Thank you for your advice.
With regard to the weight, is it really that bad when compared to the weight of the gantry? I know rotational inertia is a problem for acceleration where large diameters and masses are involved, lighter flywheels on cars & bikes don't improve power, they effectively loose mass that needs accelerating, but then we are talking several lbs saved at 10s of inches diameter there, rather than ounces at a couple of inches. I think cost will be more of my consideration here rather than acceleration, though significant, real world improvements can always change my mind where cost is involved!
I forgot to explain
A heavy pulley is a flywheel, flywheels have long been used to add inertia to systems, and if you look at the maths you will find they do it well. It takes surprisingly little effort to accelerate a gantry, if it weighs 20 lbs then a 20 pound force will accelerate it at 1G, same as if you dropped it. If you add unnecesary inertia to the system you have to watch out for that moment when you hurl the gantry in to reverse. That is the moment you risk losing steps and blowing your position. Unnecessary inertia is to be avoided for that reason.
As you like explanations I should also warn you about choosing stepper motors by holding torque alone. For some obscure reason the hobby market has settled on stepper motors with an arbitrary 200 full steps per rev, so the field coils have to turn the rotor 1.8 degrees. If you use a large rotor you get more leverage on the shaft but the pull in distance for 1.8 degrees is larger than for a small motor. Unfortunately as the distance increases the magnetic field falls away by the inverse square law so you tend to lose more than you gain.
Next problem is that the suppliers have found that nobody seems to care a fig for the actual motor performance, they only look at the holding/detente torque for the motor. Obviously the holding torque is measured with the minimum gap so big numbers are easy achieve so that is exactly what they do. What you are really interested in is the pull in torque and the inductance of the field coils.
You can "fix" a large 200 step motor by increasing the volts but that is expensive in the driver/PSU department. When choosing a motor you want to view holding torque as a simple indicator so you gett in the right ball park, what you really want is a graph of torque against speed and see how it goes with what you are trying to achive in accuracy while still giving a good rapid which gets more important as the size of the machine bed increases.
Edit: There is also a tendency to sell motors in sets of 3 for XYZ, why anyone would want to use the same motor for 3 completely different tasks eludes me.
Last edited by Robin Hewitt; 02-08-2012 at 11:58 AM.
You might find this useful: http://www.mycncuk.com/forums/faqs-p...do-i-need.html
As Robin has already alluded, stepper motor holding torque is what it takes to move the rotor from a stationary position. Once it's moving the torque available is considerably lower and essentially flat until you get to what is called the corner speed after which it becomes inversely proportional to speed. Ideally you run your motors at or just below this corner speed. The corner speed itself is dependent on the rise time of the current in the coils, and that depends essentially on the ratio of the coil inductance to its resistance. As Robin said, one way to improve the corner speed is to drive the coils harder by applying a much higher voltage (most stepper coils are rated for 3 - 5v but will often be driven from 12 - 80v) and controlling the current to avoid burning them out - constant current drive (expensive, heat producing) or chopper regulation (cheaper, but noisy and prone to other side effects).
Sadly, very few stepper manufacturers provide graphs of torque vs speed at the rated current so its always a little bit of 'what has everyone else done'.
Thanks for your reply Robin,
I was heading down the right track, I just didn't get to the end in time!
I have a real love for the old Mini and the A-Series engine (Yes this is still on track) sadly can no longer justify the expense of owning one along side my pickup truck :-( In the tuning manuals I had there was the formula that proved a few pounds off the weight of the flywheel equated to a significantly larger weight reduction in the vehicle. Obviously this effective weight loss was reduced with each gear, as the rate of acceleration decreased. Had I given myself a bit more time to think, it would have been clear to me that this holds true with the mass of the pulleys vs the mass of the gantry Doh!!!!
There are so many variables in each component, there comes a time when you have to say enough - any more information your brain will fry - and base your choices on what seems to be the best given what you know and what the rest of the world is doing in a similar situation.
