Tapping into the ~400vdc from an active pfc circuit

[justinjja]

Never mind I'll do this one without help, thanks anyway.







I need something in the range of 300-400vdc for a 3phase motor driver I'm working on.
My first plan was to just rectify a 240v mains circuit to ~340v, but I figured modding a desktop psu would work just as well.

Plan is to plug the psu in without an earth wire, so the 12v/5v grounds can be connected to the 400v ground.

I know I'll have to make sure I program the micro controller out of circuit so I don't blow the whole thing up. (USB ground would short out the circuit)
Also the case won't be earthed, so I'll have to be careful not to shock my self too.

Anything else to watch out for?

[drussell]

Do NOT disconnect the ground!

The 5V/12V output grounds should already be at ground potential anyway.

You still have to treat the 400V just like you would mains, though!  Be sure you know exactly what you're doing since it is completely un-isolated, potentially lethal, and has high current available to fry things or start fires.

The fact that you're thinking about disconnecting chassis ground leads me to suspect that you haven't thought this through completely and may not be at the level where you should be playing with the mains in random fashion.

[justinjja]

So the 5v/12v grounds are at my house ground,
but the 400v capacitor is not, that is why the ground needs to be disconnected.

[ArthurDent]

My advice is that with your level of understanding that you may end up getting injured or electrocuted if you try this, whatever your vague idea of this is.

Just forget it or hire someone who understands electrical circuits to do it.

[justinjja]

The 400v capacitor is floating with respect to my house ground.

If I tried to use both the 12v and the 400v it will short out.

[justinjja]

actually I can just power it from 240v (2 hot legs)
so I won't have to cut the earth wire to connect the 400v and 12v grounds.

That eliminates one problem

[drussell]

EGADS, NO!

STOP NOW!

YOU DO NOT UNDERSTAND WHAT YOU'RE DOING!

[justinjja]

? please explain whatever it is you think I don't understand.

[ArthurDent]

First of all your crude drawings make no sense and indicate you don't have the knowledge to play around lethal voltages.

You have got very good advice not to try whatever you think you're trying to do but you seem to be hell bent on ignoring all advice and proceeding.

Don't, just don't

[justinjja]

Just wondering if there was anything specific,
don't read into the pics too much, ms paint is not very good for this type of thing lol

[SeanB]

With active PFC the input goes through a bridge rectifier, so the negative of the capacitor after the active PFC stage is going to float at roughly half mains voltage ( actually will swing from negative mains peak to 0V per cycle, average voltage is half the mains voltage) and the positive side will be 400V above that.

Your PC power supply is designed to have the secondary voltages referenced to mains ground, and there are multiple connections to there on the board as well. disconnecting mains earth creates a lethal voltage on the case, the wires carrying the 5V and 12V supplies, and is generally not a good idea. Will create a lot of RF interference to every thing in the vicinity as well, and might cause improper operation of poorly shielded equipment. Again, the lethal voltages on a big case thing out in the open right by you is a bad idea.

Better for your design is to get a separate smaller SMPS module ( two wire supply type, so the output is floating and not mains ground referenced) and use that as your power supply for the circuit. That way you can program it with the main input PFC side disconnected and non powered.

[justinjja]

Hum, that doesn't quite make sense to me.

If cutting the ground created a lethal voltage on the case,
then the ground would have to normally be carrying some current.
and if the ground was carrying current, it would trip when plugged into a gfi.

grounding the case is to protect from something like a lose wire touching the case,
and then the case becomes lethal.

EDIT:
You can find anything on the internet:
https://www.rcgroups.com/forums/showthread.php?1170241-Using-two-power-supplies-for-higher-voltage-capacity-chargers-safety-issues

Read through a few pages, nobody died lol.

[ogden]

Hum, that doesn't quite make sense to me.

Then drop your "project" as it is said multiple times here already. You have my vote as well.

If you are not able to get that you must not connect DC negative rail of full-bridge rectified AC mains to ground (where neutral is tied to), then this project is not for you. You can [safely for yourself] try, but you will blow fuses or rectifier diodes or PFC circuit.

[justinjja]

That isn't what I said at all.
That is the reason I originally was going to cut the ground cable.

[drussell]

That is the reason I originally was going to cut the ground cable.

Yeah, but this is where things go off the rails.  That's not how you do something like this, even for experimentation.  You don't lift the chassis from ground, you modify the rest of your apparatus' circuitry to work properly while still maintaining safety and sanity.

Before you even think about running something like this directly off the mains, you need to build your required style of circuitry in a low voltage version and thoroughly test everything.  Then, once you have everything working properly in a low-voltage version, you can move on to making a safe version of the motor power side and test it with your high voltage motor.

If everything were isolated properly with a transformer, you could theoretically modify the secondary, low voltage side with your control circuitry to run at the same potential as the motor side, provided that everything to the control side was properly isolated with transformers and optocouplers with proper isolation and insulation.  I'm not sure you understand the intricacies of doing that safely yet.

You could also just as easily design it with them them both isolated from each other (essentially as your computer PSU is originally, in stock form), again, using proper optocouplers and transformers, etc. so that it would always be properly isolated, regardless of what the potential at the actual motor is relative to your control circuitry.  Then you still have danger in your motor area, potentially more dangerous than regular mains connected stuff since we're talking high current 400V, but can be done, provided you take all appropriate precautions.  Again, I'm not sure you appreciate the danger involved with high energy, high voltage circuits in a configuration like this.  This could easily be far more dangerous than the 120 volts that comes out of the wall.

One way or another, you need proper isolation!  Removing the ground and running your control circuitry at 400 volts without isolation is not the way...  That is lunacy.  In some ways, that is even worse than the old hot-chassis TVs and radios. 

[viperidae]

If you tie the negative of the filter cap to ground (which neutral is tied to) then when the active line goes on its negative cycle it will short to ground via the diode in the bridge rectifier. Boom.

[drussell]

Never mind I'll do this one without help, thanks anyway.

Well, if you insist on going it alone, please at least use proper fusing, consider a GFCI for a bit of added safety and learn what a "dim-bulb tester" is and use one while experimenting with things that might have the potential to go bang.

I blew up many a row of 7400 series logic and TRIACs and such building lighting controllers and dimmers when I was learning power electronics as a kid and it certainly helped me understand how this stuff works.  I worked with >500 Vdc 8417 tube amplifiers and such which you need to have a healthy respect for.  The first time your multimeter starts beeping maniacally at you and showing 1050 Vac on it makes you stop and think for a moment about what you're doing. 

Just remember that 400 Vdc packs a lot more of a wallop than the not-so-scary 120V that comes out of the wall.   Zapping yourself may well not kill you every time but it is a painful way to learn about high energy circuits!  Be careful!!

[ArthurDent]

This is a perfect example of the Dunning-Kruger Effect, described in the link. I just hope justinjja survives to watch this video.

[ArthurDent]

Sometimes checking other threads the person has started can give you some clues as to why they trying something that would make everyone who has some technical background cringe. Here is a post in another thread that helps explain it.

"Bitcoin mining in my garage, I have my 200A service maxed out."