RepRapPro's new printer kit (off-topic)

I was told on Friday that the orange pads were for attaching print cooling fans, but their design was not finalised yet. The belt tensioning screws are at the top of the frame.

To use the Mini Kossel heated bed (if it will fit in the space available), you would need to use a different Z probe, because the Z probing system is built into the bed supports. Maybe there would be room on the effector for one of my mini differential IR boards, [miscsolutions.wordpress.com].

Edited 1 time(s). Last edit at 05/25/2015 08:49AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I think one simple solution is to use a standard probe as could be seen here (on Cherry Pi) :
[www.youtube.com]
He uses an inductive sensor, which will work only on aluminum, but is very low cost. With a capacitive sensor, a bit more costly, it is possible to accomodate other materials.

The diameter of the bed shown on the photo is approximately 170 mm in diameter, so a 180 mm bed shall extend very slightly from the side plate.
However, I think they shall set the plates 4~5mm further from the center, which will allow the carriage to move within the plate without the external cut needed now and reduce the width of the center hole. The side plate will be stronger, which will reinforce the machine and may allow fitting a 180mm plate within the side plates. However, I need more precise data to check.
The delicate part to size the center hole is to accomodate the cooling fans.

The Kapton heater on the Mini Kossel bed is 160mm diameter. This means that you could shave nearly 10mm off the edges off the bed plate away from the 3 fixing holes, which might make it easier to fit inside the side panels.

The inductive sensor in that video looks rather large and heavy for a small printer such as the Fisher.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

For non contact sensors, M18 is the most common diameter, however, it is easy to find inductive in 12mm diameter at low cost.
For much higher cost, it also exists 8mm diameter sensors. Capacitive sensors, however, are more difficult to find in 12 mm diameter, but they exist in flat package.

Inductive sensor, direct from china, 2.3 GBP, postage paid ... That may take some time, however...
[www.ebay.co.uk]

Capacitive sensor, in a flat package, more difficult to install, but compact, 3.44 GBP, postage paid :
[www.ebay.co.uk]

Quote
PRZ
For non contact sensors, M18 is the most common diameter, however, it is easy to find inductive in 12mm diameter at low cost.
For much higher cost, it also exists 8mm diameter sensors. Capacitive sensors, however, are more difficult to find in 12 mm diameter, but they exist in flat package.

Inductive sensor, direct from china, 2.3 GBP, postage paid ... That may take some time, however...
[www.ebay.co.uk]

Capacitive sensor, in a flat package, more difficult to install, but compact, 3.44 GBP, postage paid :
[www.ebay.co.uk]

I wonder how temperature stable those sensors are. Making a capacitive or inductive sensor is easy-peasy. Making one that gives a repeatable trip-point to within 0.1mm over a 100C range is keeping me up late into the night.

---

RS Components Reprap Ormerod No. 481

I have asked RRP to have the Fisher Delta data to enter in my OpenScad simulator, but didn't get it (yet).
So, I have gathered data from the photos. Overall I am relatively confident in the number I found, while there is some uncertainty about effector offset and arm length.
Dimensions are :
Height is 500 mm. Printer is inscribed in a 280 mm cylinder . Side plates are 200x500 mm.
Rods are 10mm, spaced 45mm, with the axis plane of two rods at 125mm from the center.
Arms are spaced as the rods, by 45mm, with a length of 167 mm (+/- 2mm)
Effector offset is 21 to 22 mm, carriage offset is ~22mm. This is quite compact (Kossel mini effector offset is 20mm).
The simulation confirms the data, with the following points :
*The clearance on top is huge, the carriage stopping at 40mm from the top structure. Maybe there is some need for room for the extruder/bowden tube setup .
*The usable diameter could be more than 150mm, provided that :
- There is no fan on the sides of the effector
- lengthen slightly the arms
- Increase the height of the plate opening (by 30mm +)
- The arms are very close of the side plate while carriage is at maximum height
I really don't see how it could be possible to install part cooling fans at the periphery of the effector. Even if that need some increase of the effector size, the part cooling shall be done with a fan above the effector.
I've made a film with the animation of the Fisher Delta simulation, which you can found here :
[github.com]
The simulator needs last version of OpenScad, preferably the nightly build (they are generally stable). To play with a machine, uncomment the given bot dataset line.

