On Saturday 27 September 2008, Jon Elson wrote: >Gene Heskett wrote: >> On Saturday 27 September 2008, Jon Elson wrote: >>> Gene Heskett wrote: >>>> Maybe I'm being cautious, but if I hooked my scope to the FET's gate, >>>> and saw transition times noticeably slower than my 100mhz dual trace's >>>> own rise times for it to make that about 20 volt swing, then I would be >>>> looking for ways to speed up the driver. Even 30ns rise & fall times >>>> will be reflected in the increased temps of the FET's as the current >>>> rises toward the max ratings. >>> >>> Well, generally, you can't go that fast in full-bridge switching >>> applications. When feeding current to an inductive load, there is one >>> side of the bridge sourcing current. When you switch that side to >>> ground, the inductive effect draws current, and the common terminal >>> between the high-side and low-side transistors tends to go below ground >>> until the low-side transistor turns on completely. Most dual FET driver >>> chips can only handle a limited amount of negative voltage there. As >>> the Di/Dt rises, the voltages generated across even short, wide circuit >>> traces becomes enormous, it is easy to see 20 V across a 2" trace. >>> Circuit board layout only helps up to a point. When you reach the >>> limit, slowing down the transistors is the last fix. With relatively >>> slow PWM frequencies, a couple hundred ns rise and fall will not cause >>> excessive losses. When you get to the 100 KHz and up range, then you >>> have to work harder. >>> >>> Jon >> >> Are you referring to FET's with free wheeling diodes, or the more common >> integrated H bridges (DMOS) like the xylotex & its A3977's? > >It may not matter all that much. In the trade mages, I've seen what the >internal bussing in the power bricks look like, and they are not going >to be much lower inductance and loop area than my boards, for instance. > >> I'll have to go look at the outputs of my xylotex again when I get a >> chance, but ISTR it was quite a bit faster than 200 ns. The undershoot I >> could see, but it was IIRC in the 8 or 9 ns region and about the limit of >> my scope to define (its only a 100mhz scope, with usable response to >> around 200), and only went down 1 Vbe below ground. The clipping caused >> by substrate diode conduction was pretty obvious to the trained eye, but >> looked to be relatively 'clean'. I'm not equipt to measure currents with >> the scope so I've NDI of the actual magnitudes involved. > >Well, the typical FET drivers can source/sink a couple amps, max, and >with the larger FETS the G-S capacitance plus the miller effect really >puts a limit on the gate charge/discharge rate. Of course, it gets >worse at higher voltage, and then the inductive effect makes it still >worse at higher current. > >> I believe the A3977 stages the switching to prevent any transition shorts, >> but don't recall being able to interpret the waveforms and confirm that >> without going to a resistive load. Not having a 15 ohm resistor of >> sufficient power rating, I hadn't even considered looking at it that way. >> Jeff I'm sure would frown on doing so. > >On my drives, I have to tune the gate resistors for different transistor >types to prevent shoot-through. >It doesn't blow anything, but causes a LOT of electrical noise plus >heating. I changed transistors after having trouble getting one type, >and had to retune for the newer part number. > >> I blew the single chopper FET in the spindle motor of my little HF >> micromill, and when I looked up the number, I wasn't terribly impressed. >> 7 amps, 400 volts ISTR, in a TO-247 package yet! That is equivalent to >> floor sweepings. No such FET's are available at the local RS, so I opened >> up a couple of old computer psu's to see what might be available there. I >> picked one of those rated at about 2x the current and 2.5x the voltage & >> stuck it in. The heat sink runs a little cooler I think, and I have >> stalled the spindle for as much as 2 seconds several times since and it >> seems to shrug off such abuse without any ill effects. That will clear >> the fuse in that 2 seconds or less, but its only 2 amps anyway. > >Interesting. I had intermittent problems in that circuit on mine, >turned out to be a badly soldered resistor. I have put a DPDT switch on >mine so I can run the motor off a servo amp. I can't get full RPM, as >my servo box has a 50 V power supply, but the servo amp gets me >computer-controlled variable speed and reversing.
I'm running mine with a PMDX-106 interface, which gives me full control from its panel and from emc2.x >I have an encoder >patched on to the spindle, and can do rigid tapping, which is a hoot! >So, it can do something my Bridgeport can't do! If only that spindle had the power to do that, but sadly, its about tapped out with a 1/4" 28 tpi tap in cold margarine. Ditto for the spindle motor in my 7x12. I have cut threads with it using single tooth carbide bits, but they are a bear to sharpen correctly. That is how I cut the threads for the 1/2" acme bronze nuts in the new z axis of that HF mill. How did you attach the encoder to the micromill spindle? Patented? :) >Jon > >------------------------------------------------------------------------- >This SF.Net email is sponsored by the Moblin Your Move Developer's challenge >Build the coolest Linux based applications with Moblin SDK & win great > prizes Grand prize is a trip for two to an Open Source event anywhere in > the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ >_______________________________________________ >Emc-users mailing list >[email protected] >https://lists.sourceforge.net/lists/listinfo/emc-users -- Cheers, Gene "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) Ignore previous fortune. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
