Is your unobtainium power switch worth $5 and an hour or so of your time?

[EchoWars]

Class aB solid-state has such a large range of primary currents that I consider thermistors unsuitable. True Class A solid-state and tube amps are a different story.

I like relays as well, but the triacs (thyristors) are sooo much easier, and the dozen or so units that I've installed them into work like a champ.

 

[Watthour]

I like relays as well, but the triacs (thyristors) are sooo much easier, and the dozen or so units that I've installed them into work like a champ.

Most times the SS devices will probably work just fine. This poster has experienced different results, and I thought I'd share some of my experience with similar problems to attempt to provide one possible explanation. There is enough iron in most transformers to ride through the minor "hiccups" which may result from the switching devices, and the transformer on the unit in question may not have quite so much reserve capacity, or in other words, there was not as much overkill as we often see in this vintage of electronics.

 

[d3imlay]

Is the additional wiring routed in such a way that low level circuits could pick up noise from the power wires?

 

[c.coyle]

The distance from the AC line, the switch, and the transformer are all but inconsequential. Electricity (arguably) travels at about 186,000 miles per second (6.7x10^8 MPH). I'll let you calculate how long it takes to traverse that extra 18" of copper and see if the resultant frequency shift/time represents anything near the microwave range, let alone the RF range (Forget about the audio range, where you can apparently hear the noise).

Right. Speed of light. Got it. :yes: The point I wanted to make was that on this unit, unlike before, the triac was in very close physical proximity to the power transformer, and I thought that might have something to do with the problem.

 

[c.coyle]

Is the additional wiring routed in such a way that low level circuits could pick up noise from the power wires?

There is no additional wiring, and no re-routing of existing wiring. I just snipped the hot lead, tied the ends to MT1 and MT2, and jumped the resistor from G to MT2.

 

[c.coyle]

Most times the SS devices will probably work just fine. This poster has experienced different results, and I thought I'd share some of my experience with similar problems to attempt to provide one possible explanation. There is enough iron in most transformers to ride through the minor "hiccups" which may result from the switching devices, and the transformer on the unit in question may not have quite so much reserve capacity, or in other words, there was not as much overkill as we often see in this vintage of electronics.

And I appreciate your take on this. I am here to learn. I get the feeling you work for a power company or something similar. Of course, this is a big, massive, beautiful, 1969 Sansui power transformer, and I'd like to think it can take whatever the triac might throw at it.

If the triac is stressing the transformer, what symptoms should I be looking for?

 

[K7sparky]

c.coyle
I am curious if you used the Alternistor TRIACS that are rated for inductive loads?

You noted the noise in the speakers is still there with them turned off. Sounds like arc type (radiated) noise picked up in the speaker leads that are acting like antennas. Guessing you are using a plain vanilla TRIAC that it is ringing like crazy - ringing hasn't ever happened to me, but one of the posters noted he had had the problem BIG time years ago. If you have a scope, (isolate it) and have a look across the TRIAC and transformer primary.

I monitored the voltage across the TRIAC with a Sony-Tek and Tek 475A among other scopes. When switching on I saw no noise and vary low voltage offset. Look at #17 for traces taken with a digital camera I posted and #38 for some really GREAT traces that d3imlay posted. As d3imlay did, I also used a Fluke 87 to capture 1msec peaks when I was testing before doing the initial post.

I started out with interposing relays to protect my amps with marginal power switches some years back. Relays open / close anywhere on the sin wave just like manual switches. I used MOV and snubber across contacts, MOV across primary for arc suppression.

TRIACS switch off at zero current (I am not picking nits) - Just the way they work. With this setup the turn on voltage is 2 silicon junctions or about 1.2V so still almost no current.

I still have a few amps with interposing relays I modified before developing the TRIAC mod. I am replacing the relays only if I need to get under the hood again.

The TRIAC gate current is lower than an Ice Cube relay coil and NO inductance. Even a badly damaged power switch has a chance to be resorted enough to work with a TRIAC.