To compare, I have also run a simulation of the micro-delta of e-motion tech (France), which have the same target as the Fisher (delivered as kit for 400 Euros ~ 285 GBP - VAT included), but is a bit smaller (450mm height , inscribed in a cylinder of 270 mm), with a much reduced usable diameter of 120 mm. Part of this small usable area is due to a huge effector, with an offset of 32mm. For comparison, the Rostock max have an effector offset of 33mm. This effector is designed for classic all-metal hotend, which may explain partly its size.
See here a nice page about its assembly : [blog.voscain.com]
This machine got some success in France and is reputed printing relatively good quality parts. It intensely use acrylic parts, but the top and bottom plates are in wood. Recent improvement have seen acrylic stepper supports replaced by steel. Use of acrylic cut parts for everything rised some concerns in the forums, with the usual heat associated...
What is interesting is the overall setup, with steppers on top, which made for more simple wiring and very easy assembly and maintenance. The bowden tube looks a bit tortuous, but there is no loss in height . Heated bed (120 mm) is proposed as an option. I really think the Fisher Delta shall adopt this setup, while that will reduce the side plate bottom 'beam' size. Some tradeoff shall be done there, but as there is already significant height loss due to large top clearance, that could be balanced with a different extruder setup possible with an open top. As I have already written, offseting the side plate by 4~5mm to clear the carriage may allow to remove the side plate peripheral cut, improving the plate stiffness.

Edited 1 time(s). Last edit at 05/25/2015 05:22PM by PRZ.

Just a note, the price for the micro delta is vat included. Just for sake of comparison.

Edited 1 time(s). Last edit at 05/25/2015 05:17PM by Fpex.

---

----- Making the world smarter @ www.xetal.eu
----- Helping entrepreneurs @ www.fralke.com

I have not initially noticed the comment of David about the belt tensioner being on the top of the frame, but that may explain the huge top clearance. Freed room in 'motor on top' setup as the arms don't need to go down the floor.

I've just measured the diameter of the Fisher bed, now that we have it here back from the show. The diameter, not including the 'lugs' for the bed probing, is 170mm. So it may be possible to fit a 160mm diameter heated bed, though I have no idea on the actual size of the linked 160mm heated bed. It would be sensible to change the acrylic bed for an insulator (eg MDF), with the heated bed on top. The 150mm figure comes from the printable area.

In it's current form, the acrylic print bed is covered with masking tape, with a light coat of PVA wood glue on top. This works well with PLA. There are plenty of other compounds around that work to keep PLA stuck to a non-heated bed, which we're looking at, and the smaller print area means that warping isn't as big an issue anyway. Ultimately, the Fisher is designed as a budget machine; fitting a heated bed as standard will increase the price, but it may be possible to have it as an additional kit.

PRZ: I'll send you the correct data (and post it here) when I have it - some of dimensions you have are incorrect. As to the design, this is going to change over the next couple of weeks as it is tweaked, but the motors and electronics will stay at the bottom. For motors at the top, you still have to have wires running from the bottom to the top, eg high current power wires and communications cable (USB and/or ethernet) running up, heated bed (if fitted) power, bed thermistor and bed probing (if using the current system) running back down. Also, the whole printer becomes top heavy, and less stable. While you can reduce the length of the belt (which is very cheap), this will increase the number of parts required, to have the idler half way down the axis. Our belt adjuster at the top is very low profile, and we may be able to reduce the wiring to the top anyway. As for the 'huge top clearance' of 40mm, yes it gives clearance for the Bowden tube and other hot end parts, but perhaps also give a little more build height than the suggested 180mm. The side panels may yet be altered, and while the arms are close, they are not touching.

There are other advantages for having motors, and electronics, at the bottom; one is that it is easier to screen these from ElectroMagnetic Interference (though not many other companies in the RepRap world seem to worry about this), and that only fairly low current wiring runs up the machine (the hot end heater drawing the most current). And as for easier maintenance, once the electronics and motor are assembled, what needs maintenance? It's the hot end and moving parts that generally need it, which are all accessible. Sorry if the Fisher is not exactly the printer you want, but we are driven by the design criteria we have chosen! Please understand that we have some expertise in the design of 3D printers...

