Re: [CAD_CAM_EDM_DRO] removing backlash from rack/pinion drives wasRe: Shoptask Z-axis - quadralift?

Interesting, but as you say, it would require two adjustments, as opposed to
one. Were the mechanics of it relatively simple, or complex? It sounds like
one of those things that could be either, depending upon the implementation.

Bill

On Fri Mar 7 02:16:51 2003 paul_norton2001 wrote:

>Not my idea, I've seen it on a webpage somewhere. The motor has a
>gear, this gear drives two others simutaneously, both of which which
>mesh with the rack. Backlash is reduced (eliminated?) by moving the
>centres of the driven gears apart. The driver gear position would
>have to be adjusted to take up backlash.
>I guess it can be compared with an anti-backlash nut but with a
>solid sleeve between the nuts rather than a spring.
>
>Hope this helps.
>Paul N
>
>--- In CAD_CAM_EDM_DRO@yahoogroups.com, "ballendo" <ballendo@y...>
>wrote:
>> Bill,
>>
>> Never seen that approach. But...
>>
>> What is typically done is similar. The GEAR is split (or two
>narrow
>> gears are used if you prefer<G>) and these two "halves" are sping
>> loaded (rotationally) against each other. You can see examples in
>the
>> stock drive products catalog, which is available free.
>>
>> A 2.5 axis cnc machine would make short work of making these from
>a
>> couple of stock gears. A manual mill and a rotary table wouldn't
>be
>> too far behind...
>>
>> Hope this helps,
>>
>> Ballendo
>>
>> --- In CAD_CAM_EDM_DRO@yahoogroups.com, ccq@x wrote:
>> > Hi guys;
>> >
>> > Maybe someones tried this before, and can tell us if it'd work.
>It
>> might
>> > require a wider than usual gear/rack, but what if you took a
>> slitting saw, and
>> > slit the rack right down the middle. lock one half down as
>usual,
>> and then
>> > slide the other half lengthwise, one direction or the other,
>until
>> all the slop
>> > is taken out. Would this be workable, or just engender extreme
>> wear, or
>> > sosmething?
>> >
>> > Bill
>> >
>> >
>> > On Wed Mar 5 16:20:44 2003 alan@s... wrote:
>> >
>> > >On 5 Mar, CAD_CAM_EDM_DRO@yahoogroups.com wrote:
>> > >
>> > >> This is addressed to those who have CNC'd Shoptask machines.
>> > >
>> > >This is an interesting question that I am currently struggling
>with
>> > >myself.
>> > >
>> > >I do have a CNC'd Shoptask.
>> > >
>> > >I do not have a Quadralift, but I am certainly going to add
>> SOMETHING
>> > >to raise and lower the mill head, for reasons that need no
>> explanation
>> > >to anyone else with a 3in1.
>> > >
>> > > Frankly, as I examine pictures of the QL, I'm pretty sure
>that a
>> guy
>> > > could make one himself, of at least the same quality, and do
>it
>> all on
>> > > the Shoptask.
>> > >
>> > >The question then becomes "why use the quill at all ?"
>> > >
>> > >You see, I don't agree that the quill is necessarily more
>> accurate. I
>> > >have taken great pains to take all the play that I can out of
>the
>> quill
>> > >(removing the "drill press mode" by pinning the gear, using
>thrust
>> > >washers on both ends on the fine feed shaft with adjustable
>> preload,
>> > >etc) and I still get the inevitable play as a result of the lash
>> > >between the gear and the rack cut into the quill.
>> > >
>> > >The issue is easily seen. Bring the quill down until it just
>> touches
>> > >the work. Now try to turn the crank some more. You should not
>be
>> able
>> > >to .. you're touching the work, right ? But you CAN turn it
>some
>> more
>> > >... you're taking up the lash in the above mentioned gear train.
>> > >
>> > >Now the REAL question .. is the above seen lash constant
>> throughout the
>> > >travel of the quill ? Not on mine, and likely not on yours ..
>the
>> rack
>> > >is just not machined accurately enough.
>> > >
>> > >Even if it is, you're left with "zeroing" the work, by what
>ever
>> means
>> > >usually taken ( rice paper, edge finder, etc ), then adding in
