PIC power supply problems

[evandude]

I have a small circuit built where a PIC uses MOSFETs to switch a couple of solenoids. the solenoids use quite a bit of power (one is a 3 watt/6 volt, running it at 9v though... the other is much larger, no rating though)

I have the mosfets switching the solenoids directly from the 9v battery that acts as a power source. I have a 6800uF 16v capacitor in parallel with the battery, which is as large a cap as i can physically fit into my project... The PIC and all control electronics are supplied by a 78L05 voltage regulator...

I had some problems earlier, where the PIC would lock up on power-up, i found that it was browning out because of how long that large cap took to charge on power-up. so i enabled brown-out reset, and that took care of that.

however, now every time the PIC trips the solenoids, it is resetting itself. I know this is because the solenoids are using so much power, it is dropping the voltage levels. I added a 330uF cap in parallel with the OUTPUT of the voltage regulator, to help stabilize just the logic power levels... but this hasn't helped... I even tried upping it to 3300uF, and that still didn't cut it. i know there is no way that the PIC alone is using enough power to drain a 330uF (much less a 3300uF) decoupling capacitor... this is also especially odd considering the original board only had a 100uF decoupling cap after the regulator.

my question is: can current flow BACKWARD through a 78L05? that seems to be the only way that the voltage could be dropping that badly even with such a large decoupling cap after the regulator... I compared this board to the original one that it replaced, and there looks to be a diode between the +9v and the input of the voltage regulator. I didn't put one in when i designed it, because i never thought there would be any danger of current flowing backward through the regulator... All i have lying around are 1N914's, but the diode on the original board looks much beefier, probably rated at several amps... I don't imagine that the PIC would need that much power... but i'm also a little leery of sticking a tiny glass diode in there and running the risk of having it blow up in my face...

any thoughts?

thanks,

-evandude

 

[Sebi]

My opinion: the 9V supply (battery) can't produced enough power for solenoids. The 9V dropped under 7V and the 78L05 cannot stabilizing. :cry:

 

[evandude]

yeah, that's what I had figured... however since the only thing running directly on 9v is the solenoids, i don't care too much if that voltage drops. but i need to keep the output of the regulator at close to 5v. if the regulator can't stabilize because the input is too low, then the decoupling cap on the output should be more than enough to keep the output at close to 5v until the input can get back to a high enough voltage.
it's just that that's not happening, so i think current is flowing back from the output decoupling cap, through the regulator, and to the solenoid, which is definitely not OK.

 

[bmcculla]

How long is the solenoid activated? Even the 3300uF cap can only keep the PIC running for less than a second:

Assume 25mA current into PIC.

Q=VC Q= (5V-Vminpic)*3300uF I=dQ/dt

if the PIC will run at 2.7V then Q = .0076
So the cap can supply less than 7.5mA over 1 second before the PIC resets. Since the current draw is much higher the cap will discharge in less than one second.

If you put the PIC into power saving mode right after activating the solenoid you might be able to squeeze some more time out of the Cap.

Hope this helps
Brent

 

[evandude]

the solenoids are only getting activated for something on the order of 15 or 20 milliseconds each.

the thing is, that the original board uses the SAME capacitor across the 9v lines, and then has just a 100uF cap across the 5v line from the regulator.

the strangest part is that this board worked just fine up until just the other day, i'd tested it out a lot, and it never had this resetting problem...

i just ran out and picked up some 1 amp rectifier diodes at radio shack, i'm going to stick one between the 9v and regulator input and see how that works.

 

[evandude]

alright, i got that diode in there... it helped, it wasn't doing it quite as bad... messed around with everything for a while, don't remember specifically what i did, but it works now, so i'm not going to touch it :wink: