3D track - fun with laser-cutters

[Phil G]

Hi James
I currently don't have any windows only drop down sides, and a lifting front and back door, which then somewhat defeat the point of the enclosure. worth noting there a bank 6 fans in the back of the roof (out of camera shot) which duct all the smoke out.
But yes the camera would be multifunctional and would allow for pass though as well.
I Think I will get one and see how it goes.
cheers
Phil,

 

[Martin Wynne]

so something is different maybe inside or outside was a bad way of explaining it, maybe it's proximity to previously freshly cut material. I don't know why, but I do know its a measurably different value.

@Phil G @James Walters

Hi Phil,

Just re-reading that, and something isn't making sense. The sockets are cut first before the timber outlines, so when cutting the sockets there isn't any proximity to previously cut material.

Are you saying that the sockets are ok, but you need a different kerf offset for the timber outline where it passes close to a previously cut socket? That would be extremely complex to implement.

It seems to me the important thing is to ensure that the kerf offset setting in Templot exactly matches the kerf which the cutter is actually creating at the ply surface. I know it is probably wider at lower depths into the material, and may change according to whether the laser is travelling across the grain or along it in the different ply layers.

Perhaps you would get better results by setting the offset to zero in Templot, and using whatever kerf adjustment is provided in your laser-cutting software? That may use a more complex algorithm than a simple offset. Maybe you are using both settings and seeing a conflict?

I'm just resigned to the fact that I won't get definitive answers to any of this stuff until I get a laser-cutter myself and try it. That can't happen this year if I'm going to make progress with the chair heaving and K-crossing chairs in time for Scaleforum.

cheers,

Martin.

 

[James Walters]

I've not experienced this issue on either of the 3 laser cutters which I have used to produce timbering. 1 at work and 2 at home. I now have a third machine at home, so I will try to produce a comparative test and see if I can notice anything similar/peculiar over the 4 machines. I doubt I will, but I was going to do some tests anyway so it's worth looking into, and having a selection of machines at my disposal may help identify a pattern.

However, I must be missing something here, so please forgive me if I am, but what is wrong with adjusting the socket sizes to compensate? Isn't that part of the basic methodology of optimisation we all ought to be doing anyway.*

Also, if the socket size can be optimised using the kerf setting only, then surely any consequent discrepancy with regard to timber width/length must be negligible, and probably only a couple of 10th's of a mm. Is this an actual problem, given that the kerf cut will be tapered slightly anyway? I'm not sure I'd even notice it.

As I say, I must be missing something, and would like to understand this issue, if only to better inform my own testing.


*If these were traditionally engineered components we'd specify max/min tolerances for both the plugs and the sockets, and have consideration for the tolerance 'stack-up'. One of the benefits of the technology we use is that we can (through trial and error), effectively reduce the tolerance stack-up to 0, but we will have to experiment to optimise our parts.
Personally, I print all plugs at default settings, and achieve my 'fits' through the kerf setting only, as I find the resin prints to be the most consistent element of the process. Doing so takes only a moment, on a scrap of the material being used, each time I cut from a new piece of material.

Could the source of this problem be localised 'drying' of the material after the heat from the first pass?
What would happen it the cuts were made in the reverse order? Would this make any difference? Would the error change 'sides'?
What is the moisture content of the material before and after cutting?

Answering these questions might hint to a source of an inconsistency. If so, these would be local factors.
But again, I still don't think I understand the problem. What have I missed?

Best,
James

 

[Martin Wynne]

I've not experienced this issue on either of the 3 laser cutters which I have used to produce timbering. 1 at work and 2 at home. I now have a third machine at home, so I will try to produce a comparative test and see if I can notice anything similar/peculiar over the 4 machines. I doubt I will, but I was going to do some tests anyway so it's worth looking into, and having a selection of machines at my disposal may help identify a pattern.

However, I must be missing something here, so please forgive me if I am, but what is wrong with adjusting the socket sizes to compensate? Isn't that part of the basic methodology of optimisation we all ought to be doing anyway.*

Also, if the socket size can be optimised using the kerf setting only, then surely any consequent discrepancy with regard to timber width/length must be negligible, and probably only a couple of 10th's of a mm. Is this an actual problem, given that the kerf cut will be tapered slightly anyway? I'm not sure I'd even notice it.

