Two Sided Guitar Neck manufacturing

[scott_lipsham]

hi all

Needing some advice with respect to two sided milling.

Trying to complete my 'first' CNC guitar neck but second (back) side is different to the first (fretboard) side. They are offset to one another. The top surface of the guitar neck is off centre compare to the back side. hope that makes sense.

I have posted this on a fusion gitar forum and seems to be a common issue.

i'll try to explain.

The attached pictures show two neck models centred in my stock.

Registering pins in the stock.

I use the same CAM path to machine registering pinholes in my 'fixture'.

All my measurements indicate that everything is centred.

My process (at this stage) is:

• centre and secure the blank on fixture with tape/supaglue
• machine reg pins; then face; then trussrod slot; shallow outline contour to headstock height; rough parallel headstock; finish parallel headstock.
• turn blank over; locate on pins and glue down with tape/supaglue,
• machine tuner holes; outside contour to full depth; rough shaping across the neck; finish shaping pass along neck;

Take it off for the big reveal and the trussrod is well off centre (about 2mm or so)??

When i do the CAM for the top i hide the 'back of neck' model and set up all the CAM with visible fretboard sided model. Then vica versa for the back of the neck.

i have two different setups top & bak.

I was wondering whether i could put BOTH MODELS In the same setup??

All my thicknesses (z values) are pretty bang on eg headstock thickness, heel thickness etc.

My origin for the topside is centre of stock bottom.

My origin for the back side (inverted model) is centre of model bottom. Makes sense to me since the fretboard gluing area (top face of neck) is now sitting on the fixture.

I home my cnc at the start of each session. Jog it to the same offset centre of fixture and zero x&y as the origin. It lines up same point each time.

Sorry for the long post but it seems to make sense to me (newbie) but something is obviously not correct.

Any thoughts anyone? hoping there might be an explanation

Thanks in anticipation

cheers

[seth.madore]

Could you do a couple things for us?

1) Share your Fusion file, it will make assistance much easier.

File > Export > Save to local folder, return to thread and attach the .f3d/.f3z file in your reply.

2) Please clarify why you have two models shown in the image. Are you making two different parts, or did you duplicate the model to reflect the two different operations that needed to be done to them?

[scott_lipsham]

hi seth. many thanks

will do

cheers scott

[chrisQX8DX]

hello fellow fusion luthier 😄

a couple of things come to mind:

what kind of CNC do you have ? are you using Mach3 ? Are your axises and steppers calibrated precisely (so the machine moves exactly the intended distance on the X and Y axis?

if i understood correctly you are already making sure that you locate your stock with pins so that when you flip it around it is in the same place exactly and don't have offset due to  glueing placement inconsistency ?

You are using 'stock from solid' and not relative size box for both sides ?

You are using the same relative box/stock origin for both sides  so the zero point is the same relative to the two sides you are machining ?

If all of the above are true then i would guess you have the same issue i used to have and bang my head against the wall trying to figure out the issue for months. I would guess that it is not a Fusion issue but an issue with lost steps on your CNC. I use a very old machine and it cannot handle certain fast moves. My machine also starts up with the steppers in slightly different position and even when homing all axises the zero point is never 'exactly' the same. I also used to have a shielding problem of the wires for the steppers and that was generating a lot of noise which could also interfere with precision. (i use Mach3 and could see a lot of spikes in the signal). But once the shielding was solved the issue was better but still present.

What i finally understood is that i have to avoid switching the CNC off and on bewteen sides (so i either do both sides on the same day or just leave the machine on over night so the steppers don't move). Before i thought that if i just home my axises the the zero point would be exactly in the same spot but i think my homing switches or steppers just weren't precise enough for that. So after i figured that one out the problem was solved .. but only sometimes, some parts machined perfectly aligned on both sides .. and then some others didn't .. again bangin my head against the wall for weeks trying to figure out why oh why 😆.

Then finally it dawned on me that my machines steppers 'loose steps' when trying to move too fast on the X and Y axis at the same time and especially when trying to go through too much stock at a time (one critical area are the corners of a pocket, lets say you are milling a 6mm fileted inner corner with a 6mm mill, when it comes to the corner a larger area of the flutes is engaged with the stock so it take way more force to go through the material, especially in hard maple and ebony this used to make the steppers loose steps). So i did a lot of experiments with different tool path strategies and speeds and finally got to a point where i know what speeds, depths and moves my machine can handle without loosing steps.

