Microcontroller current sinking issues caused in HV PSU / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Microcontroller current sinking issues caused in HV PSU

Move Thread

LAN_403

V3rge

Sat Aug 06 2011, 12:10PM

Registered Member #3253 Joined: Wed Sept 29 2010, 12:21AM
Location: New Zealand
Posts: 13

I've recently built a new ignition coil based unregulated HV power supply, which I have been testing over the last few weekends and began migrating it into an enclosure today. It is driven by a PIC12F683 which puts out a variable duty cycle based on the ADC reading from a pot. There is one major issue I have been having over the last few weeks with regard to the interference generated from HV side, which was never a suprise at all. I have continuously modified the design since I started building it, trying enclosed metallic shielding methods and changing wire lengths, which had a very significant impact on the stability. Before putting into the small enclosure, the system was working absolutely perfectly, I could easily run it at 1W @ 1-2kV all the way up to 60W at around 15kV.

Now the main issue i'm having is with regard to the microcontroller, especially since the migration to the enclosure. What happens is it will run stable if the duty cycle is in a certain range (almost all dependant on external factors - spark gap size and wire lengths I have found) but once it moves out of this zone the controller will stop and start sinking current to who knows where - this is the issue that needs fixing. After this happens, the power can just be switched on and off and everything works fine, until you push it out of its comfort zone again.
The power to the microcontroller is simply filtered and regulated with a 7805, and I have monitored the regulated 7805 output voltage but do not see anything influencying the regulation, although there are sometimes very large spikes that can be seen coming through the probe, they happen quite often and do not worry the microcontroller during the long periods of time i've spent watching it.

I have no doubt all of the issues are to do with the interference generated from high voltage side, but I do not know specifically how to fix it since I haven't been able to pinpoint the primary cause of the problem.

Any suggestions as to how this can be resolved or what I can do to test the primary cause of this issue would be appreciated. I believe it is most likely to do with the voltage provided to the PIC, which is a bit high at 5.24V and a little instability from interference may push it too far, but what else can I do eliminate the effect of the HV output on the controller?

Adam Munich

Sat Aug 06 2011, 02:19PM

Registered Member #2893 Joined: Tue Jun 01 2010, 09:25PM
Location: Cali-forn. i. a.
Posts: 2242

Well, 5.1 volt zeners and ferrite beads everywhere would be the first step I'd take, and if that doesn't work try optosiolating the uC.

Electra

Sun Aug 07 2011, 10:48PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

Try putting the microcontroller board inside its own metal box to shield it.

If the problem is still there then itâ€™s being coupled by the wiring, ferrite rings and decoupling caps will help deal with this, as will using screened leads for signals and ground the screen.

2Spoons

Mon Aug 08 2011, 12:23AM

Registered Member #2939 Joined: Fri Jun 25 2010, 04:25AM
Location:
Posts: 615

I had a similar issue with an 8051 variant many years ago, in an electric fence system. Interference was causing a thyristor-like latchup on the ADC supply of the micro, resulting in the ADC supply being shorted out and the ADC reading full scale all the time. I tried all sorts of shielding, filters, clamps - nothing ever worked.
Luckily the ADC supply was through a shunt regulator, so I used another port pin to deliberately short out the adc supply when the fault was detected, allowing the ADC pin to unlatch and everything would run again.

As for solutions - I can only suggest using feedthroughs + transil clamps and filters on every wire going in or out of your box. Also make sure the lid shorts to the box all the way around, as a slot can let RF in.

V3rge

Mon Aug 08 2011, 08:45AM

Registered Member #3253 Joined: Wed Sept 29 2010, 12:21AM
Location: New Zealand
Posts: 13

I placed a ferrite bead on the postive wire feeding the regulator, straight after I tried it with this I thought I noticed a significant improvement in stability, but after a few minutes it would start regularly stalling. The regulator, microcontroller and optoisolator were all enclosed in a grounded metal shield before I inserted it into the new enclosure, it had very little effect on the operation of the circuit so I left it out. Decoupling capacitors are already in use, and sadly I don't have any small screened leads sitting around, but I will definately take a look into getting some.

2Spoons, that latching effect is a much better way of describing what mine is doing, with the exception of it happening with the ADC. My whole circuit system collapses during that condition, and everything ceases functioning. The controller will sink about 200mA during that state which is going straight through the Vdd pin, not a happy controller with that pumping through it.

I will add some more filtering from what you've all suggested when I get around to it. Thanks for the help, I really appreciate it and will see how it goes with some more additions.

2Spoons

Mon Aug 08 2011, 10:54PM

Registered Member #2939 Joined: Fri Jun 25 2010, 04:25AM
Location:
Posts: 615

Seems you could benefit from some heavy pre-filtering before the regulator. Say 100uH or so inductor plus a good 100uF in parallel with a bunch of ceramics. Use a shielded inductor, or better still a toroid - you don't want to add an antenna to your ckt. 7805 regulators are completely useless at blocking noise above 10kHz - they are aimed at mains frequencies, a simple fact which ppl often overlook.
Check your ground currents too (where they go) - bad grounding can also cause issues.

V3rge

Wed Aug 10 2011, 09:46AM

Registered Member #3253 Joined: Wed Sept 29 2010, 12:21AM
Location: New Zealand
Posts: 13

Well i've added more filtering and it seems to be alot more stable then it was under higher stresses, but it still manages to crash it depending on where it's grounding to, which I will point out below.

I have another major question or two, related to the HV side. In an ignition coil, where should the HV output be grounded to? During all this I have been grounding it to the negative side of the coil since both the primary and secondary grounds are connected and this transfers maximum power from what i've seen, and I thought would support the highest voltage. The coil primary negative is on the drain of the FET, so should I be loading the output to there or to the supply ground?
I;ve come to wonder this because tonight I was playing with it discharging sparks to metal objects sitting on the table, and the spark length was significantly longer and it was perfectly stable, and putting out MUCH higher voltages up around 30kV+ at about 10W input, although it does not transfer nearly as much power. The issue i was having using this method is that it isn't just the HV output that has such high potentials across it, small spark streams were flowing from both the 2 terminals on the coil itself towards the HV output. I turned the lights off and noticed the whole top was glowing with ionized particles, and if I pushed the power to high it was arcing from the HV output connector down to one of the primary terminals which was enough to crash the controller. Should I be insulating the primary terminals to avoid this from happening? I don't even see how the potential is building up on the terminals but the very think sparks seem to be originating from there some how.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob