Class AB amplifier powered by non-polarized SMPS - cause for concern?

[DVDdoug]

"Breaker Finder 5000"

I've stolen a copy of that picture! My brother in-law is an electrician and he'll probably get a kick out of it. I assume he has a plug-in signal-tracker. I've probably shorted an outlet intentionally once or twice. But usually, it's trail-and-error and by now most of my switches & outlets at home are marked on the back-side of the cover plate.

Then one time I was working on an outlet at work. Someone told me which breaker it was, but they were wrong and it was still "hot". I didn't get shocked but I snipped a wire and somebody in the office on the other side of the wall said a bad word when his computer shut-off! Of course, I was going to kill his power when I shut off the correct breaker anyway. I think most people had gone home for the day.

 

[solderdude]

What I'm talking about specifically right now is if the supply's output is isolated from input. I'm wondering if this is the trick which allows SMPS to be used non-polarized.

SMPS are isolated from mains. There is a transformer in there that takes care of this. It just runs on a high frequency so they can be smaller.

Now ... there is something else going on with SMPS. In order for those designs to meet regulations the output needs to be 'coupled' for HF garbage on its DC output which is directed to mains.
This is done with a small value high voltage capacitor that, if it fails, should do so in a safe manner. It is called a Y-capacitor.
In the picture below it is the blue capacitor right next to the (yellow) transformer.

Now... that capacitor is a small value so it is low resistance for HF but high resistance for low frequencies (mains + most harmonics)
That capacitor is usually connected between the 0V (negative pole) and Neutral.

When you got the polarity wrong there is some leakage between Live and the DC out. That must be very low.
It is high enough that you may feel some light buzz when touching it. Not lethal or dangerous.
That current can thus make it in the audio path. This thus is orientation of the figure-8 or mains plug dependent. You can try rotating the plug to see if it helps.

So yes, SMPS are isolated but can have some leakage and yes those can have a leakage current which can be (and usually is) 'mains polarity' dependent.

There is a difference between isolated and Class-II. Class-I (3 prong with safety ground wire) is still isolated. The ground wire is needed if parts of the device are conductive and can be touched and cause a shock when something goes wrong. It is still isolated from mains.
Class-II can be used when there is no risk of electric shock for the user. The Y-capacitor must meet standards and is supposed to be safe.

 

[AnalogSteph]

That wasn't always true. I've seen older equipment (especially tube equipment) with a 2-prong plug where the chassis is connected to neutral (hopefully). I have a live concert recording that starts with the singer saying, "I keep getting shocked off this mic." And then he says something like "Will you see if you can switch to ground them out".

AC/DC sets (including guitar amps) were notorious for this. A certain famous Australian rock band probably had ones like that in their practice room at one time. AC/DC arguably survived the longest in TVs, well into the '70s.

You can make AC/DC work in a radio or TV where the user can be isolated from the set by using bakelite knobs on commonly-used controls, and input transformers on the antenna inputs. Guitar amps would have a mains polarity switch, but it's still pretty dodgy tech that would never meet any modern safety standards. Never try adding a composite input to an RF-only AC/DC TV set or an RCA input to an AC/DC tube radio (like a typical AA5), it is super dangerous.

 

[Doodski]

Why modify a collectors item piece that is in demand. Those old Pioneer receivers are very nice and changing it without electrical code certification is not going to make it popular.

 

[Timcognito]

AC/DC sets (including guitar amps) were notorious for this. A certain famous Australian rock band probably had ones like that in their practice room at one time. AC/DC arguably survived the longest in TVs, well into the '70s.

You can make AC/DC work in a radio or TV where the user can be isolated from the set by using bakelite knobs on commonly-used controls, and input transformers on the antenna inputs. Guitar amps would have a mains polarity switch, but it's still pretty dodgy tech that would never meet any modern safety standards. Never try adding a composite input to an RF-only AC/DC TV set or an RCA input to an AC/DC tube radio (like a typical AA5), it is super dangerous.

More evidence Keith Richards is imortal

When Keith Richards Was Nearly Electrocuted During a Stones Show

Keith Richards was almost electrocuted while performing in a Rolling Stones concert on Dec. 3, 1965 in Sacramento, Calif.

ultimateclassicrock.com

 

[mhardy6647]

View attachment 316429

Glad to see it's grounded. Safety first!

