Originally Posted by noisillator  Is that the 1.7uS spec? I assume you're not fond of the National LMD18245 either then? 
Thanks for your comments on the steppers. I searched, but wasn't able to locate the exact article. I did find other mention of the "point of diminishing returns" being in the 140-160 range, but the material is dated. Maybe the one I remember was older as well.
Incidentally, I can appreciate your position regarding the engineering and performance of the Gecko drives. However, it is the purpose of engineering to solve problems. If those problems are relatively minor in terms of accomplishing the work, the additional engineering cost may not be justified. I believe that's the case regarding the Gecko controllers, at least from the perspective of a hobbyist. Searching the 'Net, I find a number of HobbyCNC users who seem entirely satisfied with their results. Also, I'm on the fence regarding resonance solutions, insofar as it may be possible to dampen it mechanically rather than electrically. Neither method is a perfect solution, but mechanical dampening appears to be an inexpensive add-on, particularly if one also has access to a lathe. In the end, this is all a matter of matching the cost to the job at hand, and to do that in a way that compliments both my DIY skills and the goals I've set for my projects. |
The Ton-min time versus the Toff time is the problem. To put a fullsteop across a 3v coil with a 36v source, the ratio must be 1:12. End of story. For
microsteps, the destination voltage and current changes! For an 8x microstep, the 3v coil might have a target voltage of like 1v for some steps, requiring a 1:36 ratio. If the driver has a minimum Ton-Toff of 1:15, the lowest it could make even in theory is 2.4v, and that's what will come out instead of the intended voltage/current. That causes 1) failure to step smoothly, 2) the static position is wrong (dampeners will not fix this), 3) other "problems", because the voltage and current dynamics are thrown off in more complicated ways than I'm getting into here.
Mechanical dampeners might help, but I see no reason to
buy into this situation. If you bought a working setup and were not prepared to gut the controller box, that's one thing. But you're assembling a setup, and can do whatever you want. $79+hours spend assembling & casing+hours spent heatsinking+possible burned out drivers+buying a power supply of a substandard low voltage is much worse than $299 for a G540.
As far as other people being "happy"... how can I say this. The standards of what it takes to "impress" amateurs varies quite a bit. Merely getting a mill to move by computer at all is amazing in itself, is it not? Things you have never seen before. But that doesn't mean the performance is actually impressive per current industry standards. In fact G540 totally changed the standards anyways.
The problem is that an inferior driver is not just "slower" at times. You must slow it down at all times because the risk of stalling can destroy your mill. The headstock stalls on lift (common), thinks its higher than it is, and next comes back down and mills through the table itself. This SUCKS. You will play a game endlessly of wondering what speed you dare to use and it may change with materials, hogging aluminum requires a lot of torque. Back when that was the way ALL drivers worked, well, people dealt with it. We don't need to anymore. If the G540 were $1000 it'd be a bigger question, but it's not.
The G540 also has a high resale value. Pretty much $299-ish! And in a few years it'll still be in demand because it does everything a driver for a mill needs to do, reliably. Much less demand for hacked-together obsolete drivers.