Thoughts on repositioning the conventional power switch.

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Ethan

Administrator
Admin
Moderator
Joined
Jun 3, 2004
Messages
1,602
Location
DC
Vitually every power supply schematic I have seen positions the power switch between the AC inlet and the power transformer primary. Often times the wire to the power switch must extend a few inches to the front or back panel--along with it a radiating field of lovely hum.

WHY?

Why not put the switch after the filter section, before the load. I see a few benefits... Equipment always plugged in, but not switched on, will benefit from having filter capacitors always'formed', no repeated surges from the power switch that may shorten the life of silicon, and also clean DC flowing through the wires of the power switch so that it could be placed closer to sensitive components if needed.

Is there a good reason not to do this?
 
For personal use, why not, it sounds like a neat idea. There might be problems when letting go of such equipment to someone else. You may want to put an AC switch in the back of the box where it stays far from sensitive circuits yet satisfy possible safety concerns.
Plus when testing or troubleshooting I would rather have an AC switch even in the back than have to yank and plug the cord constantly.
 
It's a UL thing. Or CSA, or wherever your from saftey code. But your right, there is no reason why DIY'ers could not do this. It would be the equivalent of a standby switch on a Fender. But heck, if your going to consume power all the time due to transformer losses and bleeder resistors, why not just leave the whole circuit on like PRR does with his tube stuff?

If the x-former is protected with maybe an MOV and maybe a .047 cap, there is no reason why you could not leave it on. I used to do this with my DIY Twin Reverb because the pwr switch kept either breaking from being thrown in the car after a wild house party jam session, having the drummers stuff in the back seat fall on it (why can't you drummers get your own car for crying out loud? Oh yeah, I forgot. No drivers liscense.:grin: You know how many headliners have been poked thru by hi hat stands in my car?:evil: )
or it would fail electrically. So I bypassed the darn thing. At least I knew that the tubes would light up when I plugged it in.

Lytics do have a lifespan hours rating, so that is something to consider

The mass-ive pass-ive has a steel rod that goes from the front panel to the switch mounted on the rear panel, so that is something to consider.
 
Anyone know what transformer losses would be on an unloaded circuit?
I like the idea but the power consumption and lytic lifespan issues bother me.
One other -slightly OT - thing, would those big ferrites clamped round some computer gear power inlets reduce startup 'whump' if used on a piece of (line level) audio gear, or is this not the same thing?
 
some gear has a long plastic "stick" attached to the front switch. It switches the power switch located close to the power inlet.
 
The ferrites will do nothing. They are common-mode and effective at high frequencies at that.

The power consumption is a complex thing. With sufficient magnetization inductance the current is not that large, and based on that and the wire resistance you can estimate the real power dissipated. There are eddy current and hysteresis losses in the core as well.

Cheap trafos often have considerable losses---notice how hot a wall wart with no load gets after several hours of just being plugged in.

If you are concerned just run the candidate trafo for a few (well, several maybe) hours with the secondary open and feel how warm it becomes. Then guess at how many watts that is.

Did I tell the story already (semi-senior moment here) about the product where I left the d.c. on and switched the amplifier load/bulk cap common with the power switch, to avoid the very thing (a.c. fields) that CJ mentions?
 
> Why not put the switch after the filter section

Basic rule: a "POWER" switch should kill ALL the power.

Imagine if the transformer caught fire? Turning off the "power" won't help.

Imagine if you turn off the "power" and think it is "safe" to stick your hand inside? Yes the amp may be dead, but the transformer and filter is still live.

I would almost rather have no power switch than a "fake" power switch that didn't really break power to 99.9% of the stuff inside. With no power switch, you are on notice that the thing is dangerous any time it is plugged in.

> a radiating field of lovely hum.

If both conductors are tight and twisted together, the field quickly cancels with distance. In fact we usually have more trouble with power transformers, either the one in the box or the one in the boxes stacked above/below it. You can't put boxen on top of a Fender Rhodes, and I have a HiFi amp that I can't stack with a cassette deck.

> put the switch after the filter section, before the load.