As I say, I must be missing something, and would like to understand this issue, if only to better inform my own testing.


*If these were traditionally engineered components we'd specify max/min tolerances for both the plugs and the sockets, and have consideration for the tolerance 'stack-up'. One of the benefits of the technology we use is that we can (through trial and error), effectively reduce the tolerance stack-up to 0, but we will have to experiment to optimise our parts.
Personally, I print all plugs at default settings, and achieve my 'fits' through the kerf setting only, as I find the resin prints to be the most consistent element of the process. Doing so takes only a moment, on a scrap of the material being used, each time I cut from a new piece of material.

Could the source of this problem be localised 'drying' of the material after the heat from the first pass?
What would happen it the cuts were made in the reverse order? Would this make any difference? Would the error change 'sides'?
What is the moisture content of the material before and after cutting?

Answering these questions might hint to a source of an inconsistency. If so, these would be local factors.
But again, I still don't think I understand the problem. What have I missed?

Best,
James

@James Walters

Hi James,

I'm not convinced there is a problem, but without a laser-cutter I can't be sure. If I had a machine, my intended way of working would likely be:

1. make a trial cut in the desired material at the intended power settings.

2. measure the kerf using feeler gauges.

3. set that as the kerf setting in Templot.

4. export a file to make a test piece and measure it (cutting the sockets before the timber outline).

5. adjust the DXF socket size and scaling settings as necessary.


We seem to be going round in circles. Here's some stuff I posted 18 months ago:

---

This is how it is intended to look immediately below the chair base:

The plug dimensions (pink+red) are taken directly from the prototype at 1" inside the chair base all round, so for an S1 chair the plug is 6" x 12.5" (2.0mm x 4.17mm in 4mm/ft scale).

The socket dimensions (green) are then set so that there is an easy fit or small clearance at the sides of the plug, and an interference overlap (red) at each end of the plug to create a bash press-fit. In the FDM bases this causes the plastic to be deformed sideways into the green clearance spaces, creating a firm interference fit and an accurate track gauge. In the CNC milled bases, the MDF material compresses to produce the same result. The idea of a bash fit is that it is more tolerant of variations in socket size (caused by the fixed printer resolution) than a traditional engineering press-fit.

I'm not too sure how this would work best in plywood bases, and without my own precision laser-cutter to conduct experiments it is difficult to find out. I imagine the red overlap might need to be increased to achieve a firm fit, by reducing the socket length. Alternatively you might need to abandon the idea of a press-fit and glue the chairs in place.

---

I don't think anything much has changed with regard to laser-cutting.

For FDM I have since introduced the easy assembly clip-fit option, which allows for a greater socket clearance to accommodate the variations and rounding effects in different FDM printers, while still retaining the chair firmly in the timber. But it still requires a close fit at the socket ends to ensure an accurate track gauge.

All the default socket size settings:

have been derived from trials with FDM printers, so I would expect them to need changes for laser-cutting (and CNC milling).

I think perhaps I should post a bit more about FDM printing, to restore the balance of discussions here on Templot Club. I'm sure there are more users interested in that than in using smoke and mirrors.

cheers,

Martin.

 

[James Walters]

USER=280]@James Walters[/USER]

Hi James,

I'm not convinced there is a problem, but without a laser-cutter I can't be sure. If I had a machine, my intended way of working would likely be:

1. make a trial cut in the desired material at the intended power settings.

2. measure the kerf using feeler gauges.

3. set that as the kerf setting in Templot.

4. export a file to make and measure a test piece, cutting the sockets before the timber outline.

5. adjust the DXF socket size and scaling settings as necessary.

We seem to be going round in circles. Here's some stuff I posted 18 months ago:

I agree 100%.
The above methodology works, and ought to be considered the default approach. With the new nib/snib updates on the way, I can't think of anything else that might need tinkering with in CAD outside of Templot to improve the accuracy of what we're trying to achieve.

re. posting stuff about FDM track bases - this is the fun with laser cutters thread.

 

[Martin Wynne]

re. posting stuff about FDM track bases - this is the fun with laser cutters topic.

Yes I know, but the fun with FDM topic is in danger of getting lost under all this chat about lasers:

https://85a.uk/templot/club/index.php?threads/3d-printers-fun-with-fdm.277/

I've just done a count in my workshop:

FDM printers: 6
resin printers: 2
CNC millers: 1
nibbler cutter: 1
laser cutters: 0

Which means either that I have something missing, or I need to concentrate my attention elsewhere.

cheers,

Martin.

