Cherry Pi III Is Available

Two comments on the above:

1. Checking hot end temperatures.

Many general purpose multimeters include a thermocouple probe for measuring temperatures. You can remove the filament and feed the thermocouple probe into the top of the hot end, right down inside the nozzle, to get an accurate temperature reading.

2. 2-colour printing.

Andy, thanks for your comments on the Cyclops. I didn't realise that you couldn't feed it with just one filament without it blocking up. I am considering a Chimera for my enlarged Kossel. I have some experience with 2-nozzle printing on my Ormerod, see [miscsolutions.wordpress.com]. I have no problem with the unused nozzle oozing or dripping, it doesn't. I set the standby temperature of the unused nozzle to 150C for PLA and use 10mm retraction (the default in slic3r, which I haven't tried changing because it works). My tool change macro files are set up to wait for the new nozzle to reach printing temperature, but not to wait for the old one to reach standby temperature (which takes longer). However, the non-printing nozzle does create slight marks when it crosses a piece of the print that has just been printed by the printing nozzle. The closer the nozzles are, the more this is a problem, because the less time newly-laid-down filament has to cool before the non-printing nozzle crosses it. Because of this, I redesigned my dual nozzle conversion on the Ormerod to increase the spacing as much as I could (to 22mm) without a total redesign of the hot end assembly. So I am concerned that the small nozzle spacing of the Chimera will make this problem a lot worse.

---

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].

That's interesting feedback on true dual nozzle printing. I may just turn my Cyclops into a Chimera. I believe E3D sell the individual components to accomplish this. If I do carry out this mod I will feedback on your concerns about nozzle proximity.

Andy, I dare you to do a bigger slice of Pi using a Kraken.

Quote
pugzor
Andy, I dare you to do a bigger slice of Pi using a Kraken.

Already thinking of that (Got one in a draw upstairs) but water cooling on that scale bothers me slightly. All that water has to be moved with the effector so may keep it for the CoreXY!

Doug.

ps the nozzle spacing on all the multi extruder sets from E3D are 20 mm as far as I know (certainly the kraken is).

Haha nice Doug. I was actually thinking of (one day, since I've pretty much had every problem possible so far) making a large envelope Cartesian using a Kraken. Well, probably a Chimera to start off with but with allowances for a Kraken. If nothing else it'd be cool to have a number of nozzle sizes pre-installed and calibrated, or, if doing super large prints, just having the ability to switch to another spool. Even with a Kraken you might have 1 nozzle for support material, 1 nozzle for detail, and 2 nozzles for high-flow with plenty of extra material. Unfortunately I think there's too much weight in a Kraken to be used on a delta, especially when you consider water cooling. I guess you could always have some kind of suspension system to take 'weight' off the print head - I've actually got a little piece of elastic cord around my wiring / Bowden tube just to remove some of the strain when it's at extremities.

Hmm, had a look on Aliexpress and came across these... [www.aliexpress.com] Seems like a good deal? For me, AUD $26 including shipping. Same guy has a whole heap of good bits and pieces at reasonable prices too.

What bothers me about a Kraken is the huge number of cables it needs. A simplistic approach would need 2 per thermistor, 2 per heater, and 2 for the hot end fan, making 18 in all. Some saving could be made by using a shared thermistor ground and a shared +12V for the heaters and fan, but that's still 11 wires, of which one (the shared +12V) would need to be extra thick to carry the current to 4 heaters.

So I have it in mind to design and manufacture a board that mounts on the hot end and does all the hot end temperature control. It would need 2 thick wires for +12V and ground, and about 3 thin wires e.g. for transmit, receive, and signal ground. The main printer electronics would transmit the required temperatures and occasionally thermistor and PID parameters to the board, and the board would send current temperatures along with any error status (e.g. open circuit thermistor). Or I could use thermocouples instead, so that no thermistor calibration is needed.

I would also make provision for connecting a Z probe to the board, so that separate Z probe wires would not be needed.

Does anyone think this is a good idea? The main thing stopping me is that I expect the market for Krakens is quite small, so I might never recoup the time and effort needed to develop this. I am not retired, so I still need to make a living.

