Thread: Men's Shed Router

A couple of our members intend doing some shopping for suitable cutters.
The only cutters we have at the moment are straight single flute.
With 12k RPM and 400W available, what sort of cutters should we be looking for ?
Would McJing in Sydney have such cutters ?
John

Some progress, we didn't get a job done today, but we did get the platen flat.

The first job attempted was a simple box with tabs and slots.
The sides came out with sloped top edges, and the bottom didn't cut at all.
We had problems at the code generation stage and suspect there are multiple overlaid lines in the original CAD file.

The second job was the Shed's logo generated with CamBam at home by a member.
The file would not run on the Shed's copy of CamBam.

So Plan B kicked in and the platen was surfaced.
Very interesting to observe minor chatter when cutting in the conventional direction, and absence of chatter when climbing.

The dust produced demonstrated that the next development should be a vacuum suction system.

John

electrosteam, you mentioned above about cutters, just recently on a visit to Aldi, they had an impressive set router bits with bearings in a case for around $20.

I was very tempted to buy a set but pulled up asking myself, "why do I need a set of router bits"?

Router Bits - Aldi — Ireland - Specials archive

Ignore the fact that this ad is an Irish one, these sets are here too.

Ken

I've been picking up small end mill, ball nose and v bits from China via aliexpress, normally in sets of either 5 or 10 cutters, very cheaply. Lots of sellers, and so far I've only had issues with one supplier. They only shipped half the ordered number of bits, but resent the entire order when I let them know about it. They last a surprising amount of time, especially as I often run at non-ideal feed rates. Bit-death is normally due to me stuffing up (thank god for wooden hold downs).

1/4" shank bits are normally the cheapest, which is great if somebody wants to do some fine detail work.

Ken,
Thanks for the link, the stock of cutters with the machine already includes a set just like this.

Today we didn't do much, but clarified a lot.

Firstly we identified why the X axis seemed tight at the -ve end, turns out the bridge is formed by two aluminium extrusions located by end fixings in slots in the end plates.
The extrusion separation at the -ve end was smaller than the +ve end, causing a jam-up as the spindle carriage approached the -ve end limits.

Then we squared the platen by cutting the side of the X-axis, and marked out in pencil a 100x100 matrix on the platen to aid future job positioning.
There is a suggestion to cut the markings into the platen.

The subject of understanding the UGS controls then occupied some time.

There is a UGS Homing button, but it is disabled in the default setup.
Left as disabled pending getting final limit switch arrangements (with a new control cabinet etc).

All the UGS move and reset controls relate to Work Origin, not Machine Origin.
There is no way (that we can find) of resetting the machine coordinates to 0,0,0, the machine must be driven to that location by manually jogging it with the UGS machine control buttons.
If Work:0,0,0 and Machine:0,0,0 is required simultaneously, then a power-down of the Arduino is necessary.
The sequence is:
- power down,
- move the gantry to desired location,
- power up, the sequence re-sets Machine coordinates to 0,0,0,
- on UGS select Reset Zero control - Work coordinates now also 0,0,0.

Much discussion among the few Shed members interested in the details and a new level of understanding developed.
We are in much better shape to do future jobs.
John.

Originally Posted by electrosteam

So Plan B kicked in and the platen was surfaced.
Very interesting to observe minor chatter when cutting in the conventional direction, and absence of chatter when climbing.

John

This may indicate a bit that is not parallel to the surface. Particularly if the bit is of a large diameter.

Bob,
I will have to explore further the effect of a non-parallel bit on chatter.

Our machine is shaking down nicely with operation now quite predictable and our understanding of the UGS controls now seems complete, but, the "Return to Zero" has been demonstrated to have a problem.

It appears that when "Return to Zero" is selected:
- Z axis moves in +ve Z direction the amount currently shown as the Machine Position for Z,
- rapid move in X and Y to the location shown as Work Position X=Y=0,
- Z move to Work Position Z =0.

Eg: if current Machine Position has Z=15, the machine goes first to Z=30.

This can drive the Z into the end stop (our machine currently is without the Z limits).

Just another bug in UGS to be aware of, prompting the urgent need for us to get our Z limits operating.
This is a clear demonstration of the need for a control to zero out the Machine Position Coordinates.

It also explains why we were having such a hard time understanding the machine response to UGS controls.

