[rich909]

Hi Julian --

I have a Zenbot 1624 on order which I plan to use primarily for cutting Baltic birch plywood (up to 3/4"), as well as HDPE plastic.

<snip>

For CNC production to realistically work I will probably need only about 0.01" accuracy. For 3/4" ply I plan tentatively to first try 1/8" passes with a 1/4" dia. end mill, at whatever feed and speed combo works. Here's another question: Does using higher feed rates with the Zenbot seem to contribute at all to diminished accuracy? I'd guess that heavier acceleration loads on the belts might lead to tiny amounts of over-travel at higher feed rates.

The Zenbot uses 285 ounce/in Nema-23 motors. If tool pressure is high enough, steps can be skipped/missed. NOTE: This is true for all similar machines that do not use "closed loop" servo drives. Essentially the Zenbot/Gecko controller says go to a specific location on an axis, but it has no way to know for sure that it actually gets there. That means if the axis meets resistance heavy enough to overcome the stepper motor it likely will not be at the expected spot. Such heavy resistance could be caused by debris on the axis rails, collision with a fixed object, or heavy tool pressure ...among other causes, resulting in inaccurate cuts and spoiled work.

My rule when cutting any new material is to use multiple light cuts on sample material to determine a safe optimum rate. Most of the time, I am interested in accurate "repeatability" rather than fastest speeds. For your example, in Baltic birch plywood, I'd likely start with a pass depth ot 1/32nd or 1/16th" (0.03 - 0.06) to start with the 1/4" bit at about 20-30ipm. By experimentation, you may find that you can cut with fewer passes and at higher speed reliably and get the accuracy you seek. The belt drive system has virtually NO slop and is not cause.

Another factor to consider is the weight of the router you select for a spindle. There have been several good discussions on this topic here at this thread and other threads, like: http://www.cnczone.com/forums/commer...bosch-etc.html If you haven't already seen these you should look them over. The spindle weight will have an inertia affect at the end of rapid moves. It is usually best to allow some flexibility in the table to help absorb and cancel this effect, rather than trying to make the table "rock solid".

Since the Z-axis slide is neutrally balanced before the spindle is mounted, you *may* want to add a counter balance to make the effort to raise it equal to the effort required to lower it. I understand Shaun is supplying a counter-balance with some of the new Zenbots and yours may come with one. If not there is an example implementatin here at post #79. Check with Shaun to see how your 16x24 is supplied. Normally, a Constant-Force spring is used and the spring weight rate should be about the weight of the router/spindle or a little more.

HTH.

--Rich

[famman47]

Thanks a bunch for your suggestions! I'll plan to take more conservative steps in determining maximum cut depth and feed for the accuracy I want. One of the more recent Youtube videos of a Zenbot in action (https://www.youtube.com/watch?v=1wCOSuVtoO8) shows what looks like a counter balance spring located behind the Z-axis assembly, so hopefully it's a factory-installed feature at this point. My plan is to do as many experiments as I can come up with before actually trying to produce usable parts. I'll want to determine if climb cutting or regular cutting produces any variation in cut accuracy between the top and bottom of the cut, due to tool deflection or for some other reason. One task I'll want the machine to do when manufacturing parts is to drill holes of uniform depth across most of the width and length of the bed. Since the bed itself is unreinforced plywood, I may need to machine down a spoil board to get a uniform height "working bed." Another thing I've wondered about is if there are strategies for maintaining a high level of accuracy when chaning tools on a Zenbot midway through work on a single piece. Is this generally done by visually positioning each tool according to a registration mark on the work piece, or are there other strategies?

Thanks again for your pointers, Julian

[rich909]

Hi Julian --

Sorry for the delay getting back to you on this.

Thanks a bunch for your suggestions! <snip> One task I'll want the machine to do when manufacturing parts is to drill holes of uniform depth across most of the width and length of the bed. Since the bed itself is unreinforced plywood, I may need to machine down a spoil board to get a uniform height "working bed."

A spoil board is definitely a good idea for flatness as well as to protect the plywood bed.

Another thing I've wondered about is if there are strategies for maintaining a high level of accuracy when chaning tools on a Zenbot midway through work on a single piece. Is this generally done by visually positioning each tool according to a registration mark on the work piece, or are there other strategies? Thanks again for your pointers, Julian

Depending on how you lay out jobs and the size of parts you cut, bed flatness may not be a big issue. I use a small format Zenbot partly because I tend to do "one-up" job layouts rather than "multi-up" jobs where a number of parts are cut out of one large sheet. Typically, I set up a jig to hold a part or "part blank" for a specific operation. Then I can run the job, pop in the next part and run the job again and repeat until a batch of parts has been run with that tool. The effect of this is that dead flatness is only needed over the area of the part blank, usually much smaller than the entire bed.