Ian
RepRapPro tech support

mmm... Ian, I feel some sort of discontent in your answer. There's no need for that. By essence, such printer in beta phase will be bought by tinkerers/engineers. As for your experience in printer design, for sure, but a lot of the improvements were carried by the users, so even if users comments are far from being always valid or pertinent, using argument of authority is not ok, especially in this forum.
The available data is scarce, hence the questions.
Also, getting the data from photos was sort of a deal, that I have already carried for the Tiko
I don't see the problem to compare your printer (in beta phase) to the closest one which have been sold by a few hundreds.
It's clear that heavy top is bad. however, as your design is clearly much more sturdy than the micro-delta for similar size, this is looking as something doable.
Another point could be that if your printer is fast, as allowed by its powerful processor, inertia may be more critical than for the micro-delta.

You clearly have a point on EMI. Are you using acrylic with shielding additives ?
I am also in heavy doubts about overall CE compliance of all printer kits available there, not only for EMI. Fire safety is a concern, also. Kits are yet in a legal "grey area" for CE compliance. Till someone make a trial after a fire.

I've ordered this printer as my first printer, in beta phase, so I think it does have some merits, so please don't take badly my questions and comments.
I've gone for a beta because, if I am new to printers, I have background in DIY CNC and have taken a lot of time to analyse and check available info about printers, and especially delta, which I feel will got a lot of success after main problems will be solved (software and calibration).

My feeling is this printer have the potential to be your best seller, if well debugged.

Edited 3 time(s). Last edit at 05/26/2015 09:28AM by PRZ.

Quote
droftarts
... and we may be able to reduce the wiring to the top anyway...

You could use a common wire for heater +ve and hot end fan +ve. With care over routing, you could use a common ground side wire for the 3 endstop switches. Beyond that, I've long held the belief that it makes sense to do the heater and hot end fan control on a small board on the hot end itself, especially for multiple nozzle systems - although in the absence of a Z probe mounted on the hot end, this would give little or no saving in the number of wires on a single nozzle system.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

PRZ: Sorry if my reply was robust; however, your post sounded like you were telling us our design was completely wrong! While it would be great to have a fully-featured printer, this comes with an associated cost, and our main objective is a much more affordable printer. As I have said, we will be tweaking the design over the next few weeks, but we're not likely to make major changes.

Yes, there are user tweaks and modifcations for all our machines... or you can use it unmodified. We realise that people buy these machines to do a range of different tasks, so they may customise them to their own requirements. It's the nature of RepRap.

Ian
RepRapPro tech support

Quote
dc42
You could use a common wire for heater +ve and hot end fan +ve. With care over routing, you could use a common ground side wire for the 3 endstop switches. Beyond that, I've long held the belief that it makes sense to do the heater and hot end fan control on a small board on the hot end itself, especially for multiple nozzle systems - although in the absence of a Z probe mounted on the hot end, this would give little or no saving in the number of wires on a single nozzle system.

With the cost of microcontrollers being as low as they are, it certainly makes a lot of sense to put the controlling electronics near the part they are controlling instead of on the main controller. Each board would need only power and bi-directional serial signalling wire(s) - and a single control signal wire could be used in star or daisy-chain to feed all boards with a suitable protocol. That way the main control board would not have to have any power components at all. Parts of the control algorithms could be transferred to the control boards - e.g. the motor boards would be fed the basic information regarding each move and will have its own firmware that issues the appropriate stepping signals to carry out that move. The heater boards would only need be commanded to set a temperature or report its temperature and the bang-bang, PWM and/or PID algorithms would be implemented by the boards.

A huge advantage is that it is almost infinitely scalable. Add an extruder, and you can add another couple of controller boards (one as a motor driver, another for the hotend), and set jumpers, switches or program them to give each a unique tool number. The controller board could poll all addresses at powerup to discover what its hardware configuration is. Wiring becomes very simple.

It would even be possible to use the power wires to carry the signalling, and an existing protocol could be used (the CAN bus protocol comes to mind, as there are cheap microcontrollers with CAN bus capability on-chip).