 

[K7sparky]

I didn't address the transformer stressing.

You are not regulating output. You are simply switching the power on just after each zero crossing. Input wave form is still almost perfect.

If you have ringing you will be getting more transformer heating than normal. If you have a ringing problem, I would be contacting Little Fuse. The Alternistor is designed for this type application.

In general transformers run close to full load temperature at no load. The hysteresis loss is generally much greater than the I square R loss. Special wound / iron transformers like the spiral core toroids are a case of lowering hysteresis losses.

Back in the 60s iron was cheap. Extra iron = cooler transformer. Insulation max temperature has also gone WAY up. Modern transformers run HOT at no load - they are designed on the edge (minimum iron) as the cheapest wins the purchase in production run quantities.

Years back with 3 phase SCR drives there were heating issues with both the isolation transformers and the DC motors the drives were controlling. The Modified Bridge was the worst case. Modified bridge uses 3 diodes and 3 SCRs. Back then SCRs were EXPENSIVE so using 3 diodes saved some $. It was necessary to go a frame size or more larger because of the heating caused by the hacked up waveform caused at turn on high on the sin wave - regulating at low speeds.

Short take: as long as the TRIAC is switching on at each zero crossing there is no stressing.

 

[c.coyle]

c.coyle
I am curious if you used the Alternistor TRIACS that are rated for inductive loads?

Yes. Littelfuse Q4040K7TP. My previous installs, which are still problem free, were Q6040J7TP, but the slight difference shouldn't matter.

You noted the noise in the speakers is still there with them turned off. Sounds like arc type (radiated) noise picked up in the speaker leads that are acting like antennas. Guessing you are using a plain vanilla TRIAC that it is ringing like crazy - ringing hasn't ever happened to me, but one of the posters noted he had had the problem BIG time years ago. If you have a scope, (isolate it) and have a look across the TRIAC and transformer primary. . . .

I'm going to do that later this week when I'm off. Arcing picked up by the speaker wires makes sense. This triac is at the back of the unit, therefore closer to the speaker terminals and speaker wires than the other units, which had the triac right at the switch.

Plus, I removed the .033uf arc suppression cap across the switch (it was really "across the switch" at the far end of two 10" long leads that end up at the back of the unit), but I wouldn't think that would be the problem.

Thanks for responding. EW thought it might be a bad triac. I have some spares, but I want to try figuring this out before throwing parts at the problem. I think I'll re-install the .033uf cap. If no change, I'll temporarily put the triac up front at the switch.

 

[d3imlay]

...Plus, I removed the .033uf arc suppression cap across the switch (it was really "across the switch" at the far end of two 10" long leads that end up at the back of the unit), but I wouldn't think that would be the problem.

Actually the cap on the switch will cause a problem. I found this out when I first experimented with the circuit. The cap allows a small current to flow into the gate. I was getting random misfires due to this with the switch off. For clarity please provide the schematic you're using.

The triac will turn on at the voltage zero crossing if there is a load current. however the current waveform will not be sinusoidal. It may show high peak currents near ninety degrees that could be ringing. An oscilloscope would be nice to have here.

 

[c.coyle]

. . . For clarity please provide the schematic you're using. . . .

See attached. Red arrow is the cap I removed. Blue arrow is where I broke the hot side and inserted the alternistor triac.

 

[c.coyle]

Just a follow up to keep this thread from hanging. The buzzing does indeed appear to be radiating to the speakers. Moving the unit and the speakers farther apart reduces and then eliminates it. All inputs are dead quiet. Once I get done, I'll close it up and see if that eliminates it.

 

[K7sparky]

See attached. Red arrow is the cap I removed. Blue arrow is where I broke the hot side and inserted the alternistor triac.

There doesn't seem to be an attachment.