I agree with your final statement, I had already thought that in an ideal world the motors need to be sized lower performance for Z, rising through Y to X as each axis in that order, has relatively more to do. However since this is not a racing machine where everything has to be maximum performance and minimum safe construction, I couldn't see the harm if Z&Y were a little over specified, when compared to X. As you will have seen, I have asked above if a single 3.1 Nm is enough for my X Axis.
As you can imagine I have done lots and lots of reading and seen large machines working with quite small motors where budgets are tight and small machines apparently grossly over powered when it has been clear money was little object!
My choice of motors was down to several parameters:
1) Most machines I have found in the size range I am after (and some at 8'x4') use Nema 23s, so this seemed to be a the most sensible choice for frame size. Incidentally it also fits in nicely with my choice of extrusions.
2) This page (Drives and Steppers~Solsylva CNC Plans) had the following statement: "Steppers in the ~300 oz.in. range can push a full sized router, but they lack the power to push the router to its limit. The steppers will most certainly stall before the router does." That ruled out the lower torque range of the Nema 23s. Also I have a copy of the TAB CNC Robotics book by Geoff Williams, his steel construction machine gets away with motors that are at best 220oz.in. So I decided on 300oz.in minimum.
Since this project does not have an unlimited budget, the best "Bang per Buck" method was going to have to be used. Also I chose to make the assumption that for a given frame size and therefore presumably similar construction a Higher holding torque should lead to a correspondingly higher dynamic torque.
My search led me to the 439oz.in (3.1Nm) motors from CNC4You at around £100 for four. At roughly 30% over the torque suggested by what I could find at the time, I guessed these would probably be man enough. My limited knowledge and obviously finite research makes these the best cost/performance ratio I can find. Given my resignation to the fact that finite budget machines are always a compromise, I am happy with my choice, but I am willing to be proved wrong!
Q: Does anyone know of 4 better motors out there for the same money, give or take a Tenner?
Now I'm going to be using belts to drive the twin screws the 4th motor is going to rest in a draw until I get round to my 4th Axis.
Thanks for the link, I'll check that out when my brain cools down!
No need to break your brain if you use sensible units such as the British standard fat bloke
The standard fat bloke weighs 16 stone which translates conveniently to a downward force of 1000 Newtons.
Direct drive on to a 5mm pitch screw requires 0.8 Nm of torque to lift one fat bloke off the floor.
Look at your cutter in it's collet and ask yourself, how many fat blokes could stand on the sharp end of this before it snaps?
Multiply by 0.8, call it Newton meters and anything above that is superfluous.
I'd be more worried about accelerating the fat bloke, although it's still not going to change Robin's conclusion that a smaller motor is fine for your Z-axis.
The question really is do you envisage using the motors on a larger/heavier CNC Router, or a milling machine in the future? If so then it's sensible to get the 3Nm motors as they a very likely to be well matched to the future machine. If not then there are motors with a lower rated torque, but would perform better on your machine due to their lower inductance enabling them to reach a higher speed than the 3Nm motors. For instance play around with Irving's spreadsheet to compare this 1.85Nm 1.6mH one to the 3Nm motor for your size machine. The annoying thing then is you're spending more money on a smaller motor, so if they do turn out to be better in theory you've got to be sure you wont regret not having the bigger motors in the future for something else. Additionally the 3Nm motors are fit for purpose, so if they're cheaper than the better alternative then why not?
One point about the X-axis slaving - I wouldn't compare two motors and no belts to one motor and belts, since the belts have so many advantages anyway that you should use them with either system. Given that, out of your list of reasons, the only one that applies is you pretty much eliminate the chance of racking the gantry if the screws are linked. What are the chances of one motor stalling if you have two? Jazz said he's never had his stall since setting up with one motor and I've not had problems with two motors after tuning.
