turn-off time will be symmetrical to the turn-on time (on transistor side).

No, it takes ~3s to turn on, and a hundred milliseconds or so to turn off. Without the transistor load, on turn on, the capacitor is charged to 13 volts (set by the R1/R2 voltage divider) through R1 (33K).

On turn-off, the capacitor discharges through R2 (10K).

Once the transistor base current is taken into account, the base voltage only rises to 7V, and on switch-off, the cap discharges (quite quickly) via R2 and the transistor base.

Originally posted by clem_o
You may want to consider using a MOSFET instead - Gate voltage requirement higher, and no gate current for turn-on - much easier to figure the RC requirement, and if you keep the R high and the C small, much easier to discharge quickly too...

I originally tried a MOSFET, but the turn-off was very slow (because there was no discharge via the gate. Also Vbe is fairly consistent, while Vgs(threshold) is more variable...

Cheers,

Suzy

[clem_o]

Hi suzyj,

Hmm, thanks for the corrections... I used a mosfet to drive the relays on my amp - the circuit worked pretty well, proper turn-on delay and a very quick turn-off. I'll have to dig it up as it's just on a perf board and it was something that I hacked together many years ago... I recall I used a diode to discharge the capacitor quickly...


Cheers

[Dick West]

What do you guys think of this circuit?

http://mitglied.lycos.de/Promitheus/...or_toroids.htm

[suzyj]

Originally posted by Dick West
What do you guys think of this circuit?

http://mitglied.lycos.de/Promitheus/...or_toroids.htm

Looks really cool. I particularly like the way they've used a big capacitor as half a voltage divider. If I were making one (I've designed my own (much more complex) inrush current limiter that allows a TTL signal to control the power amp) I'd use a 48V relay. Then I could use a significantly smaller capacitor.

Just one thing though - it's _not_ an anti-thump circuit. It'll stop your power amp blowing fuses when you turn it on (by reducing inrush current) but does nothing to mute the speakers while the rails ramp up and down.

Cheers,

Suzy