Design & Build of Frankenrouter

[peteeng]

Hi Mactec - I was surprised too. But I have played with the usteps (1/8th & 1/16th) and driver decay settings and it has smoothed out. The video had zero motor decay setting. It is now set on medium decay and its much smoother can't pick the diff if I use 1/16th or 1/8th. Still not happy with the z settings, working thru that today hopefully. I think it has something to do with the stiff spindle plate deforming the Z plate, it skips occasionally at near full down. I did it up very tight. More news when I know. Will do new video when smoothed out. Peter

Re - vibration. With no decay on the motor the baseboard was acting as a sound board. The low tones especially, the dust was jumping around busily. I'm making some carbon fibre sound boards for a guitar builder friend. That will be interesting as well.

[peteeng]

Evening all - Today I chased the gremlin in the Z axis. Occasionally near bottom it would skip a bit. Intermittent issues are frustrating! So I pulled the z down checked the fit of everything so I knew there was no mechanical issues. Then pulled down the motor support bearing and it was smooth as silk. But the seal is very tight on the new spacers. So I greased them up well and rebuilt it a bit loose. Then tested it and tested it, then snugged things down a bit better. Seems to have lost its tick.... I measured its envelope today. Happy with it. Never was designed to an envelope except the Z axis which I wanted to be 300mm. Its now 310mm so very happy with that. Now to design a run in program and give it a long spell moving.

Have decided to design a full sheet machine called Lanky... Will mainly use aluminium and plywood laminates but want some cast grout in there! I'll start a thread when developed a bit more.... Frankie is nearly in the wild. Peter

[peteeng]

Morning all - Frankies first parts! Clamps for itself. Pushed the 6mm 1F tool a little hard at first (pine was steaming) but backed down. Happy with the cut and speed. Now to make a surfboard fin and Frankies destiny is realised. Peter

[peteeng]

Hi5 All - Well Frankie is feeling solid. Gave it a bit of a workout today doing clamps and a collet holder. The collet pockets where done with a 3deg ramp 2500mm/min feed could go much more aggressive than that... Now to make a surfboard fin as that's what he's been built to do. Peter

https://www.facebook.com/ScootCNC/vi...41491829248657

[crjohnson]

Hi Peter, I follow you on facebook a little and saw you were doing tests on linear and corner max accel's etc. Just wondering how your testing went?

I use UCCNC and an axbb-e on my machine and still think the CV settings could be improved.
Cheers

[peteeng]

Hi CR - The X and Y are now at 100mm/s/s (0.1g) and V=5m/min (500rpm). The Z is at 40mm/s/s which is the lowest I have ever used. I think this is because of the ST6560 controller... I have two test programs. 1) is a conventional 2.5D square and circle with fillets 2) is a full 3d spline quite big I call the roller coaster (RC). The RC has a small kink in it that taxes the z axis. It usually goes "bump" at that spot. So I change up the velocity in say 200mm/min intervals until it stalls. Then back down or change the accels. In UCCNC the comp-accel seems to be related to the accel somehow. You can put in a figure but it can get overridden. So its a process of creeping up the figures until it stalls, back it down and do some normal jobs and see how it goes. Once its comfortable let it be. I could ramp up Frankies velocities now but I shall leave them, it all feels solid at the moment and jogging is not slow so let it run in a bit more.

Scoot runs Z much faster with higher accels. Scoot uses DM542 controllers fully digital and self tuning. Smoother motion than Frankie I feel. The last Scoot build I used a 48V system and it ran really smooth. Recommend running as high a voltage as you can.

Frankie uses a ST-6560 24V controller and I think the signals are a bit choppy. Slow speeds its like a dolphins sonar click, quite a bump bump bump even at 1/8 usteps. So edge your velocities up a bit every time you use the machine and then adjust the accels to suit.

I just reread the uccnc manual and the compensation accel is to do with backlash and tapping not corning accels. So once Frankie leaves then I shall be setting up an Octo BTT controller and pad to use on my small router Brevis. This will have much better motion control then the CNC systems. Peter

[crjohnson]

Did you play with the linear error max., lin. addition length max., corner error max. and lin. unfiy length max settings much?

From my understanding these actually need to be adjusted for large and small parts/radii to get the best results. I did some testing a while ago on a part approx. 85mm in diameter and changed these settings to see how jerky the machine was and how they effected the Actual Feedrate (ACT).

It looked to me that when you get these dialed in nicely the machine is much smoother and the ACT is what you programmed in.

[peteeng]

Hi Craig- If your looking for more feed speed from UCCNC then you need to consider how your inputing data into it. 1) if your using dxf transfer this can be very limiting as the code can be made up of very small increments. 2) If your using cam software then look at the tolerance settings and back them off to an acceptable level. If your doing signage then there is no point in 0.001mm tolerancing in the cam system. Once you have understood what the source is doing then look at the general settings area of UCCNC. These have to be looked at in view of what your type of work requires. Again if signage then low settings give more speed. If doing fine machining then high tolerances are needed which may slow the feed.

stop at angle is self explanatory usually set at 89degs so if the corner is 89-90degs the velocity planner will slow to a stop at the corner
look ahead is say 200 or 300 but does not affect speed unless this bogs down the processor. I run at max 300 with no issues. 300 is not many if the increment is 0.5mm thats only 150mm ahead
linear unify length max. If there are many straight lines they are combined into one line. The combining of lines max length is set by this number say 20mm. Say you import a dxf and the polyline is set at 1mm discretization then it will only add 20 segments together... also if the line is a curve that same thing is done and the actual path is kept to the linear error max figure. if the code is made up of many short moves the planner is not allowing the path to get to full speed. By decimating points and having a longer path the machine has time to get to max velocity if the accel time is there.
linear error max. As each segment is added together, the error from the averaged line is calculated and either the 20mm max will kick in or the max deviation error will kick in at some point. Then the planner stops averaging (decimating points) and starts again planning fwd. Simples??

