|
Post by lamazoid on Feb 3, 2017 13:15:19 GMT -8
Hi everyone =) Recently i've stuck with searching a simple, cheap low-voltage PWM controller IC. My requirements were quite clear (and seem to be easy to meet) - operation voltage range (at least) 4-12V, low-side MOSFET driver with (at least) 1A output, adjustable frequency, current-mode control. It's hard to believe, but almost ALL the IC manufacturers have a BIG TROUBLE making that kind of device! There is a good and cheap old IC, which meets almost all requirements except low voltage operation - that is UC3845. it is also very cheap, about 10 cents. There is a lower voltage CMOS version - UCC3805, produced by Texas Instruments/Unitrode. But it's surprisingly expensive, about 10 times (i cant understand why) and seems to be hard to obtain. There's also LM3478, which is almost ideal for described requirements, with except of high-frequency operation (not less than 100 KHz recommended). That's everything what i have found. So i create this thread for other possible finds/additions, which, as usual, could happen unexpectively
|
|
|
Post by Gary Lecomte on Feb 3, 2017 15:14:10 GMT -8
Not sure what you mean by Current-Mode control?
|
|
|
Post by lamazoid on Feb 4, 2017 0:07:33 GMT -8
Not sure what you mean by Current-Mode control? Just like in UC38xx family: there is a current sense amplifier (comparator) input, which usually connects to "sense" resistor installed between power MOFET source and circuit ground. Switching current at this point is represented by sawtooth ramp of certain amplitude. When current amplitude crosses comparator threshold, PWM signal is terminated until the end of period. This allows to keep stored energy (per cycle) in power inductor at constant, wheh supply voltage changes.
|
|
|
Post by Gary Lecomte on Feb 4, 2017 7:28:05 GMT -8
I Built up this circuit Many years ago
|
|
|
Post by lamazoid on Feb 4, 2017 8:03:52 GMT -8
wow, looks like an application "inside-out": current sense input (pin 3) used for voltage control??? while pin 2, which is dedicated for voltage feedback, used for... (i dont understand for why)
|
|
|
Post by Gary Lecomte on Feb 4, 2017 8:39:55 GMT -8
Yes maybe somewhat Crazy, But it works! Attachments:
|
|
|
Post by lamazoid on Feb 4, 2017 23:42:34 GMT -8
looks cool, but getting back to main question: do you know any of UC38xx alternatives which could run from 5v or less? The only thing i have found is LM3478
|
|
|
Post by Gary Lecomte on Feb 5, 2017 10:11:25 GMT -8
looks cool, but getting back to main question: do you know any of UC38xx alternatives which could run from 5v or less? The only thing i have found is LM3478 SORRY NO.
|
|
|
Post by lamazoid on Mar 2, 2017 7:00:28 GMT -8
Found another one, SIP2800 family, produced by Vishay Siliconix. Seems to be direct replacement of UCC3805 and even pin-compatible. but there are no russian suppliers who has stock of this IC here's what Mouser says about it: mouser.com/ProductDetail/Vishay-Siliconix/SIP2803DY-T1/
|
|
|
Post by Gary Lecomte on Mar 2, 2017 8:09:55 GMT -8
Looks like it is Obsolete.
|
|
|
Post by retiredat55 on Mar 2, 2017 11:35:46 GMT -8
Maybe they still can order them.
But max output current is only 50 mA.
So it has not : "driver with (at least) 1A output".
|
|
|
Post by lamazoid on Mar 3, 2017 3:33:55 GMT -8
Maybe they still can order them. But max output current is only 50 mA. So it has not : "driver with (at least) 1A output". yep somewhat strange datasheet... in other place, it says that output drive current is 800 mA. and another strange thing about oscillation frequency, it is specified as fixed. I'm starting to think about implementing ICs functionality on discrete logic or MCU. a combination of MCU+Gate driver+LDO regulator costs two times less than UCC3805...
|
|
|
Post by retiredat55 on Mar 3, 2017 6:05:27 GMT -8
Good point!
|
|
|
Post by lamazoid on Mar 18, 2017 5:26:33 GMT -8
found another one: LT3750. but very expensive and seems to quit production soon. there are some interesting features in that IC, like sensing secondary voltage from PRIMARY side (e.g. no voltage feedback needed). datasheet explains how that works. good start to develop custom PWM...
|
|
|
Post by lamazoid on Apr 5, 2017 10:15:07 GMT -8
Another crazy (actually not) idea: put a service step-up converter before UC3845, to bypass UVLO threshold. I've found some cheapr IC's to implement this, 34063 for example (and it's modern version NCP3063 from ON semiconductor). This should (will) work from as low as 3V supply. Total BOM cost it 2-3 times less than with UCC3805.
Actually i have already tried this approach, with ucc3805 & NCP1402A to run an HV converter off a supercap (2.7V fully charged). I have not measured overall efficiency, but it worked down to 0.8V !
|
|