Am I going crazy???

[rankot]

I'm currently experimenting with DC boost with NE555. Simple schematic, like the one here:


I wanted to see what is the difference with various frequencies and on/off rates while switching. At first I tested it with R1=33k and R2=1k, C2=100p. It worked well, and the switching frequency was as expected. Frequency 400kHz and duty cycle 97%. But my MOSFET was hot, so I wanted to reduce frequency and duty cycle, and put R1=2k and R2=68k, leaving C2 the same. That should give me 104kHz with 50% duty cycle, but I get some 2Mz(!!!) on my scope and the damn thing isn't working - output voltage is just 9V. It happened last night, and I went to sleep. Now turned it on, everything works as expected, MOSFET is cool, and output voltage just as it should be, with 104kHz frequency on TLC555 output.

WTF???

[antonis]

Can't see neither any Gate drive resistor nor turn-off accelaration configuration (resistor in series with diode).. 

[Phend]

Heatsink ??

[rankot]

Actualy, my schematic is a little bit different, I just didn't have time to make picture on previous post, so used one similar from the web. This is my circuit:



I suppose this is what you have in mind with acceleration (R20, D9, D36 and BC327), or not?

Maybe I have a cold joint somewhere, and moving the PCB on my desk just made better contact, which could explain late night's behavior. Will have to check...

[rankot]

Heatsink ??

I tried without heatsink, because I need only 25-27V in this case. I tried with IRF740 for the first time (which was hot, but most probably due to too high duty cycle), then used IRF3205 - it can handle the voltage I need and Rds(on) is 0.008Ohm, which is negligible compared to IRF740's 0.55Ohm. Now it's running for an hour and a half, voltage is stable and MOSFET is completely cool. No heatsink. I will also try to run this using 2N7000, because it can handle the needed power and using TO-220 can be bulky in a pedal.

[antonis]

I suppose this is what you have in mind with acceleration (R20, D9, D36 and BC327), or not?

Yeapp..
(+1 for improved version..)

[antonis]

I will also try to run this using 2N7000, because it can handle the needed power and using TO-220 can be bulky in a pedal.

I wouldn't trust a TO-92 package inside a closed box..
IRF3205 (TO-220AB) has R_θJA about 60^oC/W where 2N7000 (TO-92) has 312^oC/W..

[Rob Strand]

The 2N7000 version might take more work.

The funny thing is the 2N7000 is a smaller device and it might turn out increasing the frequency helps.  If the inductor is too small and/or the frequency too low the inductor current could ramp up and peak to a value not favorable to the 2N7000.

Measuring the DC input current with a fixed load is a good way judge how much loss the circuit has and that generally means the components are less stressed.


There's also a point where charge-pump circuits make more sense.

I tried with IRF740 for the first time (which was hot, but most probably due to too high duty cycle), then used IRF3205 - it can handle the voltage I need and Rds(on) is 0.008Ohm, which is negligible compared to IRF740's 0.55Ohm. Now it's running for an hour and a half, voltage is stable and MOSFET is completely cool.

Compare that to say 1 to 2 ohms (5 ohms max) for the 2N7000 and you can see how things change when you have high peak currents.