On Sat, Nov 29, 2014, at 11:37 AM, Kirk Wallace wrote: > > Thank you for the book recommendation. I have found a lot of good > information trolling through manufacturer's documents and various > websites.
I second that - data sheets and application notes can tell you a lot. > This started as just trying to scratch an itch -- bare bones > FET and motor, but I tend to get drawn in while making progress. Sounds like you are having fun. This kind of thing is my day job, I'll be happy to pitch in to the thread. > > I assume you already know why you need an isolated > > power supply to the high side driver? > > Oh yes, but I kind of resent having to wiggle the high side to keep the > driver charged. So you are thinking of the "bootstrap" type high-side supply? The one where a cap is charged thru a diode when the low side switch is on, and then the cap supplies power for the high side driver (for a while). > I'm tending to think, since a tiny AVR is so cheap, I > could add one to maintain the high side supply, and maybe do other > things since it is there. I like the idea of being able to go right up > to 100% duty without giving it a thought. Do you mean that the AVR would generate short pulses on the low side switch to keep the high side cap charged? That might allow you to command 100% without thinking about it, but it doesn't mean you actually get 100%. If you really want 100% duty cycle you need an independent floating supply for the high side. On most motor drives there are extra windings on the main power supply transformer. That supply is usually a flyback running from the DC bus, so extra windings have reasonably good regulation with just a simple diode/cap rectifier. If I was messing around at home, I'd probably buy some of these little isolated DC-DC converters: http://www.digikey.com/product-detail/en/CME0512S3C/811-2892-5-ND/4693720 At $3.81 each, I'd be tempted to get three for an H-bridge. One for each high side switch, and one to power both low side switches. Combined with 3120 drivers, you would have complete galvanic isolation between the power stage and the control. That would greatly reduce noise problems and even if you start raising the power level the risk of damaging anything on the control side would be much much lower. > Another thing the AVR could do is monitor the current sensor. What did you have in mind for a current sensor? I'm not fond of shunt resistors - again, I like galvanic isolation, and that is hard to do with a shunt. And as soon as you start doing an H-bridge you really want the shunt in series with the motor, which means you need to measure the small shunt voltage while it is jumping up and down many volts at the PWM frequency. That requires either an isolation amp or a differential amp with phenomenal common mode rejection at very high frequencies. Allegro makes some isolated current sensors that are pretty cheap ($2.70): http://www.digikey.com/product-detail/en/ACS711ELCTR-25AB-T/620-1371-1-ND/2470595 I've never used them, their voltage isolation isn't appropriate for a 480V powered industrial drive. But they look promising for lower voltage. A really nice sensor for this kind of thing is the LEM LA25NP. Datasheet here: http://www.lem.com/docs/products/la%2025-np.pdf They aren't cheap - $25 at Digikey http://www.digikey.com/product-detail/en/LA%2025-NP/LA%2025-NP-ND/513153 But if you just want one to play with send me your address off-list and I'll send you one (I have a few surplus ones sitting in a box in the basement). One annoyance of the LEM is that it wants a split supply, +/-15 volts. The Allegro sensor will run on a single 3.3 or 5V supply. > It seems > that monitoring current is important for driver protection as well as > possibly motor protection and torque control -- but, what current? Peak > current seems important to prevent blowing out components. Average > current is needed for motor load and driver over-heating. I guess I'll > find out as I go. I typically like to have instantaneous over-current protection that acts in microseconds in case of a short circuit. That is typically a comparator and a latch, no A/D converter or software of any kind. Average current for thermal protection or simply closing a current control loop can be slower and would be where an A/D converter comes in. -- John Kasunich [email protected] ------------------------------------------------------------------------------ Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server from Actuate! Instantly Supercharge Your Business Reports and Dashboards with Interactivity, Sharing, Native Excel Exports, App Integration & more Get technology previously reserved for billion-dollar corporations, FREE http://pubads.g.doubleclick.net/gampad/clk?id=157005751&iu=/4140/ostg.clktrk _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