 

[Steve_Cornford]

Hi Martin,
By agree, I meant with your 2nd option


If you cut the timbers before the sockets, the chances are the sockets would not be in the right place!
That is what I was advised by the bureau I used.
Internal cuts first, that is why they define a different colour for the internal cut lines.
When I used the bureau the Templot Kerf setting was not yet available, so I compensated by :-
a) reducing the socket width by 0.2mm (-0.1mm each side)
b) increasing width of timber by 0.2mm (+0.1mm each side)
Steve

 

[James Walters]

Hi Steve,

I agree, but now we have nibs & snibs it wouldn't matter which way round the timbers or sockets were cut, as the timbers are held in one piece and cannot move.

Best,
James

 

[Steve_Cornford]

Sorry James, I had forgotten about the nibs & snibs.
Fascinating to see your laser cutting the 3mm timbering base.
Steve

 

[John Walker]

I've not experienced this issue on either of the 3 laser cutters which I have used to produce timbering. 1 at work and 2 at home. I now have a third machine at home, so I will try to produce a comparative test and see if I can notice anything similar/peculiar over the 4 machines. I doubt I will, but I was going to do some tests anyway so it's worth looking into, and having a selection of machines at my disposal may help identify a pattern.

However, I must be missing something here, so please forgive me if I am, but what is wrong with adjusting the socket sizes to compensate? Isn't that part of the basic methodology of optimisation we all ought to be doing anyway.*

Also, if the socket size can be optimised using the kerf setting only, then surely any consequent discrepancy with regard to timber width/length must be negligible, and probably only a couple of 10th's of a mm. Is this an actual problem, given that the kerf cut will be tapered slightly anyway? I'm not sure I'd even notice it.

As I say, I must be missing something, and would like to understand this issue, if only to better inform my own testing.


*If these were traditionally engineered components we'd specify max/min tolerances for both the plugs and the sockets, and have consideration for the tolerance 'stack-up'. One of the benefits of the technology we use is that we can (through trial and error), effectively reduce the tolerance stack-up to 0, but we will have to experiment to optimise our parts.
Personally, I print all plugs at default settings, and achieve my 'fits' through the kerf setting only, as I find the resin prints to be the most consistent element of the process. Doing so takes only a moment, on a scrap of the material being used, each time I cut from a new piece of material.

Could the source of this problem be localised 'drying' of the material after the heat from the first pass?
What would happen it the cuts were made in the reverse order? Would this make any difference? Would the error change 'sides'?
What is the moisture content of the material before and after cutting?

Answering these questions might hint to a source of an inconsistency. If so, these would be local factors.
But again, I still don't think I understand the problem. What have I missed?

Best,
James

re: James comment about localised drying etc

I wonder if we a stretching the ability of our laser cutters to cut the tiny rectangles for our sockets (especially in 3mm) and sleeper/timber outlines with their lines including right angle corners. I am not sure if Lightburn optimises those constructs into separate lines.

If Templot were to break those constructs into a series of straight lines of different colours we could experiment with cutting say all the horizontal lines followed by the vertical lines on a different layer.

Is that worth playing with?

Martyn: would that be a fairly simple change to Templot?

I will have a go with my short (8 sleeper) test piece and see if I can see any difference.

Regards
John Walker

 

[James Walters]

HI John,

My comment about localised drying was just a suggestion as to what might be causing Phil's issues.
I think it unlikely, but is perhaps worth investigating. It is not something I've ever come across before.
I've had no problems with cutting the sockets, in fact I was surprised with just how well they worked in 3mm scale.

There are options in Lightburn to control acceleration and deceleration which might improve things for you, but I think that if you follow Martin's suggested approach above all should be well. That is essentially what I did, with my 3mm experiment, and the chair fit was fine.

Best,
James

 

[Steve_Cornford]

Hi John,
Templot outputs just lines already.
I dont think Phil is having trouble cutting small sockets or timbers, he is just having trouble getting both sockets and timbers the size he wants, but at present we dont know what Templot parameter settings he is using, or exactly what post export processing he is doing.

