Post by thread:[EE] power supply, what am I doing wrong?

part 1 1625 bytes content-type:text/plain; charset=ISO-8859-1; format=flowed (decoded quoted-printable) Hello Attached is the simplified circuit of a power supply I'm building. TRANSHI and TRANSLO are the high and low voltages of a transformer, VAC is the mains supply. The purpose of this circuit is to control an electromagnet, the low voltage feeding a circuit containing a dsPIC reading sensor data, the high side supplying up to 1A, 200V to the magnet coils, controlled via a MOSFET which in turn receives PWM from the PIC and its excitation circuit. The problem is that when I connect the two grounds of the high voltage and low voltage supplies smoke happens. The low voltage ground is held steadily at half the high voltage and even if I put a resistor in there an uncomfortable amount of juice flows and there�s no significant difference in voltage levels. I think I'm missing something obvious here. My transformer is isolated, the ohmmeter doesn't even show a resistance reading resistance between the coils. Shouldn't the low level supply, when the grounds are connected, float to meet the high voltage ground potential? I know I could use a 1:1 transformer on the high side but it would be heavy and big and lossy and I'd rather not. As much as I think about my supply design I can't see anything wrong with it. I could use an optocoupler but I'm very worried about adding delays in a 100 kHz PWM line which may operate down to as little as 0,1% (10 ns) duty cicle. That would also require a transformerless supply for the MOSFET excitation circuit so I'd really like to make this circuit work. Or at least find out why this doesn't work. Anyone know? Thanks. part 2 8467 bytes content-type:image/png; name=supply.png (decode) part 3 35 bytes content-type:text/plain; charset="us-ascii" (decoded 7bit)

On Thu, 2007-05-17 at 21:56 +0100, peter green wrote: > > Or at least find out why this doesn't work. Anyone know? > my first thought would be that the ground of the low voltage supply is somehow (possiblly via a peice of test or programming equipment) connected to mains ground. That would also be my guess. If you've got say the ICD2 connected to your PIC, and that's connected to your PC, which connects it's local ground to earth ground, magic smoke is the result. If you are using a laptop perhaps the power adapter for the laptop grounds the negative rail? TTYL

On 5/17/07, Herbert Graf <.....mailinglist3KILLspam@spam@farcite.net> wrote: > On Thu, 2007-05-17 at 21:56 +0100, peter green wrote: > > > Or at least find out why this doesn't work. Anyone know? > > my first thought would be that the ground of the low voltage supply is somehow (possiblly via a peice of test or programming equipment) connected to mains ground. > > That would also be my guess. If you've got say the ICD2 connected to > your PIC, and that's connected to your PC, which connects it's local > ground to earth ground, magic smoke is the result. > > If you are using a laptop perhaps the power adapter for the laptop > grounds the negative rail? Actually I've disconnected everything now, I'm testing with just the parts in the circuit I attached and it doesn't work. (I already burned the computers COM port, luckily I also have a usb-rs232 adapter).

Yikes! I'm glad you haven't been hurt by this. Yeas - its obvious. Here's the deal. Your computer and probably much of your mains-powered test equipment is grounded to the AC mains ground. When you connect the negative rail of your HV supply to ground, you are essentially shorting out either D7 or D8. Draw out the entire diagram of the input side of the HV supply. Make sure you include the neutral-to-ground link in your building's Service Entrance in your diagram. Here's another quick test to show you just how dangerous this whole thing is. Connect the AC mains to only the HV bridge rectifier input. Now grab 60W or so incandescent light bulb (your desk light will do quite nicely. Connect one side of the light bulb to the AC mains ground line, then touch the other side of the light bulb to your HV ground. Notice how bright the lamp is (about half brightness). The way I see it, you've got a couple of choices. a) Use a suitably rated AC mains isolation transformer. b) Insert some form of isolation barrier in your gate drive signal. This can be optical, pulse transformer, the neat little capacitive couplers from Burr-Brown, etc. Sorry to be the bearer of bad news. dwayne At 02:37 PM 5/17/2007, Cristóvão Dalla Costa wrote: {Quote hidden}>Hello > >The problem is that when I connect the two grounds of the high voltage >and low voltage supplies smoke happens. The low voltage ground is held >steadily at half the high voltage and even if I put a resistor in >there an uncomfortable amount of juice flows and there´s no >significant difference in voltage levels. > >I think I'm missing something obvious here. My transformer is >isolated, the ohmmeter doesn't even show a resistance reading >resistance between the coils. Shouldn't the low level supply, when the >grounds are connected, float to meet the high voltage ground >potential? -- Dwayne Reid <dwaynerKILLspamplanet.eon.net> Trinity Electronics Systems Ltd Edmonton, AB, CANADA (780) 489-3199 voice (780) 487-6397 fax http://www.trinity-electronics.com Custom Electronics Design and Manufacturing

On 5/17/07, Dwayne Reid <.....dwaynerKILLspam.....planet.eon.net> wrote: > Yikes! > > I'm glad you haven't been hurt by this. > > Yeas - its obvious. > > Here's the deal. Your computer and probably much > of your mains-powered test equipment is grounded > to the AC mains ground. When you connect the > negative rail of your HV supply to ground, you > are essentially shorting out either D7 or D8. Hello With nothing but the diodes and capacitors in the schematic I still have trouble (can't connect the sides or smoke appears). No computer connected, nothing else. Am I still missing something? You have alerted me to the need to isolate the computer ground though when in the future the circuit connects to a PC. Looks like I'll need to isolate everything anyway. Thanks.

