curtisa

Just having a look at it now. In the mean time try running the attached file. It's just the pickup and neck pocket components merged into a single Gcode, with the pickup cavity running first. Run it in the air, just in case there's something I've overlooked when I've joined them together, and your machine tries doing something unexpected (the sim in Camotics looks OK though). Rename the file extension from 'rtf' to 'nc' (forum software won't let me attach an nc file) Pickup + Neck Pocket Test.rtf

Post your pickup Gcode. If that works and the neck pocket one doesn't and nothing else is changing between running either, it has to be something different in the Gcodes between the two.

Just need to insert N115 in the position shown in the quote (after N110, before N120).

I'm inclined to think it's something else. If you look at my second diagram of your movement, the X axis will actually go east and west. The problem you appear to be having is that it won't go west far enough; the axis goes west as far as 0 and doesn't go any further. If it were something like step loss due to incorrectly set pulse widths and/or timing you'd expect to see the axis drift by random amounts in random directions when running any Gcode, not just this one neck pocket routine.

That's really wierd. Almost looks like there's soft limits truncating the movement of the X axis in the negative direction. If you accidentally positioned your Gcode within the boundaries of your machine such that the workpiece overhung the negative X limit I would've thought Mach3 would give you an error on startup saying that the Gcode was larger than the maximum travel of the machine (or something to that effect) and prevent you from running the program. Might be more of a question for @MiKro - is it possible to accidentally set up Mach3 such that it will ignore the soft limits of the machine, and instead of halting the program it runs it anyway? It is worth exploring though: All movement on the Y axis in your Gcode is in the positive direction (the cutter only goes 'more north' or 'less north'). Z axis only goes negative towards the floor, with the exception of a couple of movements where it is +0.2" above the work piece while repositioning (can you confirm that the Z axis actually does rise upwards above zero when it does these repositioning moves, or does it also get clipped at Z=0?). X axis is the only one programmed to spend equal time moving positive or negative of zero. I'm thinking there may be something funny going on with the zeroing of your workpiece and machine before starting the program, and the lack of defined coordinate offsets in the Gcode header (see point 1 in my earlier post). What happens if you take the first few lines of your Gcode: And insert an extra line N115:

I concur with @Bizman62, I don't think you're going to get that out without taking the fretboard off. Even assuming the rod is loose enough and you can get a good purchase on it to pull it towards the nut adjuster access hole, you'd still need to get the rod past the end lip of the cavity where the allen key fits in. That'd involve trying to bend the rod upwards over a very short distance as you draw it out, which will be no fun at all and could result in splitting the fretboard where the rod is being forced to bend. And then you'd have to reverse the process to push a new rod back in again. Fretboard off or scrap the neck, I'd say

So instead of doing this: It's doing this?: Ie, the machine doesn't want to go to any position where the X coordinates are less than zero and the motion gets 'clipped'?

Some thoughts on the code. I don't know if they have an impact but they stand out to my eyes: The Gcode doesn't define a coordinate offset system. I'd expect there to be either a G53 (machine absolute) or G54 (user defined) mentioned in the header somewhere, but there isn't anything. Not saying it indicates a problem, but it looks odd. It opens fine on LinuxCNC as it is and looks OK on the backplot. On line N110 the command G70 is issued to tell Mach3 to use inches as the units of movement. On line N100 G20 is also issued to tell Mach3 to use inches. The Mach3 Gcode reference suggests that G20 should be used in preference to G70. Can't imagine it makes any difference if essentially the same command (although may be is issued twice On Line N100 G90 is issued to tell Mach3 to switch to absolute positioning mode. On line N110 G91.1 is issued which tells Mach3 to use incremental mode for arc movements. As your Gcode doesn't use arc moves anyway (no lines prefixed with G2 or G3), you'd think that requesting mixed modes of positioning would be handled OK by Mach3 and it would just ignore the G91.1 command, but who knows? On line N140 G43 is used to initiate a tool length compensation of 0.8". This shouldn't affect the positioning of the X axis (it changes the apparent position of the Z axis while cutting to compensate for the amount of 'stick-out' the tool has below the collet), but is largely redundant in your case as I don't believe you can insert the cutting tool into the collet of the spindle at a predefined, repeatable length every time. Your programmed feed rates for the bulk of the cutting moves is 80 inches per minute. If I'm reading your earlier comment about the stepper motor settings you have a maximum feed rate in Mach 3 of 1500 (I assume this is mm/min?). 80 IPM is 2032 mm/min, which is faster than your maximum axis speed in Mach3. Mach3 should just limit the maximum speed while the Gcode is running to 1500mm/min if asked to do more than this, but maybe something odd happens in Mach3 when the axis velocity limit is reached while running? At the end of each spiral clearing pattern the program pulls the cutter up 0.2" above the surface and does a rapid (G00) move back to the start of the spiral and starts again. Maybe the X axis is over-travelling while doing this return move? Try lowering your acceleration speed in Mach3 for the X axis and see if that changes things. Do you have another Gcode for something similar that you know works reliably to compare with? Maybe a pickup cavity?

So you're finding that it starts off and does one spiral clearing pattern from the inside out, and when it returns to do the next spiral pattern at the next depth down it attempts to start again shifted right, and so on and so on?

I just meant, 'do want me to have a look at your Gcode file?'

Then it's looking more likely that there's some subtle difference in the way the code is being executed, rather than an issue with the driver hardware, electrical installation or mechanical foibles with the machine. Shifting the toolpath to one side by the width of the bit is interesting. Incorrectly defined cutter compensation perhaps? That could also explain why your code provided by the other user runs fine on their machine but does oddball things on yours (cutter compensation based on any tool tables you have set up in Mach3 would be particular to the end user, rather than something universal within the Gcode). Care to share your routine?

Have you tried running the troublesome routine in a CAM simulator to see if the creeping axis is occurring in the code rather than in your machine? I use Camotics to proof stuff before committing it to the workpiece: https://camotics.org

The hardware you're using to drive the stepper motors will have some kind of minimum specs for the incoming step/direction signals to satisfy in order for the motors to reliably rotate from step to step. If you're lucky your drivers will specify these minimum parameters in the datasheet. Something like this: Each driver accepts two signals to define the motor's motion - the step and direction signal lines. The step signal controls the rotation of the motor; every pulse that appears on the step line advances the motor around by one increment. The more pulses the further around the motor will rotate. Steps coming in more frequently translates to the motor turning much quicker. The direction signal controls whether the motor rotates clockwise or anticlockwise when the step signal is pulsed. If the direction signal is 'on' the motor may rotate clockwise, and if 'off' it will rotate the other way. The above table highlights some minimum timing requirements for each of the two signals: The step signal when it pulses 'on' must be on for a minimum time in order to be reliably registered by the driver as a step command (labelled as 'Pulse duration, STEP high' in the table = 1.9 microseconds). Anything shorter than this specification of 1.9usec may result in the driver not seeing the pulse and a steps can be lost. The step signal when it pulses 'off' again must be off for a minimum time before the next pulse comes along ('Pulse duration, STEP low = 1.9usec). Anything shorter than 1.9usec may result in the driver not seeing the subsequent pulse and a steps can be lost. The direction signal needs some time after being turned on (rotation = clockwise) or off (rotation = anticlockwise) before any subsequent steps commence on the step line (Setup time, command before STEP rising = 650 nanoseconds). Anything shorter than this specification of 650 nsec may result in steps being lost. The direction signal needs some time to remain in its last stable state before changing again if you want to make the motor begin rotating in the opposite direction (Hold time, command after STEP rising = 650nsec). Anything shorter than this specification of 650 nsec may result in steps being lost. The above parameters should be configurable in Mach3 for each axis/motor (I'd be surprised if you can't), but I'm basing this solely off LinuxCNC which does allow the user to configure this kind of driver information. Ideally the software should be configured with timing information no lower than the above values, but for reliability you'd probably use values 2x of those presented in the table above. Naturally, if your driver has different data use that instead - you may be able to pop the cover off the top of the CNC interface box and see if you can locate the driver ICs and do a search for a datasheet on Google based on the IC part code. Probably a bit more info than you really needed...

I think we're talking about two different kinds of 'calibration'. Mikro's talking about checking the number of steps required to get an axis to move a known distance. I'm talking about satisfying the minimum requirements of the hardware in order to guarantee correct motor performance; that each pulse generated by the software is correctly adjusted for timing and spacing to make sure that the motor advances a step without losing pace.

Initially I used parallel, but I'm now using ethernet with an external step/dir pulse management card. Whatever came with the original CNC3020, probably not too dissimilar to yours. I think TB56-somethingsomething gets mentioned fairly regularly with regards to these cheap generic CNCs. Actually, that's another thing to investigate - check that your step/dir pulse lengths are configured sufficiently long enough in Mach3 to allow each motor to fully advance between steps correctly. That can also be an indicator of lost steps when under load.

The other option if all your toolpaths work fine except for the neck pocket is that the neck pocket code has something in it that upsets the X axis zero. Maybe you have a couple of moves programmed in G91 incremental mode that aren't being cancelled properly, or a G92 coordinate offset accidentally enabled? I'd regard total replacement of the motion control hardware as a last resort. Replacing the entire drivetrain of your car because one tyre goes bald quicker than the others seems a bit excessive.

Could be mechanical - lost steps or a loose motor-to-leadscrew coupling is a possibility if the drift is only in one axis. Is this neck pocket routine particularly aggressive in terms of the cuts being made? Are there lots of hard corners in your routine where you're asking the cutter to make 90 degree turns while deep in the workpiece? Backlash is usually only going to show up as a consistent small error when reversing the direction of an axis; squares that come out looking like rectangles or circles that look like ellipses. But we're usually talking about a mm or so of difference at the most. Half an inch of loss of positional accuracy in 30 seconds of motion is unlikely to be due to backlash.

Buckeye burl is too soft and fragile to be used as a fretboard. It won't withstand the wear of being exposed to guitar strings and fingers, nor will it retain the frets once they're installed. Pretty sure cocobolo is used regularly as a fretboard, but its rarity makes it much less commonly seen. Bocote is another related timber that makes for another exotic fretboard option. There's far more options out there for fretboards than just the five you list. You're not looking hard enough. A lot of Australian desert hardwoods make good fretboards - Jarrah, Gidgee, Cooktown Ironwood, River Oak. I used a piece of Merbau decking as a fretboard once that worked quite well. Strictly speaking, any timber that has a high stiffness and hardness, and good stability once dried and cut should work as a fretboard. Maple, ebony and rosewood just get the spotlight more due to their plentiful supply (or at least before widespread over-logging started endangering supply) and the expectations of customers and manufacturers that 'traditional' timbers be used in guitar construction.

https://www.mylespaul.com/threads/does-a-straight-truss-rod-channel-really-work.406227/ The discussion here (post #3) perhaps explains better how the straight channel single acting rod worked for Gibson. The key appears to be that the body-end of the rod is set further below the fretboard than the nut-end. When the strings are tensioned up the neck will want to pull forward into a slight curve as you'd expect. Because the rod is placed at a slight negative angle relative to the curve of the neck, tightening the nut wants to pull the nut-end back away from the curve. Hence you have your adjustability to compensate for excess forward bow due to string tension. The double acting rod (and the whacky single acting one I found in my Strat neck) with the flat bar welded on top just simplify the installation process as they no longer rely on a channel that needs to be cut at a falling angle (Gibson) or some complicated curve (Fender) to work.

I pulled a single acting truss rod from an old Strat copy neck as a practice run for steaming off a fretboard a while back. It looks almost exactly the same as a double acting rod with the flat steel bar on top of the straight threaded rod, the difference being that the adjuster nut can only be done up to bend the rod backwards. Undoing the nut beyond the slack point just results in the nut falling off the end. Long story short, straight single acting rods exist, but the mechanism for making them bend the 'right' way is different to the traditional rod sitting in a curved channel. Whether the period-correctness is important to you, that's entirely up to you.

There's something weird with the raw http link provided for those two missing pictures. The link address itself is about 3 times the length of the other working pictures. Working one, for example, reads as: https://lh3.googleusercontent.com/pw/ACtC-3cnswCYVpLO2eTOG4S7qVTRDIArqNwovHAKFrFG_ySXLL6-yb9e46pvTZ7q42iuI_s9-p_dok3ILAuJuZtgT5hYO3A5vBkzJ0tpLleqyUFAmEp2Nm6h0QJWKsXgvOSQWqjjnEUBWMJou281y02RH9Ag=w947-h1262-no?authuser=0 Broken link reads as: https://lh3.googleusercontent.com/deo2B6Qytte57tL1B0rDPX2yZ9itBh4sxrpUI328Gh3KM9Oo9zMrYDwjDWT--dYiPfSVhGleRtytnr3L6jBcKhM3gSTI77LdSUBUGPHVB2fUa0067f8ncclptAxgnRpghF35Jyggaq0tebMteUZiKnqGF1AN6KXj3rJ-Z2r8pzqrZR3opJhuNCXF1RGruXgsc7vpTbuIlK9ZMTSRvhWPJ4iUCkoYh58OA6Z55aSL_8bJ-1YE30PSOhTM6q5I5Dj9lP2veAKW_xGbfArVApM7_4x34KFLxeIxqrd6r0vgy-MtfsL9Sk2HkJB-BBzMM0kCTTkdNijcujtUWgJNt0-BuqrEZFApHqi-URqmI160CMPzoFKsfoJ469lOa-gL8snoUmJRDlkhl1CmLHIs8C_MXXTwDfSnyEK4VuH3GjaeRqswAdI1rG0ANsG5fMhYjMAvtQsSSJjdrCOvV1gvf7Q9GWc972KtfUB8y1aFaFkb0bMOjETV1RvgKiyZNB1n5R7HIVjtcC7zVRxkYkVVH53Jou2th_APf9YrUf4M9slOy4-YjrUr4iWfh_2SCk_WgYEItGCAp4vlw051Zo5og54TkYGkbrGv9QhcWKMf5fqYfx3fsAyAEZ41jYNDZlOPx4dTBoddN-pyo3SFW0SpKJAOsh-phsQa8Vibtq4syyUnkXLFQd7KihmkX8DCw0HU=w1683-h1262-no?authuser=0 Clicking on the broken link takes the user to the Google account sign-in page.

Feed a little fresh solder onto the tip of the iron as you heat up the old joint. The freshly melted solder adds more surface area to the tip and helps heat transfer from the iron to the joint.

Dust is best collected as close to the source as possible. That means running tools that generate large quantities of fine particulates with dust extractors or vacuums. The mask should only really be relied upon as a backup to whatever other dust management you're using. If you can smell the dust you're generating while sanding either the dust isn't being collected effectively at the tool, your dust mask isn't fitting correctly or the dust mask isn't rated to filter out the particles you're generating. Is the fan drawing the air from the basement to outside? If all you're doing is using the fan to blow across your work area inside a sealed basement you could be doing nothing more than stirring up the dust into the air and recirculating it around the room. Any power tool that has a dust collection port should be used with a dust collector of some sort, particularly when in an enclosed space. For hand sanding, wear a mask that fits correctly and is rated to filter out dust and (if possible) do your sanding in an area with good ventilation. Keep your hands where you can see them (don't place your fingers underneath the block of wood you're cutting with the jigsaw). Any bladed tool that has the potential to come in contact with you must be operated in the way it was meant to be operated (don't stick a circular saw in a vice upside down and use it as a table saw). If the tool has a handle for both hands, use both handles provided (don't hold a power planer with one hand underneath the sole). Keep the pointy end away from you (don't hold the wood you're chiselling with one hand in front of the blade). When you power up a cutting tool make sure it's not in contact with the workpiece before it has reached full speed (don't leave the router in contact with the timber when you switch it on). When you power down a tool, let it come to rest before doing anything else (don't put that spinning router back on the workbench while the cutter is still spinning down) Keep your work area clean. Know your limitations and seek assistance where required (read the fricken' manual) It is not a toy. You will be fine

Ok, cool. Then your options are either: To make the pocket deeper AND introduce some neck tilt to keep the action low - see @Andyjr1515 and @Bizman62's posts above for the general process. The drawback is that it's hard to execute cleanly if you're unfamiliar with the tools and processes involved, but there's less work required in finish repair after the pocket has been re-cut; or, Remove material from the back of the heel to get rid of the excess bulk. The upside of this method is that you don't need to change the pocket or setup of the guitar, and you can use whatever hand tools you have at your disposal to thin out the heel. The downside is that you'll need to do significant repainting of the area if you want the re-shaped area look original.

But for what purpose? Lowering the neck into the body will have ramifications on how the guitar will need to be set up to compensate for the change in relative geometry of the neck, bridge and strings. While there's solid information being provided to you as to how to do it, it's not known what you're wanting to achieve by doing it. If all you want is the string action a little higher, there are easier, less-destructive ways of doing it that don't result in having to do major surgery to the neck pocket. If you're bothered by the look of the neck sticking up above the body of the guitar as much as it appears to be, then yes, the neck pocket can be lowered. But you will need to do some corresponding adjustment to compensate for the extra string height that will be introduced by doing so (unless, of course, the extra string height is acceptable for you). This is where the advice from @Andyjr1515 and @Bizman62 regarding angling the neck pocket with shims comes in to play. You'd also need to weigh up if the effort involved in doing so is worth it just to change the appearance of the instrument. If the heel area feels too chunky due to the neck pocket being too shallow you could always remove material off the back of the body around the heel to make the neck/body transition 'feel' slimmer, but you'd have to refinish the area to restore the look. Knowing a little more about your requirements will help us provide you with some advice on how best to achieve it. At the moment the suggested fixes (while perfectly valid) seem to be a little like rebuilding a car's suspension system to improve the road handling ability, when all that may be required is a little more air in the tyres.