Sharpen up my PWM, 4093.

Discussion in 'The Projects Forum' started by HellTriX, Aug 2, 2008.

  1. HellTriX

    Thread Starter Active Member

    Apr 11, 2008
    83
    1
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    That's the charge on the gate of the MOSFET. Which MOSFET did you use?

    You could use the three unused gates to beef up the drive. Connect the inputs of the three gates in parallel to the output of the gate that's being used as an oscillator. Connect the three outputs together, and feed them to the gate of the MOSFET.

    You could also use a voltage follower.
     
    Last edited: Aug 2, 2008
  3. HellTriX

    Thread Starter Active Member

    Apr 11, 2008
    83
    1
    I actually didn't drive a mosfet yet. I tried driving a capacitor and then a resistor, then both. I just added a 1uf (I think) capacitor between the vdd pin and ground pin and that squared off the signal nicely. But it boosted the frequency from 1.25khz to 2.15khz. (I ran out of .1uf)


    Gonna look through my mosfet junk box to see what I can find.
    I'm gonna use a mosfet to drive this 12va step down transformer.
    The mosfet will switch 120vdc and drive the 120v side of the transformer at around 1-5khz, giving me ~16vac. I will then rectify and regulate that to drive the rest of my circuity and driver chips for my main circuity.

    Right now I'm using a 12.7v zener and 2.2kohm 22 watt resistor to power the 4093. Silly way to make a powersupply maybe? hehe
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Ahh, OK. Well, it should run at reasonably close to the advertised frequency if your components aren't too far off. Lots of older ceramic caps had really lousy tolerances. Also, if you're dealing with new old stock stuff, they may have drifted quite a bit over time. You can deal with it if you measure things beforehand, but if you don't have the testers... (cap meter, for example)

    That's pretty whacky alright! :confused:

    This 12VA transformer - was it designed for 50/60Hz? If so, your performance is going to stink when you try to run a couple of kHz signal through it. If you try to run it at 50/60Hz, you'll probably get better performance, but since your rise/fall times are so much faster, you'll get some REALLY LARGE reverse-EMF pulses from it - you may very well exceed the breakdown voltage of the insulation.

    I'll wager that the first pulse the MOSFET puts through that transformer will be it's last. The transformer's reverse EMF will smite it. :eek:
     
  5. HellTriX

    Thread Starter Active Member

    Apr 11, 2008
    83
    1
    lol. I planned to use a diode to protect the mosfet.
    And I was under the impression that a 50-60hz transformer could work efficiently to 400hz. After that performance would drop off. The higher the frequency the better current you can get through em up to a point you mentioned. I'm gonna see how low I can get this lil 4093 frequency down to.

    I was going to do an H bridge to drive the transformer, and figured why bother.

    I'm not skilled enough to design a good power supply that runs from 120vdc and converts to 12vdc @ 2.5 amps. Best I could do was a sort of switch mode supply of my own design, and to be honest I haven't looked at any switch mode power supply schematics before undertaking my own. All the circuity I have operates at max of 38vdc except the mosfets. So figured I would just make a zener voltage drop to run the mosfet driver and push that through a transformer.
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Take a look at "buck" topology DC-DC converters.

    National Semiconductor and other manufacturers have wizards on their sites to get you up and going in a hurry. You can get very high efficiency using some of those designs.
     
  7. HellTriX

    Thread Starter Active Member

    Apr 11, 2008
    83
    1

    Oh trust me, I've looked everywhere. I find tons of buck converters and many with the current rating I need. But None, not a single one can handle a 120-156v input. 90% of them are below 60vdc.
     
  8. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
  9. HellTriX

    Thread Starter Active Member

    Apr 11, 2008
    83
    1
    I'm thinking the power loss in mine will be much lower :)
    Gonna keep increasing the resistor for the zener to see how little power dissipation I can do. That will run a nice mosfet/transformer and somewhere above 60% efficiency. I think my entire drive circuity might use less then 1 amp at 12.7v but I'm over engineering it just in case. Can always downsize the design later.
     
  10. HellTriX

    Thread Starter Active Member

    Apr 11, 2008
    83
    1
    Used a bts141 mosfet I found in the junk box. Hooked it upto the 4093, then wired up the transformer between my 20v and drain. Got .9 volts from the secondarys at about 50% duty cycle running at 4.76khz. Was too tired to switch the capacitor out to lower the frequency. I think I might have a headache from that frequency resonating out of the mosfet and xformer.

    I will measure efficiency of the circuit at different frequencies tomorrow. The oscilloscope showed a pretty clean signal at the transformer. No sign of big emf spikes.

    I'll draw up a schematic soon of my power supply soon as I refine it a bit more so people can see it and critique it.
     
  11. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Running that fast, the 4093 is going to have a tough time charging/discharging the gate. That's why I set it up for 1kHz to begin with. It can only output a few mA's. If you're running it slow, the transition time vs ON/OFF time is very significantly reduced, so less time for the MOSFET to get warm.

    If you wanted to experiment, you could make a toroidal transformer. You can run a small toroidal transformer at much higher frequencies than your traditional laminated plate transformers.

    Basically, the higher in frequency you go, the smaller the transformer you need.

    You're going to need some kind of regulation of your output. One way you can do that is to increase the frequency of the oscillator when the voltage is high enough (which decreases the efficiency of the transformer) and decrease the frequency when the output voltage is too low, thus increasing the efficiency of the transformer.

    Take a look at Ronald Dekker's "Flyback Converters for Dummies" page:
    http://www.dos4ever.com/flyback/flyback.html
    Not really for dummies - but there's a LOT of great info on making your own flyback supplies, test bench for coils you wind, etc.
     
  12. HellTriX

    Thread Starter Active Member

    Apr 11, 2008
    83
    1
    Nice thanks for the info.
    I did a 120.6 volt test today.

    Using a 120kohm resistor and a 12volt zener I got 16.7volts to the input of the 4093 (sounds a bit high for a 12volt zener +- 5% ?).

    I also dropped the frequency to a perfect 50.0 hz. I might tweak it a bit for 60hz, but not sure if that matters as the xformer is rated at 50-60hz.

    but the test was to make sure I could drive the 4093 from a resistor/zener. I had to disconnect my mosfet as its only rated at 60v and swap in one of my 200v fets. Then test how much power I can suck out of the transformer.

    More tomorrow.
     
  13. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Interesting. With a 120k resistor across 103.9v, you'd have less than 1mA current flowing. I would've thought you would measure much lower than that. The kind you get at RS usually like to have 10mA or so through them before they're regulating fairly stable; if you feed them less, the voltage across them decreases.

    If you're somewhere between 50-60 Hz, that's close enough.

    Have fun :)
     
  14. HellTriX

    Thread Starter Active Member

    Apr 11, 2008
    83
    1
    Mmk I have a correction. The 12v zener was providing 7.3 volts as it helps to measure DC with the DC setting rather then AC measurement :rolleyes:
    I also added a mosfet and drove the transformer with 120volts switched through the fet at 50hz with 50% duty cycle and measured 5.4 volts. It should be higher but I'm thinking that I need to drop the 120k resistor to the Zener to increase the gate voltage closer to 10-12volts to make the mosfet turn on harder.
    Also need to clean up the power lines with some capacitors. Only spent a few minutes on it today as I have too much homework right now. Still haven't blown any parts up or fried myself with 120volts which is almos surprizing if you seen my proto board. The kind where you gotta take the needle nose pliers around to bend components around so they are not shorting each other :p So I usely do 2-3 inspections and safty checks before hooking up power.

    I guess I have too much fun with this stuff.
     
Loading...