The same thing could be done on a smaller scale for 2-nozzle printers, but the benefit in reduced cabling would be lower.

---

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].

Quote
dc42
What bothers me about a Kraken is the huge number of cables it needs. A simplistic approach would need 2 per thermistor, 2 per heater, and 2 for the hot end fan, making 18 in all. Some saving could be made by using a shared thermistor ground and a shared +12V for the heaters and fan, but that's still 11 wires, of which one (the shared +12V) would need to be extra thick to carry the current to 4 heaters.

So I have it in mind to design and manufacture a board that mounts on the hot end and does all the hot end temperature control. It would need 2 thick wires for +12V and ground, and about 3 thin wires e.g. for transmit, receive, and signal ground. The main printer electronics would transmit the required temperatures and occasionally thermistor and PID parameters to the board, and the board would send current temperatures along with any error status (e.g. open circuit thermistor). Or I could use thermocouples instead, so that no thermistor calibration is needed.

I would also make provision for connecting a Z probe to the board, so that separate Z probe wires would not be needed.

Does anyone think this is a good idea? The main thing stopping me is that I expect the market for Krakens is quite small, so I might never recoup the time and effort needed to develop this. I am not retired, so I still need to make a living.

The same thing could be done on a smaller scale for 2-nozzle printers, but the benefit in reduced cabling would be lower.

one thing to note Dave is that the Kraken does not have a fan on it it is watercooled so there are 2 tubes running from it to a pump.

And I don't think you would recoup the Time Effort or expense as market is very low.

Doug

good idea tho maybe you could speak to the E3D guys and see if they would be interested in doing anything with you

Speaking of large format multi extruder printers I've just completed this PrntQB V5 (that's Print Cube) as a custom build for a client :-



It's a parallel belt core XY with a 400mm x 400mm x 400mm print area. It has a silicone mat (1000W 240V) as the heat bed with a 1.5mm aluminium spreader and 3mm glass. The design was conceived to remove the need for ground rods and linear bearings. All the bearings are mini V wheels ala Cherry Pi. The bed is driven by two NEMA 17 steppers and 8mm leadscrew. The bed is stabilised at both ends by a core XY Spectra line arrangement using guitar tuning pegs to tension/adjust bed level. There is a servo mounted switch for auto bed levelling too. this has the Cyclops installed at the clients request. I built a similar machine for myself that has the same specs but this one benefited from being the second build and is much more refined!

That's pretty awesome!!

I would be needing a new Girlfriend if I was to build anything this size.

Quote
dc42
So I have it in mind to design and manufacture a board that mounts on the hot end and does all the hot end temperature control. It would need 2 thick wires for +12V and ground, and about 3 thin wires e.g. for transmit, receive, and signal ground. The main printer electronics would transmit the required temperatures and occasionally thermistor and PID parameters to the board, and the board would send current temperatures along with any error status (e.g. open circuit thermistor). Or I could use thermocouples instead, so that no thermistor calibration is needed.

I have floated the idea of an effector-mounted PCB managing the hotend PID, fans, lights, etc via a serial link to the main controller. I also wanted to include a tri-axis accelerometer, to detect rod separation or structural failure. Throw in a kill wire, watchdog timer, and a smoke/heat sensor to complete the safety package. My only oversight was not thinking that dual/tri/quad extruders would make an even stronger case for an intelligent breakout board on the effector like you just proposed.....

Edited 1 time(s). Last edit at 03/31/2015 08:31PM by vreihen.

Speaking of larger format printers, has anyone thought of laser cut aluminium plate for either 2 ends on a cartesian or top and bottom on a delta?

I've been posing the question to a couple of laser cutters around the place and the price doesn't seem that bad. Obviously more expensive than linear rail but a piece of 500x500x8mm sheet with a few holes processed looks like it'd be under AU$100. For something closer to the size of a CPIII, it'd be around AU$140 for both sheets (400x400x6).

If you're using quite hefty linear rail with a couple of mounting points, I'd imagine this would virtually eliminate any movement. Could probably get some nice patterns cut out of it too to reduce weight (8mm is heaps thick!). There'd also be a good amount of time saved in assembly, more mounting points, etc, etc.

