Any EE PWM Motor Speed Control Geniuses Here?
erco
Posts: 20,256
A question for the illuminati here. Just received these nifty Ebay step-up voltage converters: http://www.ebay.com/itm/141054974007 3-32V in, 5-35V out. Not bad for 2.10 each. "Up to 96% conversion efficiency" 
A small trimpot sets the output voltage, fairly independent of the input voltage. Let's say we replaced that trimpot with a microprocessor-controlled digital pot and used this as a motor speed control. My favorite 24V motors are very low-current, stall at 1 A. Should work. Let's pretend this is a robotics app, differential steering, synchronizing wheel speeds by changing the drive voltage as needed.
Compared to PWMing the motors, any foreseeable advantages/disadvantages? Ideally, PWM should have more torque at lower speed by virtue of higher peak voltage, but there are a lot of high-frequency (TOO high, IMO) speed controllers that do nothing but squeal at low power settings, and send no appreciable torque output to the motor. Seems like a pure DC controller like this might have some usefulness somewhere.
I'll get to building & smoke testing it sooner or later.
A small trimpot sets the output voltage, fairly independent of the input voltage. Let's say we replaced that trimpot with a microprocessor-controlled digital pot and used this as a motor speed control. My favorite 24V motors are very low-current, stall at 1 A. Should work. Let's pretend this is a robotics app, differential steering, synchronizing wheel speeds by changing the drive voltage as needed.
Compared to PWMing the motors, any foreseeable advantages/disadvantages? Ideally, PWM should have more torque at lower speed by virtue of higher peak voltage, but there are a lot of high-frequency (TOO high, IMO) speed controllers that do nothing but squeal at low power settings, and send no appreciable torque output to the motor. Seems like a pure DC controller like this might have some usefulness somewhere.
I'll get to building & smoke testing it sooner or later.
Comments
Edit: 15 minutes nonstop running, motor mostly no load, Vout dialed up to 43.4V. The motor is barely warm, module is cooler, batteries still room temp putting out 7.3V, 530 mA no load. Under a big motor load, battery/input current increases to 1.5A. Cool beans.
Sure I'm not an illuminati nor an EE, but due to many Sitting Bull's hours at the bench and in case a heck of neck and back muscles pain do really counts for something, I believe I could try to contribute someway.
I'm only lacking of a closer front view of the five pin IC, to try decipher its brand and part #.
Or, count on your eagle eyes to look at it, and tell us what kind of beast it is.
Also two questions about the 1k trimpot:
1- Have you checked the voltage difference across its leads, along supply's full range output variation?
2- Does one of its leads( perhaps the closer to IN+, at the back view) directly connects to one of the power planes?
Any info will be of value.
Yanomani
I'd almost squeezed my eyes up to their implosion point, and all that I could find, was many shed tears.:nerd:
If you'd not given me the part number, I'm sure I didn't yet had managed to extract some meaningful characters, from these fuzzy superficial depressions.
To me, they appear meaningless, as the mumified wrinkles at Ramses's eye corners.
Looking back to Ebay's pictures and having them compared against XLSEMI's application circuit, I'd noticed some discrepancy between them:
- Picture 1 shows three assembled parts, all with 100uFx50V as output capacitors, and 220uFx35V as input ones. Picture 2 shows a single unit, displaying a pair of 220uFx35V units, at the input and output filter. Perhaps a manufacturing mistake. Must see what you will receive, when the package reach you.
I'm hoping you intend using them, with brushless motors. At least to spare board's output filter, to deal with a bunch of brush switching noise, affecting voltage regulation.
Since XL6009's internals will develop a 1.25V reference voltage accross board's trimpot leads, IMHO it's wise for you to determine the intended range of output voltages, before working out any mods.
Also I believe it's better to exert some logic control on EN (pin 2), to switch off the circuit, since weighting down the digital pot till near zero ohms will possibly destroy it, unless there is an internal limiting to the output current at FB (pin 5), not shown or stated at the datasheet.
From Xlsemi's manual, the lower limit for XL6009 output current is 100mA, to ensure feedback voltage within 1,25V +-3%. Sure, output voltage tolerance will depend on intended application, but it will add or someway depends on the way FB is controlled (stepwise, thru a digital pot), and the drift one can expect in the value of the series resistance, between FB and GND.
I didn't checked all digipot manufacturers, but there are several options, including some that could withstand higher voltages at their analog leads, and other that apparently limited that voltage, to pot's power supply values. I'm not sure, but some of then appear to be capable of managing up to 5mA thru resistor's terminals, sure taking power dissipation limits into account.
Perhaps, after deciding the intended output range for the final circuit to perform, it will be possible to substitute the fixed resistor, the one that connects VOUT to FB, with a digipot, leaving a 1% 1K fixed resistor at trimpot's place. This will limit the adjust circuit current to 1,25mA.
Hope it will help a bit
Yanomani
You dig right in! Thanks for your insightful and occasionally humorous observations, my friend. I should definitely measure the trimpot current to see what power is going through there. When I get a chance, I will let you know!