Smartrap evolution, improvements...

HI Guys,
Thanks for the huge amount of support, I will try to reply to everyone now.

RTurnock.
I have Arduino software on my PC and I have some experience with flashing firmware. Luckily enough the correct Marlin version is already on the board installed by Smartfriendz so thats one problem sorted.
If there is confusion about if heat sinks need to be used then i think i might go with the side of caution and buy and fit them. I have an old CPU heatsink that can be chopped into small pieces to fit the pololus. Then I will look into tuning the pots to get the motors running smothly.
I will get a gluestick and give it a go.

Regpye.
It's a cold bed, no heated bed yet. So i will try a bit of PVA after the gluestick to see what works best.
I agree about using heatsinks just in case it goes bad.
I will look into Marlin and see what thats about.
Serge is busy at the moment drafting documentation to send to the guys who got the new J-head lite. There is no documentation yet. I feel like I am spamming him because DHL lost my package for a week and i had to email him loads of times already about where the package was. As a side note, the J-head lite looks crap compared to the Aluminium fin heatsink head but as long as it works I am happy.

BackEMF.
Flashing: As above i have the software needed and don't need to flash the board but I will look into it.
I decided to go with heatsinks just in case it goes bad.
I will add Bison D2 to the lsit of bed coatings that I need try.

OK, i am going to get stuck into the machine now and see how far i get. I have looked at the other self levelling designs and i think i get it.
Regards
Mike

Edited 1 time(s). Last edit at 08/21/2014 05:31PM by mikefiatx19.

OK, unit is running and the heatsinks are fitted.
I have some 'Jane Seymour Wonder Set Hairspray' thats been in the bathroom for years. It works ok for sticking PLA to the glass BTW but i will borrow some pritt stick from work tomorrow :-).
So, I started a first print and i noticed that the auto calibrate nozzle is acting as follows during the print start.

1- nozzle presses on the bed until switch clicks.
2- nozzle lifts up and stops about 2mm from the bed
3- nozzle heats to around 200c
4- once 200c is reached nozzle drops back down too far and presses on the bed
5- nozzle starts to drag around on the bed printing plastic

because the nozzle drops too far after levelling the first few layers are all at bed height, it's only after 10 layers or so that the nozzle lifts up beyond the glass and starts to build upwards. So the first 10 layers just end up being a mess. I think i need to play with the marlin config file to adjust the z switch offset so that the firmware sees home as 2 mm above the click.

I tried to use the triple check levelling g29 in gcode so that Marlin homes 3 times on 3 parts of the bed to calculate the plane.
After the G29 the Zaxis was at 2.01, I then had to move Z upwards to 3.43 before it cleared the bed. The difference was 1.42 so i then used M555 Z1.42 and then M500 to save the settings but it made no difference. I will keep playing with it until it works.

I also thought (i could be wrong) that as the auto levelling nozzle has to move to be able to press the switch, i am losing some stability in the nozzle and the machines accuract may be affected. Does anyone using the no servo auto levelling function have views on this? Is the servo version better?

Well, it's late here so i am done for now.
Regards
Mike

Edited 2 time(s). Last edit at 08/21/2014 06:31PM by mikefiatx19.

Quote
mikefiatx19
HI Guys,
Thanks for the huge amount of support, I will try to reply to everyone now.

RTurnock.
. Then I will look into tuning the pots to get the motors running smoothly.

When you adjust your pots you will need a multimeter that can read in the low voltage range setting.
I have found the best and easiest way to do this is to get a second set of multimeter probes and modify them so that the red positive probe becomes a screw driver that you adjust the pot with, and the black negative probe becomes a flying lead with a pin on the end that can plug into one of the GND sockets of a spare stepper driver socket (all ground pins are connected)
When you adjust the pot you will get a direct reading on the meter, because the screw driver probe will be connected to the centre connection of the pot. It makes life a lot easier doing it this way and you can adjust all the pots without having to mess around with clips and other tools that could cause a short by coming off or just being in the way and not need to change the GND pin as they are all connected

---

[regpye.com.au]
"Experience is the mother of all knowledge." --Leonardo da Vinci

Still banging my head on a wall here.
EEPROM is enabled in Marlin but the offset is not making any difference to my prints. The head is just dragging through the plastic. As a test I even set the head offset to M555 Z4 as i knew this should be too high, the program just ignores it though.
Serge is not responding to my emails which is nice.
And there is no method to manually adjust the bed at all so I am unable to print at all. Very annoying!
Mike

Have you calibrated Z-steps per millimeter?

When the nozzle lifts up and starts around two mm from the bed, is that actually where it should go according to your G-code? If I recall correctly, the instructions on the smartfriendz site will cause you to end up with a negative number, but you need to enter a positive one -- or the other way around. Anyway, try the levelling comand manually, and then M114 to see where the printer thinks the head is.

Worst case you might create a temporary workaround by using (EDIT: G, not M) G92 to tell the printer where the head is after levelling until the cause is found.

Edited 1 time(s). Last edit at 08/22/2014 03:55PM by ZeAllan.

Quote
mikefiatx19
Still banging my head on a wall here.
EEPROM is enabled in Marlin but the offset is not making any difference to my prints. The head is just dragging through the plastic. As a test I even set the head offset to M555 Z4 as i knew this should be too high, the program just ignores it though.

Hi Mike,

I scratched my head over the z offset for several evenings. I just added a Z offset to everything in slicer, until I found the variable zprobe_zoffset in marlin_main and traced where it was used and set.

After updating the z probe offset you have to store it in EEPROM immediately (M500). If you or the host software read anything from EEPROM (M501) before storing the update, your updated value gets erased. Try this

M501

to reads the current EEPROM configuration. Look for a line such as

M206 X0.00 Y0.00 Z-1.80 ZProbe-1.95

That "ZProbe" value is what you need to change. Set it with

M555 Z-1.5 ; or whatever you want the new offset to be.
M500 ; immediately store the new offset

Then try an M501 again and look for a changed ZProbe value. That was the only thing that worked for me.

I hope this helps

HI CellJeffe,
Thanks that helped a lot. I actually needed a minus offset for some reason but I have it printing now.
I do still have an issue with the fact that the bed closer to the machine is a good height but the part further away above the roller bearing is lower. The offset doesn't take into account the angle of the bed. I can only print small items or the ends wont stick due to the bed being lower at the ends of the parts.
Regards
Mike

____________________________
Issue above resolved, i printed an adjustable leg for the bed support. I also think i figured out the 3 point probing. For some reason the ramps board was not recording the data when using the G29 and the M500 commands. I re-flashed the board and it works now.
Very nice design and very accurate first prints.

Mike

Further update to the info above. For some reason i still have issues with the uneven bed. The design seems unstable and the bed keeps having levelling issues. There is also no option to manually adjust the height so i am a bit stuck.
Is there any good solution to this?
The G29 offset seems to just point out the fact that the bed is uneven and on my Smartrap it does nothing to compensate during the build. My offset is around 1.2 mm at the far point of the bed. According to one of Serge's videos, the machine should follow the angle of the bed while printing.
Sorry for being a pain in the ass noob.
Regards
Mike

Hi,

I'm not entirely sure about the problem, but I think I've been fighting the very same problem today. It seems that Z_PROBE_OFFSET_FROM_EXTRUDER_DELTA_X might be the source of this behaviour. As far as I can determine, it tells the firmware how much further down the probing is pushing the outer side of the bed before activating the endstop, because of the the way the bed is better supported on the inner side. If your contruction is more solid than assumed (the default is .5 mm), the far end is not pushed that far down in reality, and therefore the printer thinks it needs to compensate more than it does, and ends up printing too high on that side.

I tried setting Z_PROBE_OFFSET_FROM_EXTRUDER_DELTA_X to zero and went from having the floating filament to having the first layer just a bit too flat on the bed, while the side closest to the X endstop was the same in both instances. I guess that means tomorrow I'll have to try with a value of .1 to .2 and that'll be it.

/Allan

Quote
ZeAllan
Hi,

I'm not entirely sure about the problem, but I think I've been fighting the very same problem today. It seems that Z_PROBE_OFFSET_FROM_EXTRUDER_DELTA_X might be the source of this behaviour. As far as I can determine, it tells the firmware how much further down the probing is pushing the outer side of the bed before activating the endstop, because of the the way the bed is better supported on the inner side. If your contruction is more solid than assumed (the default is .5 mm), the far end is not pushed that far down in reality, and therefore the printer thinks it needs to compensate more than it does, and ends up printing too high on that side.

I tried setting Z_PROBE_OFFSET_FROM_EXTRUDER_DELTA_X to zero and went from having the floating filament to having the first layer just a bit too flat on the bed, while the side closest to the X endstop was the same in both instances. I guess that means tomorrow I'll have to try with a value of .1 to .2 and that'll be it.

/Allan

You got it! It took me a bit to figure it out myself.

Ha, Figured it out,
I put G29 into the custom code in Slicer just after the G28 thats already there and now its all good. I was using a manual G28, G29 before each print. I popped G29 into the custom G code and now the z is following all the uneven bits. I didn't realise that slicer needed to do the G29 for it to work.
Soooo coool
Mike