A specific objection is: the amp is powered-up with a bang. If you switch before the transformer, power comes up with some ramping, even if only a few milliSeconds. Some +/- powered circuits will latch if the + and - rails don't come up about the same: with a 2-pole switch you are pretty sure to get one all the way up before the other pole even touches. You could add more caps after the switch, but the surge current approaches infinity. You could add more resistance.... but you already have unavoidable resistance in the transformer, why buy more? Also the current is less on the line side (for most modern gear that eats less than 100-200V), less contact abuse at switch-on. (OTOH, more abuse at switch-off due to higher voltage arcing longer.)

Yes, a lot of gear works as you suggest. Common practice now is to have wall-power in the box all the time. When "off", it is still on, though perhaps only with a standby supply. An ATX PC power supply (or AT&T 3B2 or modern Mac) has like 5V 0.1A available all the time. A signal derived from this supply (via cheap low-volt pushbutton, keyboard, or timer) switches-on the 5V 30A and other big-power rails. I remember when TV sets kept the tube heaters working part-voltage for "instant on". I swear my VCR's "power" switch just turns off the display and over-rides the controls.
 
[quote author="tk@halmi"]You may want to put an AC switch in the back of the box...[/quote]

For this you could use one of those IEC fused AC inlets that have a switch built in. That way you still have a power switch, inconveniently located on the back panel, but still have the load switch on the front.
 
If you do have power on the rear of the unit you could always put the gear in a rack with a power distro. I've got different kinds of gear in different racks.
pres and other tracking gear off in one unit etc. saves a bit on my bills not having to have a bunch of pre's on when im mixing. keeps the room a bit cooler too.

And I like to put a piece of sheet metal in between my psu side of the enclosure and the circuit side. just seems like a good idea whenever possible.
Sleeper
 
[quote author="Sleeper"]And I like to put a piece of sheet metal in between my psu side of the enclosure and the circuit side. just seems like a good idea whenever possible.[/quote]
Our Sr Engineer here at work pointed out to me the other day that this can create a shorted magnetic ckt around a trafo and actually cause more radiation into the circuits. Not sure that I can see a difference myself, but an interesting idea.

PRR did you use the word "boxen" on purpose? Where did you hear that used?

Peace!
Charlie
 
[quote author="SonsOfThunder"]
PRR did you use the word "boxen" on purpose? Where did you hear that used?[/quote]

"boxen" is very common nomenclature around guru-ish types. It's wordplay that stems from the fact that the plural of words ending in x is sometimes 'xen', like multiple Ox are Oxen.

VERY common in the computer world, as in 'I have a rack of UNIX boxen in that closet'
 
Our Sr Engineer here at work pointed out to me the other day that this can create a shorted magnetic ckt around a trafo and actually cause more radiation into the circuits.

Hmm... thats an interesting thought.

I usually try and keep at least 1/2" minimum between the sheet metal
(I'm using .075 aluminium to be more precise)
If you think about most tube amps you'll see the power transformer is outside the box so it seemed like a good rule of thumb practice. I think I saw SSLTech do it on his 9K pre images... I felt justified, and it's real easy to do.

I'm mostly using toroids anyhow, but...
I seem to remember a conversation involving PRR, a toroid and a tuna can, but I don't recall the exact details. I know in that case the shielding was conforming very tightly to the trafo, like a mu metal can would.

I've never done any testing to see whether or not this makes a difference, but when I have done this I haven't noticed anything negative effects.

I'm also not 100% sure whats meant by a "shorted magnetic circuit"
I guess you mean some kind of coupling.

I'd like to know more.

Sleeper
 
If the metal forms a loop like a turn of wire THROUGH the toroid then you are in trouble. Just putting a box around one is not the same thing.

Putting a bolt through the toroid and allowing it to contact the top and bottom of the chassis is an example of a shorted turn.
 
Hey BC.
That's it, that's what PRR had in the tuna can!

Even if it's only marginally effectice, I also like the look of the shield barrier.
I think it looks like caring. :wink:
Sleeper
 
Back
Top