It actually seems like a more cost effective idea for someone in AU where Openbuilds linear rail is quite costly. If I was to build a super large cartesian with a print area of 700x500, it'd cost me around $320 for two end plates (700x700x8mm) + $180 for a print bed (800x600x8mm, probably laser cut to being honeycombed or something for weight reduction whilst maintaining rigidity). Print bed would probably get a thin glass panel on top too. Only around 5m of linear rail would be needed after this but potentially reduce a lot of joins and at least 6m of rail.

All my designs are based around Bosch/Rexroth or Motedis extrusion which only costs £4.00 per metre in the UK. Not sure what Motedis charge for shipping to Oz, they are a German company. My Bosch/Rexroth supplier is KJN in the UK. I would seriously consider going down the coreXY route rather than using Cartesian kinematics. It's the most efficient system in terms of frame size to build plate size and, next to a delta, gives the best opportunity for fast head speeds. The machine pictured above prints quite happily at 70mm/s. The main costs in my previous coreXY designs have been for precision ground rods and linear bearings but I've now managed to eliminate those. I've also been comparing the merits of a leadscrew and lead nut Z lift compared to using a cheap M8 threaded rod and UHMW leadnut ( homemade by tapping an M8 thread into a 6.2mm hole). So far I can't see enough difference to warrant the high cost of a pair of 500mm leadscrews and two bronze leadnuts (about £40).

Quote
AndyCart
All my designs are based around Bosch/Rexroth or Motedis extrusion which only costs £4.00 per metre in the UK. Not sure what Motedis charge for shipping to Oz, they are a German company. My Bosch/Rexroth supplier is KJN in the UK. I would seriously consider going down the coreXY route rather than using Cartesian kinematics. It's the most efficient system in terms of frame size to build plate size and, next to a delta, gives the best opportunity for fast head speeds. The machine pictured above prints quite happily at 70mm/s. The main costs in my previous coreXY designs have been for precision ground rods and linear bearings but I've now managed to eliminate those. I've also been comparing the merits of a leadscrew and lead nut Z lift compared to using a cheap M8 threaded rod and UHMW leadnut ( homemade by tapping an M8 thread into a 6.2mm hole). So far I can't see enough difference to warrant the high cost of a pair of 500mm leadscrews and two bronze leadnuts (about £40).

Andy

Have a look at Dold (Partner with Motedis) I got 3 1mtr lengths of TR 10x2 leadscrew from them for around a £5 each Dold Leadscrews and picked up some bronze nuts locally for around £3.50 each Bronze lead nuts (this is an example appears they are out of stock of the 10mm ones) Dold also do 8mmx1.5 at 5eu for a 50cm length and the do Hex nuts for them at Eur 6.70 each (easy to print a holder for these as well)

HTH Doug

The good news is that I have finished my Cherry Pi IIIs, with the exception of the heated bed that's still winging its way from China. With this in mind I thought I'd try to see whether things are moving in the right direction when commands are sent from RepetierHost.

The not-so-good news is that the motors aren't moving as I expected, and I'm having some difficulty sorting it out. I suspect that one of my stepper drivers isn't working properly, but I'm finding it hard to diagnose the problems while all the motors are moving in strange directions! For example, I send a command to advance the extruder, and 2 of the axis motors lock up and make a strange noise. I have checked all my connections, the end-stops and motors are plugged in correctly, I have applied Andy's configuration.h file to a copy of Rich Cattell's Marlin, compiled and uploaded it to the Arduino board, and I've set up the printer in RepetierHost according to the instructions that Andy posted recently.

It doesn't help that I am unfamiliar with the way that a Delta printer works - the Prusa i3 is self-evident once you've got the configuration right, but I don't know what to expect from a Delta. For example, if you send a 'home' command, where does the effector go? Down to the print surface, or up to the limit switches? Is there any general information on-line to help me? (not necessarily for the Cherry Pi)

Is there some firmware available that would allow me to drive one motor at a time, to aid with diagnosis? For example, could I load the firmware for my Prusa i3 so that I can operate the X, Y and Z motors independantly? At least then I can try to move each column's motor in very small increments to see if everything is behaving sensibly - obviously I wouldn't try homing any axis, or do anything else complicated, as that would be all wrong for the Cherry Pi.