Cristóvão Dalla Costa wrote: {Quote hidden}> On 5/17/07, Dwayne Reid <EraseMEdwaynerspam_OUTTakeThisOuTplanet.eon.net> wrote: >> Yikes! >> >> I'm glad you haven't been hurt by this. >> >> Yeas - its obvious. >> >> Here's the deal. Your computer and probably much >> of your mains-powered test equipment is grounded >> to the AC mains ground. When you connect the >> negative rail of your HV supply to ground, you >> are essentially shorting out either D7 or D8. > > Hello > > With nothing but the diodes and capacitors in the schematic I still > have trouble (can't connect the sides or smoke appears). No computer > connected, nothing else. > > Am I still missing something? Possibly. Dwayne's explanation is right; but your circuit should work on it's own with nothing attached. So maybe it did work right until you attached something and then fried a diode. If, as Dwayne mentioned, D7 or D8 got shorted out, then D5 or D6 (or both) might have easily been overloaded and either opened or shorted. Although you can replace them if they are bad, you won't be able to connect up your computer to it safely.

Marcel Duchamp wrote: {Quote hidden}> Cristóvão Dalla Costa wrote: > >> On 5/17/07, Dwayne Reid <dwaynerspam_OUTplanet.eon.net> wrote: >> >>> Yikes! >>> >>> I'm glad you haven't been hurt by this. >>> >>> Yeas - its obvious. >>> >>> Here's the deal. Your computer and probably much >>> of your mains-powered test equipment is grounded >>> to the AC mains ground. When you connect the >>> negative rail of your HV supply to ground, you >>> are essentially shorting out either D7 or D8. >>> >> Hello >> >> With nothing but the diodes and capacitors in the schematic I still >> have trouble (can't connect the sides or smoke appears). No computer >> connected, nothing else. >> >> Am I still missing something? >> > > Possibly. Dwayne's explanation is right; but your circuit should work on > it's own with nothing attached. So maybe it did work right until you > attached something and then fried a diode. > > If, as Dwayne mentioned, D7 or D8 got shorted out, then D5 or D6 (or > both) might have easily been overloaded and either opened or shorted. > > Although you can replace them if they are bad, you won't be able to > connect up your computer to it safely. > > Somehow related to the question of the OP: In a PC are digital and frame grounds connected together and in turn connected to AC main ground? Gaston

Gaston Gagnon wrote: > Somehow related to the question of the OP: > In a PC are digital and frame grounds connected together and in turn > connected to AC main ground? Usually, yes. The power supply in your average PC grounds the chassis (to AC earth), and the chassis is also connected to low-side power ground. -- Hector Martin (@spam@hectorKILLspammarcansoft.com) Public Key: http://www.marcansoft.com/marcan.asc

> > If you are using a laptop perhaps the power adapter for the laptop > > grounds the negative rail? > > Actually I've disconnected everything now, I'm testing with just the > parts in the circuit I attached and it doesn't work. (I already burned > the computers COM port, luckily I also have a usb-rs232 adapter). Oscilloscope ground clip providing the short to ground? -- Brent Brown, Electronic Design Solutions 16 English Street, St Andrews, Hamilton 3200, New Zealand Ph: +64 7 849 0069 Fax: +64 7 849 0071 Cell: 027 433 4069 eMail: KILLspambrent.brownKILLspamclear.net.nz

On 5/18/07, Brent Brown <RemoveMEbrent.brownTakeThisOuTclear.net.nz> wrote: > > > If you are using a laptop perhaps the power adapter for the laptop > > > grounds the negative rail? > > > > Actually I've disconnected everything now, I'm testing with just the > > parts in the circuit I attached and it doesn't work. (I already burned > > the computers COM port, luckily I also have a usb-rs232 adapter). > > Oscilloscope ground clip providing the short to ground? > Bingo!! I wasn't even using the scope but surely the probe was still grounded to the circuit. Thanks, I'll sleep easier tonight :)

On 5/17/07, Cristóvão Dalla Costa <spamBeGonecdallacostaspamBeGonegmail.com> wrote: > On 5/18/07, Brent Brown <TakeThisOuTbrent.brownEraseMEspam_OUTclear.net.nz> wrote: > > > > If you are using a laptop perhaps the power adapter for the laptop > > > > grounds the negative rail? > > > > > > Actually I've disconnected everything now, I'm testing with just the > > > parts in the circuit I attached and it doesn't work. (I already burned > > > the computers COM port, luckily I also have a usb-rs232 adapter). > > > > Oscilloscope ground clip providing the short to ground? > > > > Bingo!! I wasn't even using the scope but surely the probe was still > grounded to the circuit. Thanks, I'll sleep easier tonight :) Oh boy. Never let software engineers doing hardware... :)

>You have alerted me to the need to isolate the computer ground >though when in the future the circuit connects to a PC. Well you should certainly be running the whole mains side of an isolation transformer while you do your development, before you let the magic smoke out of yourself. >Looks like I'll need to isolate everything anyway. If the circuit still lets the smoke out when you have no instrumentation connected, my pick would be that the frame of the transformer is connected to the mains earth lead, and you have connected the secondary ground to the same point, but not the negative end of the 10u 500V capacitor. Once you connect that point to mains ground the magic smoke will certainly be released from the diode bridge.

On Thu, 2007-05-17 at 22:44 -0400, Gaston Gagnon wrote: > Somehow related to the question of the OP: > In a PC are digital and frame grounds connected together and in turn > connected to AC main ground? Pretty much universally yes. The AC ground is connected to the frame in the power supply. The digital ground is connected to the frame through the screws holding the motherboard to the case, and any expansion card plugged in and screwed onto the case. Even the serial/parallel/USB port sheilds are connected to digital ground and are in contact with the case. TTYL