>the
>> lash
>> > >you know is there and hope your chosen offset is correct.
>> > >
>> > >There's a good reason why Chuck McGhee's ball screw quill
>upgrade
>> is so
>> > >popular.
>> > >
>> > >So here's the thoughts kicking around my brain on this topic.
>> > >
>> > >Since I need SOME sort of screw to move the mill head up and
>down
>> on my
>> > >homebrew QL, why not make it a QUALITY acme screw with QUALITY
>> > >doublenuts.
>> > >
>> > >Make sure the QL goes up and down with REALLY good
>registration.
>> > >Thomson linear slides is my current scheme, but there are other
>> choices.
>> > >
>> > >Then just lock the quill ( it's more rigid that way anyway )
>and
>> always
>> > >use the whole head for movement, while under CNC control.
>You'd
>> want a
>> > >hefty SERVO/STEPPER motor to get any sort of speed out of the
>> deal, but
>> > >that's not a big deal.
>> > >
>> > >Problem solved.
>> > >
>> > >Now, as to the real gist of your question, your double encoder
>> scheme
>> > >has a real problem. You can certainly wire something up
>to "add
>> up"
>> > >the encoders, but this only makes sense if you know PRECISELY
>how
>> far
>> > >your drill bit in a chuck sits below some common reference
>point,
>> as
>> > >compared to how far your endmills in a holder sit below that
>same
>> > >reference point. God forbid you ever sharpen that bit.
>> > >
>> > >Only with this precise information does your double-encoder
>scheme
>> make
>> > >sense.
>> > >
>> > >Here's what I would do.
>> > >
>> > >Make a height adjustable electronic edge finder. ( A led, a
>> battery a
>> > >resistor ) The "switch" to turn on the LED is made up of the
>> machine
>> > >and the insulated top of the edge finder.
>> > >
>> > >The circuit is closed when the endmill or drill bit JUST
>touches
>> the top
>> > >of the edge finder.
>> > >
>> > >No great precision is needed here, except to make the top of
>the
>> edge
>> > >finder very parallel to your machine table.
>> > >
>> > >Now, find the X and Y zeros of your workpiece, by your normal
>> means.
>> > >
>> > >Next, find the top of your workpiece also. This is your Z axis
>> zero.
>> > >Clamp your edgefinder on some unused part of the table, screwed
>> down all
>> > >the way, and move the quill ( under CNC control ) over to about
>the
>> > >middle of it.
>> > >
>> > >Now slowly raise the edgefinder until the light JUST goes on.
>You
>> now
>> > >have a surface in a known X-Y-Z location.
>> > >
>> > >Raise the head .1 ( to avoid scarring the surface of the
>finder ),
>> move
>> > >to X=0, Y=0, lower the head .1 and Z also now equals 0. Mill
>away.
>> > >
>> > >Tool change time, raise the head ( manual or auto ), change the
>> tool,
>> > >move under CNC to the edgefinder XY location, MANUALLY lower
>the
>> tool
>> > >until the light JUST goes on. Zero EMC. Raise the head .1,
>move
>> to
>> > >X=1 ... etc.
>> > >
>> > >My wild idea, anyway.
>> > >
>> > >
>> > >Alan
>> > >
>> > >
>> > >
>> > >--
>> > >
>> > >Alan Rothenbush | The Spartans do not ask the
>number
>> of the
>> > >Academic Computing Services | enemy, only where they are.
>> > >Simon Fraser University |
>> > >Burnaby, B.C., Canada | Agix of
>> Sparta
>


--
----------------------------------------------------------------------------
RKBA! ***** Blessings On THee, Oh Israel! ***** 4-19!
----------------+----------+--------------------------+---------------------
An _EFFECTIVE_ | Insured | All matter is vibration. | Let he who hath no
weapon in every | by COLT; | -- Max Planck | weapon sell his
hand = Freedom | DIAL | In the beginning was the | garment and buy a
on every side! | 1911-A1. | word. -- The Bible | sword.--Jesus Christ
----------------+----------+--------------------------+---------------------

Constitutional Government is dead, LONG LIVE THE CONSTITUTION!!!!!

----------------------------------------------------------------------------