It's early days so I'm not too downhearted yet, but I'd like to see that it's likely to work when I finally try to set it up.

Cheers,
David

Hi David,

If your motors are just making a noise and not moving you might want to adjust the driver to give more power to the motor. The arms on home should go up to the limit switches.

Matt

@ David J

As you intuited in your post, you could load the Prusa i3 firmware on to the Arduino you have on your delta, and test the motors (don't forget to adjust the steps/mm for the Z axis) one by one. DO NOT home, just move each axis by 1mm +/- at a time, that at least should allow you to check the steppers. And as Matt suggested, check the Vref on your stepper drivers so that each motor is getting around 1A.

Once you have made sure that your steppers are all moving in the same direction, reinstall the delta firmware and start on the manual calibration procedure.

I just wanted to point out that the delta support in RepRapFirmware (for Duet electronics) has the facility to move individual motors built-in, precisely to make testing easier. I'm surprised that Marlin doesn't.

---

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].

@Dougal1957

Thanks for the link Doug. I'll definitely give Dold a go. Ridiculously cheap.

Andy

Quote
AndyCart
@Dougal1957

Thanks for the link Doug. I'll definitely give Dold a go. Ridiculously cheap.

Andy

There very quick as well usually within a week there smooth rods are good as well.

Doug

Thanks Andy and Doug, I've just ordered a shed load of stuff for a CoreXY from Dold :/

I haven't finished the Cherry Pi yet.

Quote
David J
It doesn't help that I am unfamiliar with the way that a Delta printer works - the Prusa i3 is self-evident once you've got the configuration right, but I don't know what to expect from a Delta. For example, if you send a 'home' command, where does the effector go? Down to the print surface, or up to the limit switches? Is there any general information on-line to help me? (not necessarily for the Cherry Pi)

A Delta printer homes upwards into the endstop switches. Once it completes homing, it sets X=0, Y=0, Z=Z_max. The point in the dead center of the bed is X=0, Y=0, Z=0, and both X and Y can go negative from there.

When testing a Delta's steps/mm on each tower, a pure Z move (don't move X or Y) should move all three carriages the exact same amount. No need for special firmware. The most frequent problem with steppers is that some folks forget to put the micro-step jumpers onto their controller boards. Some manufacturers (SainSmart for one) don't even ship the loose jumper caps with their boards. Another commonly reported problem is that some Chinese jumper caps are missing the metal inside that connects the pins together, although I sometimes wonder if this isn't a made-up excuse by someone trying to save face because they didn't put the jumper caps on. I assume that you have already done the most important thing...adjusting the stepper driver current using a multimeter (and not counting turns, listening to sounds, or any other seat-of-the-pants method that does not involve a multimeter and hard numbers).....

#vreihen

I have to admit I used the turn it until it moves calibration But I would like to know how to do it properly, I haven't got the first clue how to measure the amps going to the motor, do you know of a guide that would help to show how it is done correctly. I do have a multimeter and I can just about use it for volts and resistance.


Thanks

Matt

This info is as good as any: RepRap wiki - look for the section called "adjusting and testing".

Also, have a look at this : nophead on the RepRap forum

[NOTE: look at the post after this one and work out which drivers you actually have!]

Edited 2 time(s). Last edit at 04/02/2015 08:22AM by David J.

This information is probably more useful, since it applies to the A4988 Pololu stepper drivers that are more common than the Stepsticks:

http://www.reprap.org/wiki/Pololu_stepper_driver_board

I used 0.4V on my A4988's, which gives 1.0A of output according to the formula. The motors do not need to be in motion to do the measurement.

My own caution is that some people slip with the screwdriver while turning the trimpot and short out their stepper drivers. If you don't have a plastic/ceramic screwdriver, please remove power before turning the trimpots.....

That's a useful page - are the Chinese knock-off stepper drivers usually Pololu clones? They seem to have the same component layout.