 

[Philbo King]

It's easier to filter the higher frequency, and as a bonus the switching frequency is normally above the audible range.


Almost all power supplies are isolated, whether they are earth-grounded or not. The primary & secondary transformer windings are separate. It's illegal to sell anything non-isolated in the U.S. unless it's "double insulated" with no possible electrical connection to the power line You could build a radio that way as long as the speakers are built-in and there are no electrical connections other than the power cord. But the power supply in an amplifier has to be electrically isolated from the power line because there are connections to the outside world.

That wasn't always true. I've seen older equipment (especially tube equipment) with a 2-prong plug where the chassis is connected to neutral (hopefully). I have a live concert recording that starts with the singer saying, "I keep getting shocked off this mic." And then he says something like "Will you see if you can switch to ground them out". Apparently they get it fixed but there is a (different) story of at-least one musician killed on stage. ...But there is also that false story about Sixto Rodriguez committing suicide on stage by lighting himself on fire!

I have a '65 Fender amp like that, till I removed the killer cap and put in an IEC cord.

 

[Timcognito]

Glad to see it's grounded. Safety first!

It's not mine (from the net). Electricity scares me. I don't like stuff I can't see.

 

[mhardy6647]

It's not mine (from the net). Electricity scares me. I don't like stuff I can't see.

Well, I was being facetious -- as I assumed you were, too. Presumably the switch applies a dead short across the line, which should identify the breaker for the circuit quickly. That's probably not the best way to do it, though.

So... electricity is like... well... a lot of things. A healthy respect is indicated, but fear is a bad approach. Fear can lead to accidents, as can overconfidence.
Shorts across the AC mains are best avoided, circuit breakers or no. Also, never assume a circuit is wired correctly (as @fpitas has already mentioned). "Trust but verify".

Similarly, with great care, it's perfectly possible to work on a live circuit... but it's always a better choice to de-energize the circuit unless it's utterly necessary not to.

 

[XpanD]

Be careful with those Korad power supplies -- they're super useful, but some models (like my KA3005P) kick out a seriously nasty voltage spike when powered on. Not familiar with your specific equipment, but worth keeping in mind. For what it's worth, I usually connect my leads after I've dialed everything in, and after I've enabled the output.

 

[Timcognito]

Well, I was being facetious -- as I assumed you were, too. Presumably the switch applies a dead short across the line, which should identify the breaker for the circuit quickly. That's probably not the best way to do it, though.

So... electricity is like... well... a lot of things. A healthy respect is indicated, but fear is a bad approach. Fear can lead to accidents, as can overconfidence.
Shorts across the AC mains are best avoided, circuit breakers or no. Also, never assume a circuit is wired correctly (as @fpitas has already mentioned). "Trust but verify".

Similarly, with great care, it's perfectly possible to work on a live circuit... but it's always a better choice to de-energize the circuit unless it's utterly necessary not to.

Amen Brother

 

[restorer-john]

e careful with those Korad power supplies -- they're super useful, but some models (like my KA3005P) kick out a seriously nasty voltage spike when powered on.

Can you post a 'scope shot of the secondary transient on power up- I was just about to order a few of the higher voltage/current ones myself? Your 3005P has a proper transformer in it, yes? (not SMPS)

 

[mike7877]

SMPS are isolated from mains. There is a transformer in there that takes care of this. It just runs on a high frequency so they can be smaller.

Now ... there is something else going on with SMPS. In order for those designs to meet regulations the output needs to be 'coupled' for HF garbage on its DC output which is directed to mains.
This is done with a small value high voltage capacitor that, if it fails, should do so in a safe manner. It is called a Y-capacitor.
In the picture below it is the blue capacitor right next to the (yellow) transformer.

Now... that capacitor is a small value so it is low resistance for HF but high resistance for low frequencies (mains + most harmonics)
That capacitor is usually connected between the 0V (negative pole) and Neutral.

When you got the polarity wrong there is some leakage between Live and the DC out. That must be very low.
It is high enough that you may feel some light buzz when touching it. Not lethal or dangerous.
That current can thus make it in the audio path. This thus is orientation of the figure-8 or mains plug dependent. You can try rotating the plug to see if it helps.

So yes, SMPS are isolated but can have some leakage and yes those can have a leakage current which can be (and usually is) 'mains polarity' dependent.

