EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Chryseus on September 24, 2012, 05:23:42 pm
-
Currently I'm working on an adjustable high voltage boost converter, so far everything seems to be more or less working however I'm having issues with the output regulation, I made this design some time ago and when I tested it on a breadboard it worked fine, my friend has also built it on a breadboard without problem, I however decided to solder it down to some copper clad PCB.
I've double checked all the connections and pinouts are correct so I don't think that is the cause.
(http://electrical-workbench.co.uk/images/boostconverter-schematic.png)
* Q1 collector is now before the inductor
* R1 has been removed (the upper one not the pulldown)
The problem is when I connect the inductor the output shoots up to around 130V when I set the pot R2 to provide an output of around 50V, after about 5 seconds the voltage slowly drops down as C2 discharges through R5, once it reaches the set voltage however it goes back up to 130V.
The oscillator is a simple 555 providing a fixed 15.5kHz signal at almost 50% duty.
At the moment the only solution I can think of is increasing the load to ensure C2 discharges quickly, this is not what I really want however as I only ever intend to draw around 50mA at 250V at the absolute maximum.
:-\
-
Make C2 smaller, try 10uF to start with. Change the poor 1N4007 to a high speed diode, it is way too slow for this application. Add a 1k resistor from the base of Q1 to ground, this will help the transistors switch off a little faster, and will handle a little leakage current. Make the duty cycle from the 555 a 10% high pulse. That will definitely help. You will have to add a diode between pins 7 and 6 of the 555 to get the duty cycle under 50%, anode to pin7. This will change the frequency up somewhat.
-
Make C2 smaller, try 10uF to start with. Change the poor 1N4007 to a high speed diode, it is way too slow for this application. Add a 1k resistor from the base of Q1 to ground, this will help the transistors switch off a little faster, and will handle a little leakage current. Make the duty cycle from the 555 a 10% high pulse. That will definitely help. You will have to add a diode between pins 7 and 6 of the 555 to get the duty cycle under 50%, anode to pin7. This will change the frequency up somewhat.
Making C2 smaller is a good idea I will try that next, the 1N4007 is perfectly fine for switching at 15kHz but I will try get a faster diode, I'll add the pulldown on Q1 and see what happens although I don't expect much change, my main suspect at the moment is the op-amp, it almost seems as though it gets stuck on for a brief period.
-
You need to limit the rate of charging to enable the loop to cut off the inductor from switching. Thus a lower duty cycle which will reduce the energy transferred so that the loop can react and control it.
-
If you're going to swap-out the op amp then I'd suggest trying a comparator with a push/pull output instead. This is what you're using your op amp as now and op amps don't always make the best comparators.
-
decoupling caps needed near the opamp.
Why did you choose a 100uh inductor , seem high for this design
On the fairchild data sheet for the FET look at the page of graphs and especially the one titled". On-Resistance Variation with Gate
Voltage and Drain Current"
-
Thanks for the replies guys I'll try out some of the suggestions in the morning. :D
The 100uH inductor is just a placeholder I've got a big pile of inductors to try out.
-
Why not use a 34063 chip to drive a high voltage FET. There are plenty of circuits for this chip on the net for the 34063 which is dirt cheap. Most automobile USB plug ins use them. I have used Clare ixd604 series FET drivers to connect the 34063 to the FET. This guarantees rapid switching of the FET to keep down heating of the FET.
For your purpose you would to use the inverter version.
-
Maybe use a 3843 or similar chip to do proper PWM control?
-
As others have stated earlier, try to use a proper PWM controller IC to do the job, possibly something with "current mode control".
Also you cannot chose arbitrary inductor values and output filter capacitors. You have to properly size them based on switching frequency and output voltage ripple that is acceptable for your application. I strongly suggest to use a power MOSFET instead of BJTs.
-
I know a MOSFET or IGBT would be more efficient however I don't have the cash to spare at the moment to get one, I do however have some supplies laying around so I'll try dig one up.
I know I could use a proper PWM IC to do this but I'm trying to make this as minimalistic as possible, high performance is not the goal of this project.
I'm building a simple adjustable frequency and duty signal generator at the moment so I can tweak it for optimum performance, using a different inductor has seemed to cure my original problem.
My friend using the same design has managed to get 520V out from 12V at low load current, which despite the simplicity I find quite impressive.
Update:
Found a MOSFET, removed Q1 and placed a pulldown on the gate of Q2.
Works very good now, tested it with a crappy 10kHz signal generator at 80% duty, puts out 260V.