With both the motor selection and drive mechanism, in the end either system is fit for purpose so there's no point in me suggesting that you should use one or the other. What is worth suggesting though is if you do stick with trapezoidal screws, then consider substituting the nuts for the 'wonky bearing drive nut' method I started out using:
Very low backlash, low friction and I bet you could make it with hand tools. There's certainly room for improvement on my design, some of which is discussed in my build log.
Hopefully I'm wrong, but it seems from your previous posts that you're considering not machining at least one end of the leadscrews presumably to save cost? If so forget it since the accuracy will be so poor.
That is excellent, when I stop laughing and the tears have cleared, I'll see if I can find how many fat blokes can stand on the average cutter, should be an interesting Google!
Now looking at Irving's stepper motor tutorial, will then see if I can estimate the mass of my machine components and see where I go from there. Since I'm quite a bit heavier than a BS Fatbloke I keep thinking I could snap (or at least bend) most things that will fit in a Kress collet!
Do we have shear force tables for typical cutters anywhere? Is it a common figure maufacturers supply? or do we just go by Dia of the cutter and its material spec?
I'm guessing snapping during plunging is much less of an issue as the material is in compression. So am I right in thinking that as long as the shear force for a cutter is not exceeded things should be OK?
Ok lets bring this back to real world use and implementation.?
While the motor calcs like Irvings are great they are really only good for giving a ballpark figure IMO.!
Unless every aspect of the machines friction and resistance to movement etc are known, which can vary wildly dependent on design, build quality and component quality then it's very difficult to accurately choose the perfect and ideal motor.
Also like Irving pointed out it's very difficult to get real useful motor torque curves graphs from manufactures.? When dealing with cheap motors this is no accident either because they would show that they are really quite below what people might expect.? Only the better motor manufacturers tend to give them and these cost far more money than the average DIY user would want to spend.!
In real DIY built in the shed, garage, basement or coal house and soon to be back of a boat.!! . . things don't often work out like the Calcs or graphs would have you believe so some (plenty) sticky wiggle room needs to be factored in.
Then there's electronics side and in-balance that using different sized motors brings to the control box and it's power requirements.?
Lets take this machine or system for instance.!
If ball-screws are used then it may be possible to use 1.85Nm motors which will run on 40-45V using 50V drives for 2 axis but the 3rd twin screw setup will need at least larger 3Nm motors and 65-70V and 75V drives. This means 2 different PSU's costing extra money and space.
The only advantages to the smaller motors will be they spin slightly faster which really in all probability won't be required, they will accelerate slightly faster but again the little extra will be negligible in real use for this machine, the total cost will be slightly less.
The while on paper the larger motors will accelerate slower and spin slower in real world use the difference won't make much if any difference to this machine other than cost. This is offset by the extra PSU required and not so massive.
The larger motors on the other hand will give great over build "wiggle room" while not having any real detrimental affects to machine performance.
Yes While it's very true Bigger is not always better in this case the difference won't be so much it affects performance to be worth bothering about.!
Then there's the longevity factor.? An over spec'd machine tends not to be so stressed and can run well within it's limits giving long life to both motors and drives.
That's my view on the motor choice obviously others have there's.!
Regards the twin screw and 3.1Nm then yes if you use ballscrews with 10mm pitch and gear either 2:1 or 1.5:1. This will increase torque and resolution while still keeping the motors in a nice RPM range well below the corner speed while cutting. The sacrifice will be rapid speed which to quote John S only any good for "dick slapping".
Like Jonathan I urge you to use ballscrews and the difference in cost if bought from china won't be very much and really lead screws can't be compared with even cheap ballscrews unless high quality lead screws which cost more than ballscrews.!
Use 16mm as well not 12mm.
Also if you use the 90x45 profile from KJN then the BK/BF blocks fit the slots and can be bolted straight on. Also the BK/BF blocks from china are ridiculously cheap and I see no point in making your own.?