Steve

 

[Martin Wynne]

If Templot were to break those constructs into a series of straight lines of different colours we could experiment with cutting say all the horizontal lines followed by the vertical lines on a different layer.

Is that worth playing with?

Martyn: would that be a fairly simple change to Templot?

@John Walker

Hi John,

It would be simple to put the sides and ends of the sockets in different colours, and/or the timber outlines and nibs and snibs.

However, that is not necessarily going to be the horizontal and vertical directions. A lot of track is curved, or the template is rotated at an angle, and/or the timbers are equalized. In a full track plan some templates might be close to being vertical on the screen.

A lot of the initial plug track discussion has been thinking in terms of single straight turnouts, but my mind has been all along on full track plans and complex formations, with timbers and chairs at all angles.

cheers,

Martin.

 

[John Walker]

Well, I just tried it and it has made quite a difference.

I have cut 8 sleepers in 3mm scale. 4 as normal, the other 4 with the sockets broken into 4 lines and the sleepers with separate long edges. My sequence of cuts was:

one side of sleepers
socket ends
one side of sockets
nibs/snibs
other side of sockets
other side of sleepers

For the sleepers cut as normal the sockets and sleeper gaps fell out. I had set my kerf to 0.3 and my nibs at 0.7 x 0.7

For the others both needed to be pushed out. I had no kerf to measure but the sleepers are slightly wider making me think that the actual kerf was narrower than normal. Since the bits didn't fall out I suspect the corners are not as clean and maybe we would need the lines to cross as for the snibs.

I will continue with that experiment.

Regards
John Walker

 

[John Walker]

Hello Martin, our replies crossed.

For my next experiment I will use a curved template!

I was imagining what Lightburn and the cutter might do when presented with a rectangle. My guess was that it would be more of a strain to cut a right angle corner than two straight lines even if they are not vertical or horizontal. (To be honest I hadn't thought about that). I guessed best practice was to cut the straight lines separately and, if possible, extend them as you have done with the snibs. Once the lines are separate it is possible to adjust the sequence of cuts to minimise any local issues.

I agree that the plug track comes into its own with complex formations. With straight track my method for converting your DXFs to resin printed sleepers/timbers vertically worked fine but it would be a nightmare to extract them from curved track to print vertically. BTW they have not curled or shrunk as yet.

One of my reasons for moving to laser cut timber was that I would get the grain but now realise that strictly that won't work for curved track either. However given that the curve is not very great I don't think it is a serious problem.

As for my nibs, they fell out at 0.5 x 0.5 but haven't at 0.7 x 0.7. Next time I'll try your new default of 0.6 x 0.6 and revert back to a kerf of 0.2.

Regards
John Walker

 

[Phil O]

Could it be that Phil's lasers are working upside down?

 

[Steve_Cornford]

Hi John,
Did you include the ends of sleeper lines as part of your nibs/snibs line cuts?
What was the fit of chairs like in the two versions?
Steve

 

[John Walker]

Hello Steve

Yes. To minimise the changes I needed to make I left the nibs, snibs and sleeper ends as one and just separated the long edge of the sleepers.

No discernable difference. The chairs have not changed and are still too tight and difficult to install.

My plan is to get the sleepers right and then adjust the sockets and chairs. I don't want to tackle too many variables at a time.

Regards
John Walker

 

[Martin Wynne]

@James Walters @Phil G @John Walker

Would it make any sense to use the laser-cutter in a similar way to a milling machine?

i.e. An initial roughing cut to remove the bulk of the material. Then go round again to remove another few thou as a finishing cut, possibly using a different focus/power/speed setting.

This might result in a more precise finished size. Or I might be talking rubbish.

Martin.

 

[John Walker]

Hello Martin

I am not sure what we can do in Lightburn to change the settings from one pass to the next. However I guess we could generate more than one DXF from Templot with different kerf settings so the first one does the rough cut a second one cleans up to the required line. That would require two independent cutter runs. I don't know if the registration is accurate enough to produce the desired result.

Maybe James et al know the answer to that without actually trying it.

It occurs to me that we might have a problem if the pieces we intend to fall out twist or partially fall out between the passes.

Regards
John Walker

 

[Martin Wynne]

The chairs have not changed and are still too tight and difficult to install.

@John Walker

Hi John,

James seems to have good results in 3mm -- he just needs some rail:

cheers,

Martin.

