Manual Y Calibration and Y Offset.

[bergerud]

Has anybody tried calibrating the Y and the Y offset by reading the calibration instructions? As was with the X calibration instructions, there seems to be mistakes. Read on and correct me if I am wrong:

Calibrate width option: The instructions say to calculate the difference between the actual board width and the machine measured board width (6.1-6.05=0.05 in the example), choose "narrower" or "wider" and key in the difference (wider and 0.05 in the example). These instructions assume that the current value is 0.000. In actual fact, the number you key in will become the new value by replacing the value that was there. For example, if the current value was 0.04 and you expected to add another 0.05 because the board was being measured 0.5 too narrow, you would get 0.05 not the 0.09 you need. You have to add the two together and actually set it to 0.09.


New value = Current value +(real width-machine width)




Not only are the instructions wrong, but the programming ideas of narrower or wider further confuse the issue. If you need the board to be measured wider, you need to add the difference to the current value, while if you need the board measured smaller, you have to subtract the difference from the current value. The narrower and wider is supposed to take care of the negative sign for you but that only makes mathematical sense if the current value is 0.000. (You only get the negative of a number if you subtract it from 0!)

Calibrate Y-offset option: The instructions to change the Y offset suffers from the same mistake as the calibrate width option. The new value must be added or subtracted from the current value. The "toward" and "away" only make sense if the current value is 0.000. There is another simple mistake as well. In the example, the offset is determined by subtracting the two side offsets of a rectangular carved region. It is implied that this difference is the amount that the Y offset needs to be changed by. Well, I think it should be half of that! Imagine the right side is 3/4" and the left side is 1/4", then the difference is 3/4"-1/4"=1/2" yet it only needs be moved 1/4" to be centered.


New offset = Current offset+(right offset-left offset)/2


(Of course, there is an easy way to use the "narrower", "wider","toward", and, "away". Simply change the current values to 0.000 before you even start!)

[RMarkey]

I think I intended the cal values to be reset to 0 before calibrating. I see how the offset doc is contradictory, saying 'subtract the difference' then also saying 'adjust by xxx', the latter being correct.

[bergerud]

Thank you for responding. I have to admit that I am confused by the recalibration process. Before the manual Y and Y offset calibrations, I went through the calibrate the edge sensor routine. When I did this, it not only changed the Y and Y offset values but also their defaults. (I had assumed the defaults were factory set and could not be easily changed.) One would hope that after the edge sensor calibration, that the Y would be calibrated. In fact the Y measurement was now off even on the board and cuts it just made. What was the point of the edge sensor calibration if I have to change the values? I do not know if the machine is really this difficult to calibrate or if I am just missing something here. I have never been able to get the machine to center in the x direction. I would be interested to know how the machines are calibrated at LHR.

[RMarkey]

Machines are calibrated with 2 kinds of aluminum plates measured to 0.0001". Various holes are positioned in the plates and a sharp metal bit is used and wired to the plate to detect the slightest electrical touches.

For X, the techs go back with a wooden board to test the cal process. They may do a manual X cal to get the measurement within 0.050".

The edge cal was the original user cal, with manual overrides added afterwards. Generally, the edge cal should get the machine within 0.020".

[FWMiller]

I know this is an old thread but I tried to do a y-offset calibration for the first time on my machine. Naturally I went to the Carvewright website and downloaded the calibration instructions and followed the instructions. I toyed with it fruitlessly for an hour before giving up and searching the forum to find this thread. They do not give you any option to calculate your own value and plug it in. Plus, unlike the other cal values, the "default value" is overwritten as soon as you store a new value. I found that you can not plug in a negative value so the only way to get a negative value is to choose the correct direction and plug in the right number and let the firmware do the math. At one point I though I had figured out that the machine took the number that I typed in, divided it by 2 and then subtracted or added it to the old value depending on whether you chose closer or farther from the keypad. However the results I've gotten since seem to indicate that this is not the case. It's definitely not doing a simple add or subtract. I've managed to get the value back to zero and then following the process using the measured difference, half the distance and both directions and I still can't get it centered in the y. I either get continued increasing error or it keeps bouncing back and forth from one side of center to the other. Can some tell me what the best method is to get this correct? And how close should I be able to get to centered in y? When I started I was off by about .070 on a 5 inch wide board.

Thanks,

Frederick

[bergerud]

I have not done it for awhile. I think understanding the first post on this thread is your best shot at it.

[FWMiller]

I think I understand pretty well the concept in the thread. What I don't understand is what the machine is doing with the number I enter. If I knew what calculation the firmware is doing it would be easier to figure out what to enter. Because the value that the LHR instructions tell you to enter is wrong. If the value I read in the calibration offset is simply an offset number in thousandths of an inch then I can easily calculate what the value needs to be. Then I'll just have to plug in numbers until I get the stored value where I want it.

[bergerud]

I think the simplest way is just to set the offset to zero. Drill a hole, measure how much you need it to move and use that as the new offset. (I am having a dim memory about rebooting the measuring process. I think you have to release the board and re-clamp it before you remeasure.)

[liquidguitars]

I did this not long ago and the " near of far" makes no sense to me... I just had to guess and input numbers.

[want2b]

Noticed there was a difference of the sides when I did the 2 sided carve that 'Berg' had contributed of the 'bullet'. At first assumed it was something I had done in the cutting of the boards. Did another 2 sided carve and found I had the same problem(haven't done many 2 sided). When I spotted this post was when I learned of the 'Y' axis offset, width always measured OK, didn't have enough understanding/experience to realise there was another option. Tried to follow the instructions that I also found under 'calibrate' on the Carvewright site and as noted above by others my first attempt moved the carved area too far. I also did not reset to default before doing carving the 'square' and using it to measure the difference from each side. I also feel that the number that should be input is 1/2 the measured difference. Unfortunately while playing with it all my keypad stopped working on the 'middle horizontal' numeric row. Can work around a lot of things by using the arrows but it will not accept anything in the numerics without the . first. Will post after I get that fixed. Meanwhile, the calibration document definitely needs an update for the next user.

Rick H.

[want2b]

Tried my machine this AM and everything seemed OK, later in the day after it got warmer and more use(strictly playing with the calibration issue) it has again started to not respond to the same row on the key pad, deal with that later.
When I checked my settings they were 'default .069 and current at -.002. I ran a test square carving 1" wide by 5 " wide on an 11" wide by 30" long 3/4 board. Found there was .050" difference from the carve to each edge.
Best I can figure the toward/away option is in relationship to an 'assumed' center/zero. The only use I can think of for the default is to get you back to a known setting after you do a lot of adjusting. I did NOT have enough nerve to reset anything, just worked with what I found. What I feel is that if you need to adjust you should move according to the current setting. What worked for me eventually was realising that my math needed to be based on where I wanted to end up, not on the Toward/Away option.
If a Default is .060 and a Current is .040 and you measure a need to move in the direction of the keypad of .030 then you would think it would be a Toward option. Any input using the Toward option will end up with a -(negative) result, Away entries do not result in a (-).
The actual input should be an Away With the entry of .025. Using .015 as 1/2 of the .030 distance and moving back from the Current of .040 setting still leaves it on the Away side of the base center/zero. Basically the Toward/Away do not mean the direction of movement, just which side of the center.zero you need to end up on.
Not sure why but as I keep playing with the numbers my Default keep changing. started at .069(an Away #) and has ended up at .054(still Away). You may have to play with this to make it work, my final with a digital micrometer was within .002 side to side but that again is based my pressure squeezing on pine.
Thought about resetting everything but wasn't sure what all I might affect so I wimped out, hope this helps somebody.

Rick H.