How to suppress turn-off transient?

[scorpio]

Hi all,
I'm using a single ended tube amp (wil EL 84) for the mid-high section of my system, turning this on is nice and smooth, but I get an awful transient burst when I turn the thing off. I've tried cleaning the switch contacts, it does not help. Now, I'm turning it on an doff by removing the mains plug, which is a little better but still gives a fairly loud transient. Is there anything that could be done to help moderate or eliminate this? Is there a possibility that it is due to the switch only and a replacement would be better than a cleaning? The amp is a second hand purchase, so I have no clue if it was doing the same when new.
Thanks

[yggdrasil]

My guess is that this is a property of this amp/construction.

If this is the case, you will need some kind of mechanism to either disconnect the speakers or short the output while turning off the amp.

Which amp is it?

[scorpio]

Johnny, it's an Aurexx Crystal, I got this cheaply on e-pay, it sounds surprisingly good, but I'll probably move it to a less sensitive application if I can't sort out this issue. It's a very simple singele ended affair, there's a wiring scheme available on the net, if that is something that could be useful for someone knowledgeable to answer the question.

I have the simple "power" amp verison, no volume control.

Cheers

[yggdrasil]

You could allways ask here: http://www.diyaudio.com/forums/forum...p?s=&forumid=5

I'm not familiar with the tubes, so... If you don't get any useful help I have a simple suggestion for a shorting (or almost shorting) mechanism.

[Ian Mackenzie]

You really need a fast acting turn off device to protect the loudspeakers.

It also depends on the cause. It probably has to do with the time constant of the power supply discharging too quickly.

[scorpio]

You could allways ask here: http://www.diyaudio.com/forums/forum...p?s=&forumid=5

I'm not familiar with the tubes, so... If you don't get any useful help I have a simple suggestion for a shorting (or almost shorting) mechanism.

Hi Johnny, is your suggestion less potentially hazardous than Grumpy's? I certainly do not fit his description of competence, so would not be trying it... Thanks

[yggdrasil]

Ian is right about installing an automatic fast acting device.

An easy way to test this would be to connect a power switch between the output terminals of the amplifier parallell to the speakers. With the switch you can short the output of the amplifier before turning off.

[grumpy]

Would it not be better to have a multi-pole/throw switch that dumps the amp into a
resistive load than to short it out? Rest of the amp circuit will still see this transient,
regardless... seems like a bad thing that should be resolved at the source if possible.

BTW, my earlier post really could be dangerous -if- the cap shorted... you'd not be
able to turn off the amp. Better place for the suppression device/circuit would be
-after- the power switch to ground... still gives a place for the transient to go,
but takes power offline as long as the switch works. If the cap shorts there, you'd
just pop a fuse or breaker.

I should have thought twice before posting.

caveats regarding required knowledge, experience, and being potentially lethal still apply.

[boputnam]

ETA Systems offers rackmounted devices with sequenced Edison sockets. You simply plug in the devices in the order you need them to power-up/-off and you are all good. You can even dial-in the wait-time desired (1, 5, 10 or 30 secs).

[grumpy]

Hmm.... could, if the outputs are s/s switched at zero volts or... if it uses
mechanical relays, at least some of the filtering stays connected on the
outlet side. My understanding of the problem was simply an amp turn
off transient issue, as opposed to a pre/amp sequence "pop", which a
device like this certainly would fix. :dont-know

[boputnam]

My understanding of the problem was simply an amp turn off transient issue...

Doh! You're right - this won't do nuthin' for that...

Carry on...

[scorpio]

Hmm.... could, if the outputs are s/s switched at zero volts or... if it uses
mechanical relays, at least some of the filtering stays connected on the
outlet side. My understanding of the problem was simply an amp turn
off transient issue, as opposed to a pre/amp sequence "pop", which a
device like this certainly would fix. :dont-know

You're right Grumpy, my current problem is only when the amp is turned off and I get this loud nasty transient. Being a tube amp, it has a very smooth and noise free start up.

Thanks for all your suggestions guys, i don't feel competent enough to sort this out myself without the risk of creating a shock hazard, so i'll call my technician. If he can't sort this out cheaply, the amp will be retired until I decide to set up this single driver rear horn system I've been thinking of for years. It would use a 'cheap' Fostex full ranger for which I would not be concerned by this issue.

cheers

Cheers

[yggdrasil]

Would it not be better to have a multi-pole/throw switch that dumps the amp into a
resistive load than to short it out? Rest of the amp circuit will still see this transient,
regardless... seems like a bad thing that should be resolved at the source if possible.

It would be better for the amp, or at least less stressful for the output transformer. On the other hand it is difficult to know if parallelling a low resistive load (e.g. 1R) will protect the speakers enough.

If amplifier protection is crucial the switch can be rewired so that one or both speaker wire leads are switched. Close the switch to connect the speakers to the amp. Open the switch when turning off the amp.

Looking at the schematics I guess that the transient comes from discharging C5, C6 and C7.

[Hoerninger]

Simply disconnecting the speaker is not enough, as the tranny may not run with an open circuit. Alternativly a resistor (about 16 ohms) should be switched to the output.

speaker off + resistor on --> mains off

(The whole is best done with volume down.)
___________
Peter

[grumpy]

yes, that was the intent of my "band-aid" suggestion, actually switch the output from the
speaker to a resistive 'dummy load' (not put in parallel). My concern is it would just
mask the problem.