It is a good idea to set up a registration mark to establish absolute x,y zero (as you mentioned) when you are producing a single part or a batch. If something goes wrong and you need to re-zero the machine in the midst of a batch it will be very helpful. Some people use a corner of a part blank for this, ...but it may be easier and more accurate to use a 1/8" drill rod in the collett that fits a 1/8" index hole located in a "waste" area on the part blank. Then you switch to the tool required for the job when the machine is zeroed.

HTH.

--Rich

[famman47]

Using a hole for indexing sounds like the way to go. I'm still waiting on delivery of my machine, but I suspect that a lot of the questions I'm coming up with I'll be able to answer for myself by experimentation. One thing I'm really interested to find out is how resistant the motor/belt drive setup on the Zenbot axes is to slight, unintended repositioning due to the forces and jostling of tool changing. Have you ever noticed anything like this? I'm sure I'll probably be able to figure this out pretty quickly through experimentation. I suspect that if the machine is turned on and the motor controller is sending an "idle" signal to the motors, that the motors full strength would have to be overcome to manually reposition the spindle. But perhaps it's not a safe tool change/indexing strategy to leave it turned on. I imagine that for the purpose of supporting accuracy with tool changes, the most reasonable long term solution is the installation of homing switches, so as to allow the control software to get the cutter re-indexed (another learning hurdle)... Lots to learn, Julian

[rich909]

Hi Julian --

Using a hole for indexing sounds like the way to go. I'm still waiting on delivery of my machine, but I suspect that a lot of the questions I'm coming up with I'll be able to answer for myself by experimentation.

Yes, I am sure you will quickly discover shortcuts & operating protocols to suit the type of jobs you frequently do. While you are waiting for your machine, you might want to sign up at: https://groups.yahoo.com/neo/groups/UZUGR_CNC/info ...if you haven't done so already. You will find some archived info there that may be useful, and you can make contact with other Zenbot owners who share your interests.

One thing I'm really interested to find out is how resistant the motor/belt drive setup on the Zenbot axes is to slight, unintended repositioning due to the forces and jostling of tool changing. Have you ever noticed anything like this? I'm sure I'll probably be able to figure this out pretty quickly through experimentation. I suspect that if the machine is turned on and the motor controller is sending an "idle" signal to the motors, that the motors full strength would have to be overcome to manually reposition the spindle. But perhaps it's not a safe tool change/indexing strategy to leave it turned on.

If you are reasonably careful you will have no problem changing tools/cutters after the Zenbot has been initially ZEROED for the session. One exception to this is that if the tool lengths are different, you will have to re-zero the Z-Axis. Some bits have a depth ring installed so that when you mount different tools, they extend the same distance from the collet.

If you are using a router as the spindle, turning the spindle motor on and off is done manually. There is normally NO RISK of the spindle motor starting during a change of tools/cutters. Assuming you are planing on using Mach-3 {or an equivalent} and with the toolpath program stopped or paused, it is easy to change a tool bit without overpowering the "position holding" ability of the steppers, assuming reasonable care and good access to the spindle. You could force the XYZ out of ZERO but it would take considerable intentional effort to do so. The belt drive system has ZERO slop, so unlike some other types of systems there is no "lash" that would be affected by a tool change.

I imagine that for the purpose of supporting accuracy with tool changes, the most reasonable long term solution is the installation of homing switches, so as to allow the control software to get the cutter re-indexed (another learning hurdle)... Lots to learn, Julian

Homing and limit switches are a considered to be a convenience by some, but they are not really essential. Actually, you may find it easier to set manually zero off of an index spot on the spoil board at the start of a session. In practice, home and max limits are often considerably smaller than the size of the work table and you may not always want to set the origin of the job at the same XY location.

BTW, there are many other considerations regarding spindle and software selections, some of which may depend on the type of work you plan to do and the materials youwill use. Some of those may be answered by earlier posts in this thread, but you will also find related discussions at the yahoo user's group mentioned above.

HTH.

--Rich

[famman47]

Received my Zenbot 1624 a couple of days ago, have it set up and just "air-cut" the roadrunner code for the first time. Very cool! I'll go ahead and start posting some questions and observations over at the Unofficial Zenbot User Group, but I'll probably be back to this thread at some point to add pictures and information to be more publicly available.

Julian

[rich909]

Hi Julian and all --

Received my Zenbot 1624 a couple of days ago, have it set up and just "air-cut" the roadrunner code for the first time. Very cool! I'll go ahead and start posting some questions and observations over at the Unofficial Zenbot User Group, but I'll probably be back to this thread at some point to add pictures and information to be more publicly available.
Julian

Any recent updates or projects to share with us?

--Rich