 

[John Walker]

I asked one of the 3millers on our zoom last night to take some rail for James to the exhibition this weekend. I hope he remembers!

Regards
john Walker

 

[James Walters]

I've cut 5 templates today from my challenge project by way of an experiment to prove/disprove laser accuracy.
I'm sitting here scrutinising them now, and I can't see a problem at all.

My S1 socket widths are 2.14mm, and the socket lengths are 4.2mm. Sleeper widths are 3.35mm, and timbers 3.9mm these
are averages from measurements taken from the 5 timbering bases. All were cut with default settings. and 0.2mm Kerf setting.
Material was 3mm laser ply.

I took the cuts easy at 15mm/sec 100% power. I could have gone faster. All sockets have square corners and very neatly cut.

Chair fit is good for width, but there is a slight (very slight) end-play on the S1 chairs I tested, which can adjusted for using the adjust socket size dialog.

Best,

James

 

[James Walters]

I asked one of the 3millers on our zoom last night to take some rail for James to the exhibition this weekend. I hope he remembers!

Regards
john Walker

Thank you John, I'll be there Sunday and was going to take along the 3mm experiment as I know there are a couple of 3millers who will likely be with me on the SECR Soc stand. I've printed a 3mm wagon to take along too.

 

[James Walters]

Here are 3 of the timbering bases I cut and mentioned yesterday.
Bounding rectangles and end snibs added manually in Templot using background shapes.
The stuck 'chad' in the middle one is cut through, and it will poke out, I must have missed it before taking the photo.

 

[Phil G]

Hi James,
I have been away a couple of days and there certainly seems to have been quiet a few posts during this time.
I plan to do quite a bit of experimenting in the next few days, one experiment I want to try is putting a piece of 3 mm ply in the over for an hour or so to dry it out. the downside is maybe the ply will warp too much don't know yet.

One question I do have though when your quoting your sizes, are they measured at the top of the cut face or the bottom of the cut face?

The reason for the question is I am getting a very noticeable taper on my kerf which seems to change depending on speed and power settings( not enough test done yet) but does seem to be notably impacted, just to give you some numbers I can just get a 2 thou feeler at the bottom of the job but at the top its between 16 thou and 20 thou depending on the settings I have tried so far that is on both 2.5 and 3 mm ply. ( cant measure in mm as I don't have metric feeler blades)
Have you ever measured your kerf taper? or for that matter do you have a noticeable kerf taper?
cheers
Phil,

 

[James Walters]

Hi Phil,
That's a good question.
I've just had another measure-up.
My taper seems to consistently measure in the region of 0.1mm to 0.15mm in 3mm ply.
Given that the kerf takes the form a 'V' shape, i.e. narrows through the cut, I think this might be the source of the 'problems' discussed with regard to accuracy of timber and socket dimensions.

The timbers, which are cut externally will tend to widen from top to bottom. Using my kerf taper as an example by 0.1mm each side.
Whereas the sockets which are cut internally, will tend to narrow from top to bottom.

Thus, if measured from one side only, they will always differ by the kerf taper.

Given that we are interested in timber width, and socket fit from the top of the timber only. I would recommend sizing the Kerf setting in Templot to achieve the desired timber width, and using the adjust socket size dialog to control the chair fit, which is what Martin has been saying all along.
I'm not convinced that setting the socket kerf independently offers any significant advantage to the above process, other than allowing us to get to nominal sizes more quickly. Which might be OK if the plugs are glued.
However, it would still require adjustments to be made following measurements of the chair plugs, if we want to achieve accurate press-fits (and gauging) which takes up back to the adjust socket size dialog.
Personally, I'd vote for one less variable, but I'd be interested to hear others thoughts.

Of the different lasers I've tested and used, all produce kerf tapers of differing amounts. This is best managed through the speed and power settings. Also consistent focussing is critical, as is cleanliness of the lens (and mirrors if using a CO2 laser).

Accurate material setting is also critical. Not all plywood is equal, some will burn more readily than others, which is why it is important to test each piece as the kerf taper will change depending upon the material cut.

Best,
James

 

[James Walters]

Hello Martin

I am not sure what we can do in Lightburn to change the settings from one pass to the next. However I guess we could generate more than one DXF from Templot with different kerf settings so the first one does the rough cut a second one cleans up to the required line. That would require two independent cutter runs. I don't know if the registration is accurate enough to produce the desired result.