Repairs are also easier - if you blow a motor controller you replace only one board instead of needing to do component-level repair or replacement of an entire controller or expansion board. The controller becomes more simple with no need to have a large number of I/O on chip.

Dave

I understand PRZ. I am building a Kossel XL by replicating what a friend is doing, and the results he gets are ok bit not what we both expected. Most design we see and try are not very stable and a crash on a Kossel can be very very catastrophic. The fisher is small, but seems rather robust in its frame and it makes very attractive. Even I am considering taking a beta one and see how it evolves. If only I could find a spot in the house,

---

----- Making the world smarter @ www.xetal.eu
----- Helping entrepreneurs @ www.fralke.com

Quote
Fpex
I understand PRZ. I am building a Kossel XL by replicating what a friend is doing, and the results he gets are ok bit not what we both expected. Most design we see and try are not very stable and a crash on a Kossel can be very very catastrophic.

[Even more off this forum topic] What sort of problems are you having?

Here is a pic of my Duet-controlled Kossel, which started printing yesterday and is working very well so far.

Quote

The fisher is small, but seems rather robust in its frame and it makes very attractive. Even I am considering taking a beta one and see how it evolves. If only I could find a spot in the house,

Yes, the one I saw at the show seemed very robust. I couldn't make the frame wobble by pushing the top.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Apart from not great choices in the original Kossels design, at least our Kossel xl are too sensitive to being touched (and go out of axes with the effector slamming somewhere). Mind that we are using open beam and print really really fast. The result is that my Kossel is on hold till the one of my friend gets dismounted and several plastic pieces redesigned. Except that now all our FDMs are down (all but the ormerod suffered severe head failure either from experimentation with a new head of a thermistor slipping out of the a j head ... ) and work gives little time to fix them :-(

I have been looking at adding panels in a way that it would not add too much weight on top causing oscillations. But as I am not mechanical designer, it is not easy.

PS: looks like you are using the very same design. Most defects came out for us after a month of intense use (for some work stuff) at my friend pkace. In fact we are bulding two to try and play around with both at different times.

Edited 1 time(s). Last edit at 05/26/2015 02:09PM by Fpex.

---

----- Making the world smarter @ www.xetal.eu
----- Helping entrepreneurs @ www.fralke.com

Quote
Fpex
PS: looks like you are using the very same design. Most defects came out for us after a month of intense use (for some work stuff) at my friend pkace. In fact we are bulding two to try and play around with both at different times.

I am using metal corners and a mains-powered heated bed, and the dimensions are of of my own choosing, but apart from that the design is fairly standard. I've been printing with a smaller version of this machine for the past 4 months without difficulty, but the use has not been intensive. I am expecting the Traxxas joints to wear out eventually.

Edited 1 time(s). Last edit at 05/26/2015 03:46PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

dc42, what dimensions you can print?

Quote
Treito
dc42, what dimensions you can print?

300mm diameter, and >= 480mm height all over the printable area. Horizontal extrusions are 355mm, verticals are 1m.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Our printers, except the Zcorp because is in a remote location, are all used a lot. lately easily 18 hours a day for weeks.

Btw what profiles you use (I think you mentioned already) as our problem could be using the open beams that are lighter than we anticipated.

---

----- Making the world smarter @ www.xetal.eu
----- Helping entrepreneurs @ www.fralke.com

Quote
Fpex
Btw what profiles you use (I think you mentioned already) as our problem could be using the open beams that are lighter than we anticipated.

I use Motadis 2020 for the verticals, 2060 for 2 of the base horizontals, and 2020 for the remaining 5 horizontals. I am using the I-profile, although it wasn't the best choice because it doesn't fit the Robotdigg aluminium corners unless the slots are widened a little.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I love delta!!!Looking farword to the details of Fisher.

Quote
dc42

Quote
Fpex
Btw what profiles you use (I think you mentioned already) as our problem could be using the open beams that are lighter than we anticipated.

I use Motadis 2020 for the verticals, 2060 for 2 of the base horizontals, and 2020 for the remaining 5 horizontals. I am using the I-profile, although it wasn't the best choice because it doesn't fit the Robotdigg aluminium corners unless the slots are widened a little.

