Reduction gearboxes on servos-good thing?

[Splint]

Hi,
can anyone give me an idea on how suitable reduction gearboxes are on servo motors in regard to backlash as compared to a belt and pulley type reduction. I've spotted some servos with reduction gearboxes and timing pulleys on the output shaft and I don't know if the backlash will be excessive to drive an x or y axis on a router. Can the mesh of the gears be adjusted to regulate backlash?

Thanks
Splint

[Al_The_Man]

For CNC you want to look for low backlash planetary gear boxes like Alpha, Bayside and others, the high reduction allows economic sizing of the motor, although the gearbox pushes the price up unless you go the ebay route.
But on large heavy gantry's the like of over 200lbs it is pretty much essential.
Al.

[RICHARD ZASTROW]

Splint It depends on your accuracy requirements. A proper belt drive assembly will have no backlash, but you will have to provide adequate motive power. If you are limited to lower power and torque building gearing, you must deal with reduced accuracy. Your compromise will give you the answer.

[Adobe Machine]

My X and Y axis Retrofitted Tree Mill have Bayside Planetary servo gearhead
reducers ( 5:1) Did not use on the Z axis. (Quill)
The inspection report generated by the manufacture shows the maximum back lash in the units is 1.08 arc minutes. Thats a very small backlash, and in fact the 26 or 27 year old machine shows less than .0007 even when I push it real hard. Not a real fast machine (can cut alum at 75 IPM), but really has some omph cutting alloy steels.
I did overhaul/re-load the ball screws with oversize balls, the fixed angular bearings were in excellent shape, so left them alone.

Adobe (old as dirt )

[Billhill]

I'm scratch-building a CNC machine for metal work, and wonder how much force is needed to quickly machine carbon steel.

My DC servo motors have 300 oz in of torque, RPMs to 4000.

I expect my Z-axis will weigh around 50 pounds.

I've been looking into planetary gearheads to get useful torque, and have been thinking of ratios in the 6 to 50:1 range.

Any opinions on Mijno, Rino, Neugart, Diequa planetaries? How about Harmonic Drive reducers?

Thanks in advance for your opinions.

[RICHARD ZASTROW]

Billhill, Gearheads can be very accurate and increase torque relative to reduction and gearhead efficiency.

Remember, you must also consider pulse count relative to gear reduction. Odd-ball ratios can be a problem when multiplying count per rev and lead of screw etc. Fortunately, I don't have to deal with that. I ask the electronics whizz-kid what is acceptable and design the gear reducer to match.

Another consideration: all that torque can destroy small planetary gear teeth if the gearhead is too light duty. Some of them have teeth the size of the serrations on a fingernail file.

Good luck on whatever you decide.

[Billhill]

Thanks for the replies.
I stuck a toe in the water with one affordable planetary gearbox from Bane Bots, with a 12:1 ratio.

While it came almost ready to use (disassembled input gear, no motor coupler offered), the input gear is about 15mm dia, and serious looking. I was surprised not to get a page of instructions, or a specification sheet, maybe it was an oversight.

If I can find those, I may buy more/different units.

I'd recommend the vendor anyway for trying to do a good job. It looks workman like, though nicked and scratched a bit, showing vise marks, etc. I don't care too much about that.

I did care about the price and the many ordering options, which are excellent.
Thumbs up for Bane Bots. Thanks.

[Al_The_Man]

Originally Posted by Billhill While it came almost ready to use (disassembled input gear, no motor coupler offered), the input gear is about 15mm dia, and serious looking. I was surprised not to get a page of instructions, or a specification sheet, maybe it was an oversight. though nicked and scratched a bit, showing vise marks, etc. I don't care too much about that. .

It very much sounds like used equipment? What is the make?
Al.

[Billhill]

No, I don't believe it is used. The output shaft, which is about 4" long is pristine, as is the input gear (loose) and key. It is advertised as new, and I believe them.

They typically supply the FIRST Robotics competition teams, and those building battle-bots, both of which are low-duty cycle applications.

But, before I decided, I looked over their documentation online, and chose a dual-stage ratio - 4:1 followed by 3:1, to evaluate their product, and I'll give it a thumbs up.

I think they use a CNC unit for housings, and assemble gears and bearings to order. My opinion, not necesarily a fact. If that is true, it is OK and a great vendor to work with for us prototype people. Even if I were rich or had tons of customers, I'd still work with them, instead of Atlanta or Mijno or CGI or Rino who all think you're made out of money or don't care to work with the little guy.

Thanks for your concern, though. As I get more experience, I will update the thread. I sent them a copy of my post today, asking for a quote for a motor coupler and input shaft, and I will more thoroughly cruise their site looking for specs & installation instructions.
Aloha - Bill Hill

[Mike Everman]

Just my 2c:
With gears, if you ever want to switch to linear encoders later, then you are far better off with a belt reduction.

Belts are zero backlash, but only if you define backlash as a "zero stiffness deadband". I have 4:1 reductions here using Gates GT2-3mm (the style that would be best for indexing accuracy, according to Gates). People do not normally encounter the error in a belt reduction. Simply put, there is tangential tooth clearance that "walks around" as the pulley turns, such that teeth march from contact with a forward face to a rearward face. It's no small amount, considering the 1 arcmin gearheads you're talking about.

The reduction I have is on a large camera pan-tilt, where indexing accuracy had to be less than .12 deg TIR. Plain Jane out of the box, standard clearance pulleys are going to get you that or more, so call it 7 arc minutes!

The way to get rid of that is to order zero clearance pulleys (not widely available). In my case, the pinions can be found low or zero clearance, (and it's far less an issue at low tooth count), but the bigger pulley is hobbed to order.

Engineering at Gates stated the obvious: the pulley needs to be hobbed to a larger OD, effectively thickening the teeth to reduce clearance. They gave it a value of .004" over nominal. The first 200 or so were hobbed to nominal and were blowing my spec. The quick fix, and you may get away with this generally in a DIY project is to cut down the OD, which buries the belt teeth in the bottom of the pulley grooves. Not an ideal mesh for maximum life, but it works.

If you do that to pulleys you have around, I'd suggest some compatible lube.

[Billhill]

great post!

Thanks.