Maybe James et al know the answer to that without actually trying it.

It occurs to me that we might have a problem if the pieces we intend to fall out twist or partially fall out between the passes.

Regards
John Walker

That's quite easy to achieve John.

The XPos and YPos measurements give us the geometric centre of the object being cut (grey dot in the centre of the selection).
Provided a second file was contained within an identical bounding box then then ensuring these measurements were the same should achieve the desired result.

Alternatively, and perhaps easier is to adjust the offsets by colour in Lightburn. Double-click the Cuts/Layers tab, and make the adjustments here.

Best,

James

 

[Martin Wynne]

Given that the kerf takes the form a 'V' shape, i.e. narrows through the cut

@James Walters

Hi James,

That is puzzling me. Given that the laser beam is focused to a sharp point on the timber surface, I would expect it to widen above and below that point, like a letter X. Which would suggest the kerf would be an inverted V shape, and widen through the cut, rather than narrow.

What am I missing (apart from a laser-cutter )?

cheers,

Martin.

 

[James Walters]

Hi Martin,
It depends where the focal point is, which is up to the user. But typically the kerf will look like this.

If the laser is focussed to a fine point on the surface of the material, then you're right the kerf will be an inverted V, but the kerf angle becomes more pronounced and the cut quality poorer.
Typically, the ideal is to focus the laser on the underside of the material to get maximum energy through the part. Hence a V taper.
If the beam is focussed on the top then the heat dissipates as it passes through the material which tends to need a higher power and slower speed which in most cases is a disadvantage.

I think that's the theory. Happy to be told otherwise though.

 

[Martin Wynne]

Typically, the ideal is to focus the laser on the underside of the material

@James Walters

Thanks James. That seems surprising to me if you are trying to achieve maximum accuracy at the surface, with sharp edges.

If the focus point is below the surface, it surely means that the beam never will focus while moving, because the beam will be obstructed by the smoke and debris from the material above?

All this does suggest that for plug track at least, the machine focussing is critical, possibly more so than the power and speed? Also the importance of keeping the lens clean, if the actual socket dimensions are being cut by an unfocussed part of the beam.

Do the usual cutters have a Z-drive, allowing the focussing to be controlled by gcode from the software (as it would be if I attach a laser to my CNC machine)? Final socket sizing adjustments could then be done by changing the focus, rather than by re-exporting the file with a changed kerf setting or socket size.

cheers,

Martin.

 

[James Walters]

On the larger CO2 machines the Z axis control is achieved by raising the table. Those with autofocus will raise the table using a stepper until either a proximity switch is triggered, or a beam of light is broken. Those without autofocus tend to have a DC motor (in my experience), I cannot speak for all machines.
I believe the focus parameters can be programmed into the G code, although I've never tried. (The G code is easy to edit in Lightburn, should anyone wish to.)
Diode lasers intended for the hobby market tend to have a manually adjustable Z axis, and the focussing is achieved by placing a calibrated block between the workpiece and the laser module. My tests with the Creality Falcon 2 and the Atomstack machine, suggest these 'calibrated' blocks are just a starting-point, and some experimentation is required to get the best results.
I have my own version of this tool which I use for testing.

A work-around, is to flip the part in Lightburn and cut with the top side of the material on the work bed of the machine.
However, in the case of what were doing here we will still have opposing taper angles on the timbers and the sockets.

Of all the variables, focus setting is probably one of the most important, yet least understood (especially by me). I think it would be impossible to account for it in Templot, as all machines and materials are different.

I think that in practice all of this is a non-issue, we have all the controls we need in Templot already! I've spent a lot of time thinking that I could improve upon the vectors I got from Templot with all sorts of tinkering in CAD, but it is just not the case.

 

[Martin Wynne]

I think that in practice all of this is a non-issue, we have all the controls we need in Templot already!

@James Walters

Hi James,

I thought that too -- but that was before i knew about tapered kerfs!

It seems that many users of laser-cutters are going to end up with sockets too tight or too loose if they just switch on and press GO. There seems to be a lot to learn and test in order to get accurate results.

Would it help if the end of the socket is indented like this (exaggerated) ?

This to create an interference fit, i.e. the plywood needs to be deformed when pressing the plug into place. This would centralize the chair in the socket (for accurate gauging) without requiring a close tolerance on the socket dimensions.