Just for a little more info Mine is essentially the same as DC42's except I used OPENBUILD's Vee-Slot extrusion which fit's nicely with the metal Robotdigg Corners and my horizontals are 500mm long.

The frame is extremely rigid but would be very easy to add cross bracing to the sides should it be required.

I have slightly less height than Dave due to using the Robotdigg carriages which take Closed loop belts and the longest I could find are 1524 mm which gives me a build volume of 380 mm diam max by approx 300 high.

Doug

The open beams I use at the 15x15, I guess that could be the source of problem of wobbling. Probably with some cross panels of sacrificing height it could be sorted out,
Why the motor (mounted at the base) turn after some heavy use at high speed is something else, we need to change the plastics for it.

---

----- Making the world smarter @ www.xetal.eu
----- Helping entrepreneurs @ www.fralke.com

Here are the current Fisher dimensions. I'm not quite sure of the terminology PRZ used, so ask if there is anything that is not clear, or if there are any other dimensions you need. Please remember all of these dimensions are subject to change!

Height is 426mm.
Side panels are 3mm acrylic, 190mm x 426mm.
The opening in the side panel is 55mm from the bottom (same height as bed surface), 138mm x 200mm.
The centre of the panels are 89mm from the centre of the printer. All other acrylic parts are 5mm thick.
The bed is 170mm in diameter (ignoring bed levelling lugs).
Smooth rods are 8mm diameter, 45mm spacing between centres, axis plane of rods is 123mm from centre.
Arms are spaced as the rods, at 45mm. Length is 160mm.
Horizontal projection radius of arms (not sure exaclty what you mean, but I think...) 81mm
Carriage offset is 19mm. Effector offset (from centre to axis plane of arms) is 23mm.
Nozzle tip is 21.5mm below the axis of the effector.
The top height of the carriage movement, where the microswitch is tripped, has the effector just about level with the top of the opening in the side panel. With the nozzle 21.5mm below this, the maximum nozzle height is 178.5mm at the moment, and gives a maximum full radius (150mm) build height of around 155mm. From there, the build area would cone up to 178mm.

Ian
RepRapPro tech support

So, I was off from a large value (15%) for the height, but more accurate on other parameters. I should have drafted the vanishing point, which I had not done. I forgot the scale compensation between the top and the bottom of the image for the vertical dimension.

That make the Fisher in quite the same envelope as the micro-delta, but with columns on a larger diameter (on the micro-delta, the carriage protrude somewhat outside the columns).

So, my conclusion about the height margin was wrong.
I've modified the parameters in the OpenScad simulator and rebuild the film (done with VirtualDub), here :
[github.com]

With these design parameters, the margins are thin and 150mm diameter is reached with arm at 12.5°, so this is the ultimate diameter with this setup. Longer arms may allow larger diameter, but shall reduce the available height. Tradeoffs, as usual.
It is difficult to have any definitive opinion of what is doable without having the fan setup, which is quite critical on a so small machine.

Good play with that, Pierre

Sally has just posted some new videos of the Fisher: [www.youtube.com] and [www.youtube.com]

Ian
RepRapPro tech support

Quote
droftarts
Sally has just posted some new videos of the Fisher: [www.youtube.com] and [www.youtube.com]

Ian
RepRapPro tech support

Nice! A few comments:

- You could reduce the Z dive height to make the bed calibration faster. It's adjustable in my fork, see the H parameter of the M558 command. In release 1.09a I made it relative to the Z probe trigger height and reduced the default value to 3mm.

- It looks like you are using 4-point delta calibration, and not the 6- or 7-point that my firmware also supports. I guess this means that the build is accurate enough for 4-point calibration to be sufficient, which is a good sign.

- The print speed of the orthogonal calibration piece is slower than I would expect from a delta, apart from the first layer and the tall vertical part at the end. Have you tried increasing the print speed?

While we are on the subject of delta printer videos, here is a video of my delta running auto calibration using the mini differential IR height sensor: [www.youtube.com]. The printer is described in more detail in my blog at [miscsolutions.wordpress.com]. I'm sure it cost me lot more than £199 in parts though!

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

So so tempting!

---

----- Making the world smarter @ www.xetal.eu
----- Helping entrepreneurs @ www.fralke.com