Was not the first result with Google, but looks like you mean this one:
But it doesn't look good. The main loop and ADC runs with 4.9 kHz and the PWM with 39 KHz. Modern switching regulators uses hundreds of KHz, which results in smaller coils, and a cycle-by-cycle control loop, which improves the output ripple and stability, and with integrated current limit etc.
-- Frank Buss, fb@frank-buss.de http://www.frank-buss.de, http://www.it4-systems.de
On a sunny day (Sun, 01 Feb 2009 19:43:32 -0500) it happened Hammy wrote in :
Yes that would work, those comparators have some restrictions though, the programmable reference is in big steps for example. That would be OK for a current limit, but a big course for voltage adjustment, assuming I also want to be able to control the voltage via RS232. Else some external resistors would fix that.
On a sunny day (Mon, 2 Feb 2009 12:41:53 +0100) it happened Frank Buss wrote in :
Yes that is the one.
The current limit is easy to add, but takes some components. Anyways, I will play with this, see what I can do.
Yes both of those FET's are good for low to medium power applications up to the 50V plus range. There are even better performance ones but for the price I can't find any better. Delay and rise/fall times combined less then 70nS typical; and single digit gate charge Qt at
5Vgs.Another inexpensive switch is Onsemis low Vce sat NJT4030PT1G and complement NJT4031NT1G. I got these specifically for using for direct drive off PIC's (1 to 2 amp area) ;30 to 40 cents cut tape. You have to use a speed up cap to bypass the base resistor but they are good for at least a couple hundred kHz switching. They are also in the
4-SSOT-223 Package so they can at least dissipate more then a couple hundred mW depending on the copper area for a heatsink, and / or use one of the smd heatsinks.
I've considered the possibility of using a uc for DC/DC controller. Mainly though I think it would only be worth while if you could use an existing PIC in the design this would depend on what the PIC's doing though; If it could handle being pulled away constantly to adjust the output voltage. If it could be done using a low cost 10F or 12F PIC (the 10f222 has ADC) might also make it a realistic option.
Well a lot of controllers are getting extremely pricey LT is a prime example who the hell uses 5 to 13 dollar controllers? Onsemi , TI and National still make some reasonably priced VMC and CMC controllers NCP series from Onsemi and the TPS series from TI these req minimal external components.
To be comparative cost wise as an example if you do build say a HYSTERETIC converter with a PIC these are the cheapest controllers a NCP3063 which is an upgraded 34063 could be had for 50 cents cut tape national makes similar ones they basically just need a cap diode and inductor for upto about 1.2A output.
If you call 250 uA "on".
Best regards, Spehro Pefhany
-- "it\'s the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com
On a sunny day (Mon, 02 Feb 2009 08:50:46 -0500) it happened Spehro Pefhany wrote in
Oops, my error, yes, 4 V needed for about 15A,
I stay with my IRLZ34N. It can do a lot of current 30A, 55V. 35 mOhm.
20A at 4 V gate. 6A at 3 V gate, I will be using < 3 A here. It has an avalanche protection zener diode with 16A rating build in. And I have some :-)
I don't think anyone would want a FET that fully enhances at a typical
1Vgs. This would mean the gate oxide could easily be punctured. You would also have to bring the gate below the source to ensure it is fully off.They should put Vth at the 10% rated current. It's just good for quick comparison between Fets as long as you realize it's just defined for a small current.
Well it depends on what your using it for. I would choose the ST one if it met all the req before the IRF one it's ¼ to 1/5 the price (for the same DPAK ).The ST one is much faster and has reduced drive req.
The only plus with the IRF Fet is it's lower RDSON which is to be expected larger fet. But any efficiency gains in conduction losses may just be lost in switching losses. If I didn't need the extra power handling of the TO-220 package I would use the ST part.
On a sunny day (Mon, 02 Feb 2009 10:43:47 -0500) it happened Hammy wrote in :
Yes, for a larger production those MOSFETS would be a better choice. I am just collection some stuff to make prototype, found some nice ring cores, have all the parts now.
What I will do is make an RS232 controlled PIC that just sets the PWM, so without the feedback, to see what happens (already did run some Ltspice simulations, really simple circuit), with these inductors. So that should control Vout, then I can think about I limit, found a small ring core I can use a current transformer in the source, that should give enough voltage to trigger one of those PIC comparators. When short circuit tests are passed I will think about the feedback PID loop.
I was referring to your 1.4MHz case :-)
Most TVs were pretty well muffled in that respect. I have here a Dell printer (that always keeps drying up so I only use its scanner as backup) and the power supply runs at an audible frequency. It's terrible.
-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
Yes, the PSoC that Klaus suggested would IMHO be the best avenue if it absolutely has to live inside a uC. In essence that would make it analog again but nobody would see that :-)
Another option might be to use an input and invert that straight onto another output port pin. No idea if the PIC can do that and whether it has Schmitt inputs, but that would allow to build a nice PWM control around a Schmitt inverter. Should be very feasible with FPGA, but only if it's needed for other purposes already because those are pricey.
-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
The Rds(on) table entries under the "max" column on top of page 4/14 are the only values you can rely on for industrial designs. Anything else is like Russian roulette.
[...]-- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.