A similar interference bash fit is used on the FDM press-fit, but in that case we have a deformable plastic material, which is not the case for plywood.

So that's yet another setting -- amount of socket indent. Set to zero if not wanted or doesn't work.

cheers,

Martin.

 

[John Walker]

Thanks James for the help with navigating through Lightburn. I notice that in Device settings I can enable the Z axis and then run a Focus test. Have you used that? If the Focus test actually works then it makes the calibrated block look rather crude.

I have cut 8 sleepers in 1.5mm ply for 3mm scale and 3mm ply for scale 7. Both are curved, the 3mm one at 3ft radius and the S7 one at 6ft.

I used the default kerf of 0.2 and the new default nibs of 0.6 x 0.6. Handled with care the sleepers remained in the fret in both cases.

For 4 sleepers in each case I separated the long edges of the sleepers and slots but did not extend those edges across the corners.

On the "standard" 4 sleepers the pieces in the slots and between the sleepers fell out. On the "adjusted" ones they needed a little push. My feeling is that the "standard" approach produces a cleaner corner. If Martin adopts the separate lines I suggest they need to overlap at the corners but I haven't tested that, yet. I will if required. However I am not convinced this has made any difference, with the reservation below.

The short piece of track started pointing vertically. The first 4 sleepers were "standard" the second 4 were "adjusted". That means the "adjusted" ones were at a more acute angle to the vertical. It may be that as the angle increases the cut is not so clean. Also, measuring the length of the first and last sleeper from the Templot generated DXF there is a slight difference in length, in the 2nd decimal place, so we may be experiencing an accumulation of very slight errors.

If I set my calipers to 2.5mm and 25.5mm I can convince myself that my 3mm sleepers are all on size. They vary by a small smidgen below that. My feeler gauge suggests my slots are 0.8mm wide as intended. Gauge is difficult to measure as my chairs have been in and out so often that they are breaking up. I am planning to revisit the chairs next. I am also going to move back to my standard Phrozen resin instead of the water washable variety.

For 3mm scale my thoughts are to use the (new) default Templot settings for the laser cut sleepers except for adjusting the slot width. I will play with the focus. After my experience with ballasting the 3mm Society laser cut sleepers I would prefer an inverted V because it helps to bind the ballast around the edges of the sleepers.

I measured my scale 7 sleepers at 59.3 (59.5) x 5.84 (5.833). The sleepers seem to gain a little length as they (or I) go round the bend. I can live with that. My chairs fitted nicely but it is under gauge at 32.7 instead of 33. My settings were 6mm/sec and 70% power. Based on my material test Martin's settings would not have cut right through. This could be a material difference and/or the focus.

Regards
John Walker

 

[James Walters]

@James Walters

Hi James,

I thought that too -- but that was before i knew about tapered kerfs!

It seems that many users of laser-cutters are going to end up with sockets too tight or too loose if they just switch on and press GO. There seems to be a lot to learn and test in order to get accurate results.

Would it help if the end of the socket is indented like this (exaggerated) ?


View attachment 8482

This to create an interference fit, i.e. the plywood needs to be deformed when pressing the plug into place. This would centralize the chair in the socket (for accurate gauging) without requiring a close tolerance on the socket dimensions.

A similar interference bash fit is used on the FDM press-fit, but in that case we have a deformable plastic material, which is not the case for plywood.

So that's yet another setting -- amount of socket indent. Set to zero if not wanted or doesn't work.

cheers,

Martin.

Hi Martin,
I think that could be a good suggestion. The ply we are typically using with laser cutters i.e. laser ply is quite soft when compared with say birch ply. I'd say it's definitely worth a try. It might make for a more compliant tolerance for a decent chair fit.

Users would still need to be accurate though with their focus set-up.
I fear this conversation might be over-complicating things a bit, although I think optimising a laser is far simpler than optimising an FDM machine.

Best,

James

 

[James Walters]

Thanks James for the help with navigating through Lightburn. I notice that in Device settings I can enable the Z axis and then run a Focus test. Have you used that? If the Focus test actually works then it makes the calibrated block look rather crude.

The focus test in Lightburn is great, but you need a powered Z (or autofocus) axis to use it, typically a larger CO2 machine, in which case it is the cutting table which goes up or down rather than the laser head.

