Welcome, Bob.
Fortunately or no, there is plenty more where that came from, heh. I suspect as an EE, you will have very little issue picking up the design cues; there are some similar concepts (resonance is resonance, in either field, for example; more mass in the structure raises the resonance point...)
I would have to agree about design methods; I once knew an mechanical engineer who insisted on utilizing a fine thread on the exterior of an hydraulic cylinder to join it to a vertical plate, which then seized in place the very first time assembly was attempted (was his personal favorite join method). What was wrong with a traditional flange and bolt circle, I am still not quite sure (bangs head).
I have always heard good things about the Sherlines; I myself had the Taig lathe (there were no local Sherline dealers) and that was far more rigid than the much larger Chinese units I was looking at, and it provided great service within the work envelope. I guess I should not really discuss the time I tried to reduce some 17-4 stainless bushings in it, however... (screech).
There are some awesome resources available on the web devoted to the Sherline units. I almost adapted the Sherline bronze worm gear 4th axis, but in the end, it had too much backlash for my intended uses (just the nature of solid worm gear units; split or loaded are a differing story). How did you find yours?
Since I am jumping from topic to topic, have you encountered the
Micro-Machine Shop or
5 Bears in your travels; some interesting information on machine construction, tooling and measurement on those two (although NOT for the Grizzly unit you were asking about).
I gather Hoss was a professionally employed machinist at one point, and has a fair grasp of design mechanics as well. Definitely one of the more prolific posters out there on the subject of tinkering and original design. He built his own 4th, and after a lot of research, it is likely I will be one of the ones to follow in his footsteps. He specifically wanted (as do I) a unit that works in lathe and index mode, in his case utilizing the headstock of a surplus lathe and separate stepper motor. He has built such for his (now discontinued) X2 projects, and was porting the designs over to the Grizzly unit.
He was also making a 5th axis (should be on YouTube under Hossmachine), using 3D printed housings and metal gearing (although I do not know how serious he was about it). I do not currently have the room for the larger machine, but I have been thinking of picking up his DVD just to peruse his designs as well.
If you would like to see the absolute extreme end of a Do-It-Yourself 4th Axis, research the
InTurn on YouTube. It is a heavy duty dual operational mode unit, much in the same vein, and is gorgeous. The fellow had a long build post on the Mach3 forums.
With regards to feeds and speeds, along with work-area, here is how I did it. There is a nice product called G-Wizard by Bob Warfield. It is not new, and has a free trial variant. It is based on a subscription model, but works for hobbyists thus: once the initial paid subscription expires, it runs in a Lite mode; any hobby machine with 1 HP or under spindle can continue to use it. Here is the link:
CNCCookbook; the site has a considerable amount of additional CNC information as well.
There are other calculators, such as the web version of
FS Wizard, which also has a long development period, and free. There is also a Pro (pay) variant. Each program has interesting features.
With work area determined by measurement of X,Y, and Z travels, I simply went online, found a calculator for the most common materials I expected to be working (4140 steel and aluminum) and used it to tell me what a chunk would mass in the work envelope I had on the machine (utilizing average density of the alloys and form factor).
Then I went to G-Wizard, selected various tooling, and told it to calculate what feeds and speeds would be generally acceptable (one highlight of G-Wizard is it can de-rate the machine based on perceived rigidity; handy for those of us with less-than-rigid Chinese machines).
Once I had that information (highest mass, lower feeds and lower mass, highest feeds) I could proceed to work out the required torque and RPM limits to push said block around using similar links to what I supplied earlier.
As it turns out, the motors suggested by Hoss et al. worked out just fine. I can drive the machine harder than I feel comfortable running it.
Speaking of microstepping, that is something else that should be taken into consideration when system building. It is great at smoothing motor operation, reducing motor resonance, but not so great for precision or
incremental holding torque (at least in the high numbers; some hold anything past 8x as being a waste, Mariss of Gecko has stated he feels anything past 10x :
Link). Here is a
favored link on the subject of microstepping.
Since steppers are not exactly known for ultimate precision, even those with a 0.9 degree step instead of the standard 1.8 (and good luck finding reasonably priced AND powerful 0.9 units), but with careful selection of the lead or ballscrew and control of backlash, the positional errors can easily be kept under those generated by the drivetrain.
Speaking of drivetrain, I just remembered that
CNCfusion also apparently has a ballscrew and mount kit for the Grizzly-slash-BF20; might be of interest.
For a motor controller, I use the Dynomotion KFLOP and KSTEP (16x microstep) with standard 1.8 degree motors, and they can place the motors better than the dovetail ways can hold position (I need to learn to scrape the dovetail in; the gib shows wear on a narrow band instead of the whole surface).
The Gecko drives have adjustable microstepping, unlike the fixed value in the KSTEP. The Geckos are also independent drives, so if you destroy one, just the one needs replacement. Some find that useful. There are other systems: Granite Drives, DMM Tech are two that come to mind.
Dave, again, is of the same mind as myself regarding closed loop operation, but experience has shown a properly designed and operated stepper system does not lose steps. While I have equipped my motors with CUI AMT capacitive encoders for close loop velocity and positional operation, I have not yet actually engaged them. A fair amount of people are reporting them to have considerable noise at certain operational settings, as well as some dithering at rest (here is a
link to show how bad it can get). While they apparently can be made to work, the US Digital E5 or similar encoders (actual glass) should be superior. So far, I am not missing the capability.
In the future, I may become absurd enough to use two control loops; one for actual table position (glass or magnetic scales), and the other for the motor velocity (controlling the motor position alone does not account for backlash in the drive train). The perceived accuracy is likely to greatly exceed the actual repeatable positioning of the mill, with the rough castings, stiction and deflection properties under load.