There is no additional wiring, and no re-routing of existing wiring. I just snipped the hot lead, tied the ends to MT1 and MT2, and jumped the resistor from G to MT2.

Is the power switch still switching the power lead as it sounds or switching the lead to G as it should be hooked up?

 

[c.coyle]

I have solved the hash problem. Details here.

 

[PViking]

No, what youre doing here is quite sound.
While many relays are used, triacs have been used quite a bit in the
appliance industry. Many of the Microwave ovens I've serviced used triacs
to switch on the primary of the high voltage transformer. The most simple
of the lot, use one of either below with an external snubber. Not that you
need it here, but you could add a small sensitive gate triac in the
mix driving the other, which would bring down the current across the
switch contacts down into the 10s of milliamps or less. If it were needed,
another device that could be used to handle the transients would be a
bidirectional diode. But apparently such details have been taken care
given the device you've chosen.

My apologies for bumping this old thread.

I tried to rebuild my Sansui AU11000 power switch, with plans to use the TRIAC design from this thread, but it won't stay together well enough physically due to damaged rocker contacts. I have a different switch I can use but it's designed for very low current guitar pickups. If I can get it down into tens of milliamps it will survive I think. As far as maintaining look and feel of the amp, the 3-way (not 5-way) Fender style selector switch middle position would correspond with the OFF and one extreme position would work for ON. The tactile feedback is even pretty much like the original. I would need to physically restrict movement to the other extreme position and maybe mount the switch at a slight angle vertically to more accurately match the original positions of the lever.

The Sansui switch has a flat blade lever that fits through a slot in the chassis and a slot in the amp face plate and the Fender selector switch has similar but shorter blade lever. I plan to have my son spot weld the original lever to this one so I can use the original knob. He welds razor blade edges together for fun and says he can do this.

I would be happy to add another sensitive TRIAC but I don't know how to do this; I'm hoping someone can point me to the right information because I can't seem to even word my search terms right with my limited knowledge.

Also, can increasing the resistor in series with the gate reduce the gate current surge if there is one? The data sheet for the Q4025L6 lists 2A peak gate current, unless I'm misreading that. I'm pretty sure that would fry this switch even if momentary.

Is such a switch just too wimpy for use as a power switch or can this be done? Are there better similar ones?

Thanks for any ideas.

 

[Hyperion]

Class aB solid-state has such a large range of primary currents that I consider thermistors unsuitable. True Class A solid-state and tube amps are a different story.

I like relays as well, but the triacs (thyristors) are sooo much easier, and the dozen or so units that I've installed them into work like a champ.

At the risk of it looking like I am 'teaching my grandmother to suck eggs' - I'd just like to make it clear that a Thyristor is not the same as a Triac, and only the Triac is suitable for this application.

 

[PViking]

Kindly see my bump on the previous page of this thread, #75.

 

[Hyperion]

Kindly see my bump on the previous page of this thread, #75.

If your power switch is so poor it can't even carry a momentary peak of 2A then you need get a replacement switch. As reducing the current to even 100mA isn't going to help much. Pretty much any switch with insulation and contacts rated for mains voltages should be able to carry this current, I would be wary of using that Fender switch for that reason, but you could try it and see how long it lasts?

 

[PViking]

Hyperion, thanks for replying. I too am wary of the Fender switch and I think the only way to make it work is to have another TRIAC to drive the gate of the power TRIAC. The TRIAC directly driven by this switch would need a very low peak gate current I guess, to avoid scorching the wimpy contacts. My knowledge of electronics just isn't enough to figure out how.

 

[roger2]

@ PViking, perhaps you could reconsider repair of your original switch? I would be leery of a double-triac simply because it has not been tried, as far as I know.

I saw pics of the contacts in your thread. IMHO they are good enough for to be used if the triac mod were to be implemented. Of course, other (better informed) opinions should be solicited.

As far as the mechanics of the switch, if you post pics in your thread maybe somebody can come up with an idea. Lots of creative thinkers here...