This is explained in the manual. So in short back these figures off and the velocity planner will run much faster. You can set up a "rough" config and a fine config. So you could do roughing faster then flip to the fine config for fine work etc... Peter

[peteeng]

In facebook I did a little exercise -here is the images. Peter

[peteeng]

Hi All - I set up a test block of aluminium for Frankie today. happy with results. I set up a round pocket and cut down in 2mm depths with a progressive chip load. CL=0.025 then 0.050 then 0.075 then 0.100mm. Was very happy with the cut on the first three. The CL=0.1mm left a burr but the bottom was good. The issue I had was with the ramp entry. Initially I had 3deg and a slow feed but Frankie wobbled with that. I then used 1deg and 1000mm/min and that was good. I then used 1.5deg and we were back to the wobble so 1deg it is. Will have to set up a square pocket now to see how the settings go in a corner. I think 0.075mm maybe max... The tool was a 1F 6mm bit at 24000rpm. For carbide it seems 300-600m/min surface speed is OK. But I have read lots of data and that seems to vary from 150-800 depends on the source.... was nearly a dry cut I used a mix of water and metho.... So now need another bit of metal to cut. I dropped into the local balustrader to pick up some scrap but he was out. Maybe next week in my rounds.... Plus I have to get an air blast in there. The recut chips sound bad!! Peter

[peteeng]

Hi All - Frankies lucky today aluminium and carbon fibre in one day! I cast a CF block awhile ago and it had a dry sport. So I decided to exorcise it today with Frankie. Cutting the dry stuff was weird and fluffy but it got there. Now to make the pocket a little bigger to clean up the last bits. Then onto a surfboard fin... got to figure out the flip from side to side. Peter

[peteeng]

Hi All - Today I started thinking about load testing and stiffness testing of Frankie. So I got out my luggage scales and started loading Frankie. But the scales time out fast so don't measure for long enough to be useful So I'll have to find a another load cell or scale to do this. Peter

[peteeng]

Hi All- Heres an article I've written on aluminium laminates. Where I want to publish it does not take pdfs so its here so I can link it. Peter

[peteeng]

Hi All - Its getting closer to handover, so I tidied Frankie today with proper labels, closed up the electronics and brushed it off. last thing is to make a fin. So have started working out the machining logic and toolpaths. Even the RH screw rule was needed today. Have started on Lanky. It has big feet! 3000mmx1500mm. Will start a thread soon. Peter

[peteeng]

Todays job for Frankie was making a surfboard fin. This is what the Maker wants to do with it. So learnt a bit more about Fusion and about what no2's workflow will look like. Unfortunately the tip split so its a sawoff fin. But happy with progress. I think a vacuum clamp is the go. But we'll get through a couple of fins then review. Peter

[pippin88]

If you orientate the grain differently, you will reduce the chance of splitting the tip off. Grain running lengthways along the fin longest axis

Looks good otherwise

[peteeng]

Hi Pippin - The split was in the timber billet from the start. Its a trial part so I didn't pay attention to grain orientation. Its just to prove the toolpaths and workflow. The real fins are fibreglass no splits there! I have thought ahead for a new workflow that does not require the cavities. But It's up to the new owner to develop his workflow not me! He wants to machine 5 or 6 pieces at a time so fixturing will be different. But my brain latches onto these things and goes crazy until a solution appears... I want to get busy on Lanky asap.... Franky is behaving well and its time for it to leave the workshop. This exercise has pushed my learning into Fusion CAM deeper and Fusion assemblies deeper so is very worthwhile. Peter

I've done foils before in mecsoft and had to figure out how to handle path extensions, edge containment and edge transitions so they are not too jerky and maintain midline accuracy. Now I'm going through that process in Fusion....

RE: Fusion assemblies - the penny has dropped with Fusion assemblies and now I can move things around (move), join them together etc. Also cracked the internal external component stuff. Now I want to play with different joints and motion....

[peteeng]

Morning All - Today I flipped the fin and fitted it to the RHS cavity. Decided to use double sided tape but the one I have is 1mm thick so reworked the cavity 1mm down. All fitted up nicely. So then finished the RHS of the fin and it looks quite good. The trailing edge is very thin and splintered quite a bit. Really happy with the midline it was near perfect. I have an idea for a fixture that does not require cutting cavities so may give that a go. I'll need to cast some FG or use something that can create a thin edge. I may vacuum epoxy saturate some MDF that works well.as a tooling board. Then I think were done.... Peter

[peteeng]

Hi All - I was looking at the fin and flexed it a little and it snapped. The trailing edge is actually "hollow" so really really thin. So I'll look at the model and add some rails to keep the foil more convex along its trailing edge. I think I'll do a stack of polyester and CF for the next one. I have grey and white PE felt and carbon. May even put some red cloth into it for some interest. I'm setting up to cast three types of Tetriums for test so may as well cast a fin block as well. The tests are for steel fibres, aluminium powder and ALOX. Hopefully get that done this week... Peter

The Tetrium-S (steel fibres) spindle plate is doing fine. I was surprised when setting up earths that it was conductive. I would have thought the epoxy encapsulation would have stopped all of that. Plus there's only about 20% by volume steel in it... But continuity was good so earthing was easy.... Plus the top 6mm was CF...

[peteeng]

Hi All - I've cut Foil No2. Used no backing mould just self supported. Went OK but next time will put a small tab near the tip. This will be a good training exercise for the new owner. Peter