There is a difference between isolated and Class-II. Class-I (3 prong with safety ground wire) is still isolated. The ground wire is needed if parts of the device are conductive and can be touched and cause a shock when something goes wrong. It is still isolated from mains.
Class-II can be used when there is no risk of electric shock for the user. The Y-capacitor must meet standards and is supposed to be safe.

To verify my understanding, the transformer isn't an isolation transformer, and the Y capacitor is responsible for the LF isolation - correct?

 

[mike7877]

Be careful with those Korad power supplies -- they're super useful, but some models (like my KA3005P) kick out a seriously nasty voltage spike when powered on. Not familiar with your specific equipment, but worth keeping in mind. For what it's worth, I usually connect my leads after I've dialed everything in, and after I've enabled the output.

I haven't blown anything up yet fortunately, but I'll check with my scope some time before the next time I use it for something that's not charging a battery (I usually use a switching supply for that, but I lent it)

 

[mike7877]

Can you post a 'scope shot of the secondary transient on power up- I was just about to order a few of the higher voltage/current ones myself? Your 3005P has a proper transformer in it, yes? (not SMPS)

I've had two versions of the 30V 5A Korad. Still do, actually. I broke one, unfortunately, when modifying the fan circuit (they're noisier than they need to be, needs resistor + capacitor to filter the supply voltage, which is PWM'd at a ridiculously low frequency for some reason, causing a loud RRRRR. I forget the values and they're heatshrinked now (in the working unit), so if you want to do the same you'll probably have to do some trial and erroring). The way I broke it was so stupid , too - dropped my screwdriver when I started to work on it, and it was on . I wasn't trying to work with it on lol, it was a mistake -I forgot

Both units have rather large transformers, and weigh 8-10lbs

I think if they were going to switch the design to a SMPS powered linear regulator, that'd be a big enough change for a new model

 

[solderdude]

To verify my understanding, the transformer isn't an isolation transformer, and the Y capacitor is responsible for the LF isolation - correct?

The transformer IS an isolation transformer. Rectified and switched mains side is fully isolated from the low voltage output of that transformer.

The Y capacitor (not drawn in the circuit above) is responsible for no isolation for RF so output and mains are directly coupled for radio frequencies is high enough resistance at mains frequencies to ensure the mains leakage current is below the maximum allowed current.
The feedback/regulation in most SMPS is done with optocouplers.

 

[mike7877]

The transformer IS an isolation transformer. Rectified and switched mains side is fully isolated from the low voltage output of that transformer.

The Y capacitor (not drawn in the circuit above) is responsible for no isolation for RF so output and mains are directly coupled for radio frequencies is high enough resistance at mains frequencies to ensure the mains leakage current is below the maximum allowed current.
The feedback/regulation in most SMPS is done with optocouplers.

Gotcha, now I get it. Thank you!

I remember measuring 90V on the metal chassis of my Dell Latitude in 2013 (when it was plugged in, obviously), and being able to feel a buzz. The voltage dropped easily, so obviously wasn't capable of high current.

Do you happen to know if the leakage current of certified power supplies (UL etc) is kept to at/below the GFCI limit?

 

[solderdude]

It should be well below the GFCI limit.
Most should be below 0.1mA
Y capacitors usually are in the 1nF - 10nF range.
Most are around 2.2nF so at 60Hz form a 1Mohm resistance

Even multiple SMPS in parallel should not trigger a GFCI

 

[XpanD]

Can you post a 'scope shot of the secondary transient on power up- I was just about to order a few of the higher voltage/current ones myself? Your 3005P has a proper transformer in it, yes? (not SMPS)

Yup, it has a transformer.

I hadn't actually measured my own unit yet, only recently got a scope -- was going off of reports/scope shots from other people.

Here's some results from mine when turned on:

...and when turned off:

Hope I got these right, still figuring out how to use/interpret this thing. (and oscilloscopes in general)

Interestingly enough, I don't see any real spikes when I enable the output on mine -- it only seems to freak out when I press the proper power button.

 

[solderdude]

Measurement artifacts which may not be there in reality.
Could be leakage currents, could be induction in wires, could be improper placement of the probe or its ground wire, could be incorrect grounding scheme of DUT and scope.
Such are the difficulties with measuring HF switching devices, .or even having any HF emitting stuff in the vincinity.