I cut my own little calibration block, like a staircase, rising in 1mm increments. I test by cutting a short line at each height setting in a piece of material until I find a focus length which works well for what I need. I don't think a precision of less than 1mm makes much difference.

There are times (when engraving) when better results can be achieved by deliberately going out-of-focus, but let's not go there for now!!

 

[Martin Wynne]

Hi Martin,
I think that could be a good suggestion. The ply we are typically using with laser cutters i.e. laser ply is quite soft when compared with say birch ply. I'd say it's definitely worth a try. It might make for a more compliant tolerance for a decent chair fit.

Users would still need to be accurate though with their focus set-up.
I fear this conversation might be over-complicating things a bit, although I think optimising a laser is far simpler than optimising an FDM machine.

Best,

James

@James Walters

Hi James,

All done:

But the tricky question -- should it be on or off by default?

And if on, by what amount? The default I have set there is 1/3rd inch scale. That's at each end, so 0.22mm shorter socket in 4mm/ft scale.

I will try to get 244a out soon so that you can test some of this latest kerf stuff, and leave the chair heaving half-done until later.

I agree about frightening folks with unnecessary complications, but where do we draw the line?

cheers,

Martin.

 

[Phil G]

Do the usual cutters have a Z-drive, allowing the focussing to be controlled by gcode from the software (as it would be if I attach a laser to my CNC machine)?

Hi Martin,
As usual your ahead of the curve, it very much looks as though the next generation of lasers will have not only a z axis, but will also have full surface touch probe technology. X-tool are already doing this with there S1 model. X tool also seems to have gone toe to toe with lightburn on software, right now there point of difference is the Z axis capability, referenced both as a tool to compensate for material warp but also to allow for etching of very undulating surfaces, There example being a corrugated profile.IE you Z probe the surface just like on FDM printers and the the z axis through the software will compensate over the mesh surface area created by probing.

re your comment about the kerf shape, its always a taper shape, what I think may change depending on power /speed variable is the angle of the taper.( not sure yet more experimenting required.)

if you want a milling machine analogy, for the focus point part, think of it as a taper cutter where the focus point would equate to the actual height of the work piece, moving the cutter up and down in z would have an effect on the top and bottom measurable values of the gap created by the cutter but not the actual taper angle. Same thing with setting the focus point.

Re pockets or sockets and profiles, at this point in time I am more concerned in understanding why the socket and outside shape are not a consent kerf than how to fix it.

interestingly I have just found out lightburn has both an internal cut and external cut kerf offset capability, not sure how to use this just yet but it has to be provided for a reason.
cheers
Phil,

 

[Phil G]

Hi Guys,
More thinking on my part ( very dangerous I know ) I think the reason the clip fit work so well on laser cut timber (when the right size socket vales are found) is the clip is acting as the inverted version of the taper IE think of it more like a cross section of a Morse taper lock system.
The issue is there are only 4 points of contact, and two of them (the gauging two) don't have taper at all. which would logically suggest gluing after fitting is a good idea.
cheers
Phil,

 

[Phil G]

It seems that many users of laser-cutters are going to end up with sockets too tight or too loose if they just switch on and press GO. There seems to be a lot to learn and test in order to get accurate results.

Hi Martin,
May I suggest we start a new topic thread called some thing like "understanding laser behaviour", the idea being anybody with a laser can contribute to there findings and offer there settings. Thus in the future as new people join we could direct them to known good machine setting values . It will always be a starting point but better then starting blind. In that way there is no need to make any Templot changes to factor in a myriad of possible machine variables.

I am thinking a bit like the lightburn forum you start by stating which machine your talking about i.e. Creality Falcon 2 22W or 40W. Or even which brand type of laser head your using I.E laser tree 20W working on a Masuter pro CNC router. or laser tree 5W mounded to an FDM printer.

This is important as different laser head driving systems, will have different max speeds and very likely different acceleration and deceleration settings as well.

If you think about this a bit deeper, there is no way a machine running at say 20mm second in a straight line, will be able to cut a pocket 2mm x 4mm nominal at that same speed, there just is not enough time to ramp up and the damp down to achieve max speed.

ergo if we are using the same power setting for both straight line cuts (say sleeper profiles) and pockets but we are not using a constant speed what effect does it actually have on dimensional stability?
cheers
Phil,