I now finally have perfect alignement for both sides consistently just by managing the speed and complexity of the tool paths. It's a difficult problem to figure out especially in wood where to slow a speed can burnish or burn your stock. For me Mahogany was the best wood to do the tests on since it doesn't burn easily. I just did a series of tests of a sample part with complex toolpath at different speeds until i found the max speed i could use without my steppers loosing steps.

Maybe this helps somehow ... if you set up and reference you stock problem i really doubt it's a Fusion problem.

[scott_lipsham]

Hi again

the file is a bit of a mess so hope it makes sense.

the two diff models reflect th etwo diff operations - top and back.

so....in the NewNeck_Top setup i make the 'top'of the model visible, do all the cam with the centre bottom stock point as origin.

next .. in the NewNeck_BACK setup i hide the 'top' model and make visible the 'Back' model and do all the CAM for it. In the second setup though the already machined top (ModelTop in previous setup) is now Model Bottom which i use as the origin or reference face for the seond setup.  HJope that makes sense. 🙂

The issue is that when it all done the çentered Truss Rod slot is well off centre compared to the back.

A lot of other luthiers use the bottom left corner as origin. Just doesnt make sense to me when everything seems centred? My heights are good in the finished product so it seems just an issue in X&Y.

file attached. really appreciate the support

regards Scott

[chrisQX8DX]

---

@scott_lipsham wrote:

I use the same CAM path to machine registering pinholes in my 'fixture'.

---

another thing just came to mind:

when you are drilling the registration holes in both sides of the stock, do you drill all the way through the stock or do you drill first one side reg holes then flip the stock and drill the other side holes ?

That was a problem i used to have that when drilling from both sides the reg holes weren't exactly in the same spot. i now drill the registration holes all the way rthrough the stock from one side so i'm sure that the reg holes on both sides are in the same place.

[chrisQX8DX]

just checked your file and your origins are off .. my screenrecorder isn't working but if you just select the NewNeck_TOP and NewNeck_BOTTOM setups you will see the origin moving.

This is because your are using 'Stock box point' as origin on one side and 'Model box point' on the other side. On the finishing pass of the back your using the 'Tool orientation' option to set another origin .. why ?

Also you are using 'From precedent' as stock on the other side, i would use 'Stock from solid' on both sides.

so i think it is a fusion issue after all, i never use ''Model Box point', always 'stock box point' .. personally i always use the corner of the stock and then flip the axises. I then just rotate the view. If you want to rotate the stock in manufacturing then look into 'Manufacturing model': https://knowledge.autodesk.com/support/fusion-360/troubleshooting/caas/sfdcarticles/sfdcarticles/How... that way you can set up two different models for each side, maybe easier than rotating the view. i use both methods since manufacturing models is a cool new feature, but i'm used to just rotating the view and flipping the axis/boxpoint so i use that most of the times.

There is no need for two different models, work of the same model and either rotate the view or create a manufacturing model and rotate the model there. IMHO rotating the view and just moving the box ref point to the diagonally oposite and flipping the axis is the easiest less prone to alignement errors method.

i corrected the setup, you still need to check the toolpaths but at least the alignements are right: https://a360.co/350TGUA

[scott_lipsham]

hi chris

and THANKS heaps!

ill start off with a Tele he said...it will b esimple he said.....faaaark! 🙂

i bought the model as a quickstart and found a few errors but so didnt notice that the mopdel is off centre.

nice pick up thanks.

didnt make any sense too me.

i have a generic chinese QuickCNC with a RichAuto DSP controller. ive used MAch 3 on a CNC i made at the school i teach at so am familiar with MAch but am quite enjoying this DSP thing. havent had any issues at all until now. done some fretboards all very accurate so was really hoping it wasnt a machine issue - user error again 🙂

 i used the model point on the second setup because i though that would b eeasier to reference my heights off that face since it was now glued to my fixture. In my mind it works - except for origin being off centre bit!

i drilled the fixture board from both sides and have just noticed it looks a bit ordinary accuracy wise so am doing another as you suggest.

so i thought ......machined the registry holes into the wasteboard, same into both sides of fixture, then stock top....boom it all line up like magic! 🙂 ill try again

im going to pursue the center origin again just to see but if no joy then as you suggest ill do the side. manufacturing model suggestion - great thanks. i will give that a go.

thanks for fixing the file. really appreciate it.

Where are u at Chris. I am south east coast in Australia.