I'm having a bad time with my RAMPS setup at the moment - had a good look at the bottom of the RAMPS shield and found many tiny solder blobs and streaks. Cleared all the ones I could see with a high-power lens, but I couldn't look under the connection pins for any more.

I'm not over-impressed with the stepper drivers either - for example, the trim pots don't have end stops, so it's possible to put the wiper into "no man's land" off the conductive track. My Prusa has G3D stepper drivers that are far better made (they also have a trim pot instead of R4, which allows me to set it as I see fit). I may well invest in a set of those...

(BTW: Don't lecture me on "you get what you pay for" - I know!)

Edited 4 time(s). Last edit at 04/02/2015 08:26AM by David J.

I'm making progress - all but one of my motors are now running properly, with the Z being the exception. At least that localises the problem and gives me a fair chance to fix it!

Many thanks to vreihen for the hint on setting the drivers - it looks like mine are Pololu clones. That, plus changing the RAMPS board over to a spare one, seems to have got me further down the line.

Now for some further investigation (but my main suspect is the Z motor's driver).

LATE UPDATE: Yes, it was the driver! All motors running as they should after putting in a spare driver.

Edited 1 time(s). Last edit at 04/02/2015 10:36AM by David J.

Eh guys

Just had an E-Mail from E3D announcing a new hotend for £21.50 see E3D V6 Lite ok wont go as hot or fast as a full V6 but will still do ABS at reasonable speeds.

No reason at all to go for Chinese copies now eh.

Doug

Quote
dougal1957
Eh guys

Just had an E-Mail from E3D announcing a new hotend for £21.50 see E3D V6 Lite ok wont go as hot or fast as a full V6 but will still do ABS at reasonable speeds.

No reason at all to go for Chinese copies now eh.

Doug

Damn... just bought an E3D v6 in 1.75mm! I could have managed quite nicely with one of the new ones...

Quote
dougal1957
Eh guys

Just had an E-Mail from E3D announcing a new hotend for £21.50 see E3D V6 Lite ok wont go as hot or fast as a full V6 but will still do ABS at reasonable speeds.

No reason at all to go for Chinese copies now eh.

Doug

I noticed these too last night when considering an E3D V6 conversion. I can still get Chinese copies for less than half of that, delivered, with thermister, heating resistor and fan.

There's actually a hotend that's produced locally to me called AluHotend. Has anyone used these? They seem to be JHead compatible but a little more refined than the standard design. They actually seem to be a bit of a hybrid between an E3D and a Jhead. [3d-industries.myshopify.com] They're about AU$50 each, so not cheap, but it'd be quick delivery for me and maybe an alternative if I didn't want to swap out my effector.

On the bright side I think I've found what the issue was with my hotend. The inner PTFE lining of the PEEK was a) too short by about 2-3mm and b) maybe didn't have the correct inner diameter (there was a LOT of space for 1.75mm filament to move, my guess was it was 3mm ID?). Those two combined meant that it was very easy for filament to get jammed from any point after going through the plumbing fitting. I ended up drilling out the PTFE lining so I could fit some of the 4mm OD 2mm ID PTFE tube in. There was still a slight jam just after the plumbing fitting so I put a bit of a chamfer on the inside of the 4mm OD PFTE tube so it was like a funnel. Seems to work well now. I'm only worried about it not lining up super well when it gets to the nozzle entry... but chances of success are a bit higher now I think. I'll give it a whirl and report back.

Woohoo!! It appears I had a bad board afterall - new arduino/ramps seems to have taken care of the Y axis max end stop issues I was having.

I'm still having issues with movement - I only have manual movement in the forward left quarter (z axis is fine, X only moves left of center, Y only moves to forward of center).

When homed repetier shows X at 115, Y at 115 and Z at 237, any ideas?

Edit 4: FINAL EDIT - found the issue, under repetier, under printer settings/Printer shape - instead of Home X and Home Y being set to MAX, I wasn't aware there was an option to set them to 0, once that was done, it now works in all directions.

Edited 5 time(s). Last edit at 04/04/2015 12:45AM by Zzyzxx71.