We tried without success to run a complex CamBam generated file produced at home by one of the Shed members..
The Gcode file provided had X and Y offsets of 4400 and 2600 respectively.
We tried in vain to find something in CamBam that would reduce these offsets.
The Shed member will re-do the code generation at home and investigate further how to remove arbitrary offsets.

John

weird behaviour for the Z axis John. Does UGS have a console that lets you view the gcode commands being sent? might be interesting to see what is sent when return to zero is set. Normal behaviour would be to move to the Z safe height, then traverse X & Y, then Z zero. Is there an option somewhere in UGS that allows you to specify the Z safe height?

sorry for the double post.
this may be vaguely useful:
https://discuss.inventables.com/t/return-to-zero-issues-possible-ugs-issue/20083/2

basically it looks like you need to set G28.1 before you hit the return to zero:
from https://discuss.inventables.com/t/ugs-losing-zero/15817
The UGS "Return to Zero" button uses the G28 G code command. This is a position that is held in the Arduino EEPROM and doesn't change even when the machine is powered off and back on. To use the "Return to Zero" you need to set the position for G28.
Power on you machine, home the machine, jog to the location that you want to be zero and enter the G code command G28.1
This sets the position in Machine space. Now any time that you press the "Return to Zero" button the X-carve should move to the location that you set.

foobillious,
Thanks for the links, at least I now know our observations are not crazy.
You also did better than I did on the searching, I looked at the Inventables site but did not pick up your results.

We also observed attempted Z+ve movements in the 100s of mm.
The value seemed to increment each time the command was issued.

I remember the console streaming output on UGS showed the extreme movement, but I cannot remember if a G28 or similar code was issued.

Grbl includes parameters for a safe height and a clearance height.
One would think these heights would come into play for a Return to Zero movement.
The documentation describes retention of the G28/G30 commands, and I will investigate those next week.
The incrementing Z height values observed could indicate a G28/G30 command issued each time.

Personally, I would think it unacceptable that the operator (one of our Shed members) should remember to do something particular in this case.
The UGS screen button should do the logical thing.

I noted earlier that UGS has a Set Work Position to Zero button.
Surely, any usage of G28/G30 would be included in this command.

I will investigate if there is a way of implementing a start-up procedure that ensures the various retained values are correctly set.

I will be watching for the next version of UGS with great interest.

(Contact with the machine only on a weekly basis really sucks !)
John.

I believe its possible to set macros up in UGS, but I don't use it.

foobillious,
I have tried the macros and I think there is quite a bit of potential in their use.

I loaded my Duinotech UNO, a Jaycar clone, with Grbl 0.9j and UGS 1.0.9 on my laptop at home.
This seems to work fine, it responds to commands and increments the machine position just like the router in the Shed.

I have noted the console display of all the UGS commands:
- Reset Zero G10 P0 L20 X0 Y0 Z0,
- Return to Zero G90 G0 Znn, then, G90 G0 X0 Y0 Z0
where 'nn' varies as the suspected Bug interferes with correct operation.
No sign of a G28.1, and the G10 will need some time to study.

I have tried some G28.1 and G28 commands, but the Bug still interferes.
Fascinating watching an ever increasing value for 'nn' as the Return to Zero command is repeatedly issued.

Also, note that once Z has got to its commanded height, the path to Zero is a simultaneous movement in X, Y and Z.
One can watch all three coordinates decrementing at different rates until all three get to zero at the same time.
What is the point of having a Z height clearance value if you then ramp it towards Zero ?

I will do some more testing tomorrow (if I can find the time).
But I have done enough with this to regard it as either a genuine Bug, or a crazy unworkable arrangement ( for a group of inmates in a Men's Shed ).
Perhaps it is time to look for a UGS replacement.

John.

Did some more testing with same result.

Searched for other reports and found that the UGS code writers have identified "Return to Zero Runaway" as Bug #308.

https://github.com/winder/Universal-...-Sender/issues

Now to write up a procedure for Shed members to follow that does not stir up this bug.
John

I'd regard that as a bug. UGS is open source, so I guess you could let the devs know and hope they fix it, or fix it yourself and send them the fix to incorporate.

There are a number of other options beside UGS. Chillipeppr is one (and there is an offline version now, which I use with my tinyg). ChiliPeppr - Hardware Fiddle

grblgui looks nice:
https://github.com/cody82/grblgui

and octoprint apparently supports grbl now:
OctoPrint.org

apart from chillipeppr I've not used any of them. now you have some test hardware you could give them a trial run