Been making acoustics for a while now but have bneen making a concerted effort to get my head around Fusion! Running out of brain cells!! 🙂

i dont know why i selected th e'tool orientation 'option??? might hav ebeen late, had a micro sleep and started clicking things! who knows. 

i'm a bit unclear on the whole corner origin, flip the axis etc just yet. im on a more trial and error journey it seems with fusion i think lol. i have to punctuate the countless sometimes pointless hours in front of the screen with some CNC machined proof that yes its still wrong!! 🙂 really enjoy it though.....and i'm teaching it to kids as well lol

cheers again

you happy if i send you a message if i get stuck again?

regards Scott

[chrisQX8DX · Accepted solution]

hey Scott, glad i could help 🙂 for sure hit me up whenever you're stuck, been building mainly electrics of my own design for the last 5 years or so ... just recently completed my first acoustic but there wasn't so much CNC involved (except for neck and bridge) and i really enjoyed the process by hand. I'm based in Berlin.

For the origin thing, i only make objects which are different from both sides, sometimes i even machine from 3 or 4 sides. The easiest i think is if you start by centering the object acurately in a square stock box, but it doesn't even have to be perfectly centered (which isn't really possible at all if you have an assymetric object) as long as you correctly reference you stock box. it would be easiest if i could record you a screencast but Screencast has always been hopelessly broken on my Mac, it almost never works. trying to get it to work once again but here isn't even a download link for the latest version anyhwere on the Autodesk sites the onnly versin i could find is the one i have and it's from 2019.

[scott_lipsham]

love to see aphoto or two sometime! you sell them or anything?

im keen to do some design too but will practice on some tried an dtrue stuff for a bit yet.

ive been watching austin shaners tutorials on youtube - you fmiliar with him? especially interested in the parametric stuff where you just change a few key references to generate a new fretboard for example.

lots of great resources out there. such an  exciting time to be doing this in - nice combination of tech and handskills.

chat soon 

have a great day - midnight here - still in front of fusion .....again! 🙂

regards scott

[chrisQX8DX]

ok so managed to get screencast to work. here's a video showing 3 things:

1. Converting bodies to components. Unless working with very simple objects you always want to work with components. It cleans up your workspace and makes it much easier to manage different parts of a model. Read up on components here in the forum or in the knowledge network, understanding Components in Fusion is really a must so spend the time at some point it will help a lot with everything you do in Fusion.

2. Enabling Design history. This is optional and there re pro's and cons of it vs. direct modeling (which is what you are using) but for guitar building i think it's a must. It allows you to change things after the fact, so for example let's say you are building a neck from scratch from a sketch you did in Fusion. You modeled it all the way and finished cutting it and now you want to do another one just like it just with different tuners (which may need bigger holes) a different neck width, a different filet for the edges, a different neck profile, differnt scale length, etc. If you have design history enabled you can then change all these things easily and the model will adjust. With direct modeling like you do there isn't really an easy 'way back' if you want to change things after the fact. The caveat is that using Design history is not super easy and you can also break your design history and end up with a load of yellow warnings in the timeline. But just like with components it's well worth the effort to learn using it properly and it will make your life so much easier once you got the hang of it. It's also super important that you understand sketching, especially constraints in Fusion properly so that's another learning topic to add to your list 😀

3. Then i show you how i would set up a cam setup with the 'flip & rotate viewport' method. (Note here: You will see me zooming all the way in to be able to select the corner point of the stock box, that is a bug in Fusion: when you have your box corner set on the world origin you have to zoom all the way in to be able to select the point, zoomed out you won't be able to select it.) I hope you can see/understand what i'm doing here: to set up the other side, rotate the view and then select the box corner which is now at the same location than the other side. The axises will point to the wrong directions at this point (because you just rotated the view, not the origin or the model), so all you have to do is to click on the axises to 'flip' them so they point in the same direction as on the other side.

hope that helps 🙂

i have no idea how to embed this screencast in the forum, it won't let me saying it's invaliid html, seems to be a common bug since years here with no solution

https://autode.sk/3tp2ugn

it's also really bad quality, not sure why, hope you can still see sth

[scott_lipsham]

Hi Chris

Man that is awesome - thank you so much for taking the time. I will watch this on the wekend. While the quality isnt so great as you say, i will be able to follow it thru along with the notes yu have written. Does screencast not have sound? ive never used it.

Once gain - many thanks for the effort. I will send yu a photo of the finished neck when i get a chance to run it again on the CNC.

regards Scott

[scott_lipsham]

hi Chris

hope all is well on ur side of the world.

i did another neck tonigt with the corrected model. machined a new fixture. i perservered with the centre stock idea and it was bang on perfect! im goig to do a couple more that way. I figure i dont need to be too particular about the stock prep  - see how it goes

Soooo happy - thanks for all ur help really appreciate u taking the time. I watched the vid an dtook some notes - really helpful

thanks again

regards Scott