Electronics > Beginners

MC34063 high voltage dc-dc boost converter

<< < (18/19) > >>

MagicSmoker:

--- Quote from: dazz on January 27, 2020, 06:30:14 pm ---
--- Quote from: MagicSmoker on January 27, 2020, 06:22:33 pm ---Basically, any switchmode converter in which power is supplied to the load during the switch off time is going to be unstable without a load, and the boost converter does just that, same as the flyback.

--- End quote ---

It's the sudden current spike at the inductor when the magnetic field collapses that does this, right? I noticed the effect got progressively less noticeable as I decreased the inductance of the coil. I'm at 33uH right now.
--- End quote ---

Kinda/sorta/almost. The nature of an inductor is to resist a change in current, right? When the switch turns on in a boost (or flyback) converter current builds up (relatively) slowly and linearly, and when the switch turns off the current that was flowing in the inductor wants to keep flowing, creating whatever voltage is necessary to maintain said flow. Basically, when the switch turns on the boost inductor acts like a current sink and when the switch turns off it acts like a current source.


--- Quote from: dazz on January 27, 2020, 06:30:14 pm ---Building a constant current load with a mosfet now to test it under load. What kind of load current should I expect to make it stable?
--- End quote ---


Usually somewhere between 1% and 10% of full load power is necessary to maintain regulation. I don't know how your ratio-extender kludge with the '34063 is going to behave in real life, however; it certainly didn't behave well in the LTspice simulation...

I would also recommend against a beginner using a MOSFET current sink (load) to test switchmode converters because that pits two control loops against each other (the current regulation loop on the constant current load and the voltage regulation loop of the switcher), however given the '34063 is a hysteretic controller there is less risk of inciting full blown oscillations here (with the inevitable release of magic smoke soon thereafter).

dazz:

--- Quote from: MagicSmoker on January 27, 2020, 08:36:15 pm ---Usually somewhere between 1% and 10% of full load power is necessary to maintain regulation. I don't know how your ratio-extender kludge with the '34063 is going to behave in real life, however; it certainly didn't behave well in the LTspice simulation...

--- End quote ---

What's wrong with it, please? I thought it was performing pretty well in the sim with a very respectable (I think) 85% efficiency


--- Quote from: MagicSmoker on January 27, 2020, 08:36:15 pm ---I would also recommend against a beginner using a MOSFET current sink (load) to test switchmode converters because that pits two control loops against each other (the current regulation loop on the constant current load and the voltage regulation loop of the switcher), however given the '34063 is a hysteretic controller there is less risk of inciting full blown oscillations here (with the inevitable release of magic smoke soon thereafter).

--- End quote ---

That might explain why I can't seem to get it to work with the boost converter. This is the one I'm trying:

MagicSmoker:

--- Quote from: dazz on January 27, 2020, 09:52:07 pm ---
--- Quote from: MagicSmoker on January 27, 2020, 08:36:15 pm ---...I don't know how your ratio-extender kludge with the '34063 is going to behave in real life, however; it certainly didn't behave well in the LTspice simulation...

--- End quote ---

What's wrong with it, please? I thought it was performing pretty well in the sim with a very respectable (I think) 85% efficiency
--- End quote ---

It made the '34063 model run in bursts with wildly varying pulse widths from cycle to cycle, some of which were way too long (ie - allowing current to build up way too high in the boost inductor). That's a good way to blow up FETs. That said, this could be the result of a dodgy '34063 model. I didn't investigate the issue too deeply (not at all, actually...).


--- Quote from: dazz on January 27, 2020, 09:52:07 pm ---That might explain why I can't seem to get it to work with the boost converter. This is the one I'm trying:
~snip
--- End quote ---

That circuit is almost guaranteed to oscillate because the op-amp is saddled with driving a capacitive load directly (the gate of a MOSFET is, essentially, a capacitor). At the very minimum there needs to be a resistor in between the output of the '358 and the MOSFET gate (take feedback directly from the op-amp output, however); 100R should work fine here. Also use a FET with an appropriate voltage rating for the task (the IRFZ44 is rated for 55-60V max).

That said, it's hard to argue with the cheap-n-cheerful approach of just tacking on a bunch of 330k/0.5W metal film resistors in parallel as needed to provide about 1mA of load each at 300V.

dazz:
Thanks again, MagicSmoker. I'll try that resistor in series with the gate. It's the same principle as a grid stopper, right? I have a proper mosfet for the job there, an IRF740 that can handle up to 400V.
I removed the 100uF input cap of the converter and that further improved the stability of the circuit. Under a 11mA load, with 111V at the output through a 20W 10K resistor, it is stable, but it's not working properly: with 12V at the input, the current draw there is 730mA. So that's 1.2W at the output and almost 9W drawn at the input. Horrible. Vref is stuck at 1.34V which makes no sense to me. Looks like the MC34063 never switches off. I believe most of the wasted power is being dissipated at the 50uH inductor I have in there right now, because it gets very hot while the mosfet runs cool as a cucumber.

The inductor is a toroidal one, about .77 inches of diameter and AWG 20 wire (or something like that)

This thing is starting to look like a lost cause.

VEGETA:
I was thinking about making an isolated DC-DC converter to get 6-8.4v from li-ion pack and convert them to 12v DC at say 200mA max. I found this IC MC34063 to be cheaper than LT8300 and I found this little SMD transformer: https://lcsc.com/product-detail/Power-Transformer_TDK-ATB322515-0110-T000_C415291.html

I wonder if this minimal circuit works with these parts?

I don't need fancy stuff, just continuous 12v isolated supply and I want it to be less than read-to-use modules like the 0505 or 1212 ones. These ones will require more circuitry to fit my application such as boost or buck to make the signal suitable for each module.

Looking forward to your feedback!

Note: I couldn't find an ltspice model for MC34063, where and how can I make it work?

thanks!

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod