MakerFarm 12" Prusa I3V - my build, usage, and mods

Got tired of the z-axis end switch moving with the wire tie. A pair of very rare 3-48 x 3/4" machine screws fixed that issue.

The love/hate affair with PLA continues. Plate glass with the ubiquitous purple washable glue stick does wonders....it's too good!!! I broke a blade trying to remove a print. No glue = no stick and ick.

Trying something different - parchment paper. Yes, the one and the same baking paper. So I tried a print on plain parchment, unheated. No joy. Parchment heated - no joy here. Parchment, glue and heat - looks to be sticking. Now to see if I can get the first layer printed as this is the largest base print I have done to date on this.

So here's a trick I found out. The glue will come off easily in water. Yes I'm an idiot and didn't realize it was washable even after sitting for days!

So just run the glass print under water

Another thing I have started to do now: I bought some 1/8th inch aluminum plates and I just put those on top of the glass using the same clips. I did this because I will soon hookup auto leveling but also cause they are slightly flexible and it makes the pieces PPP right off.

For you I think just finding with water is the better solution. Nice to have a fresh glue layer each time anyways and if you bought some extra glass pieces you could swap them out and keep the prints going.

@milkypostman: Good ideas. Nice thing about aluminum...it heats up and cools off real quick. May have to give that a try.

Also, today I made a huge mod. Dropped the heat bed relay. I was finding that when the bed heater came on (bang-bang style), the first and second layers were messing up big time and getting patterns in them. So, I had some IRF4905 MOSFETs from a different project. I took one of them, and with the aid of a 10k, 1/2 watt resistor and a small heatsink, made a high current switch. The RUMBA board does low side switching and this is perfect for this. The IRF4905 turns on when the gate lead is grounded which makes this such an easy mod. Put the gate lead to the heat bed negative terminal and the middle leg of the device went to the positive side of the heat bed and the right leg went to +12. The 10k resistor is mounted between the outside legs and keeps the device off when not called for. So, turned on PID mode and retuned. So nice not to hear a relay clicking on and off. The next big question is, does the print clean up or not? I am currently using two supplies, one 50 amp supply just for the heat bed and a bank of batteries to run the rest (they were there, so use them). What a day.

Since the incident with the last thermistor and its fail, I decided to go on a hunt for replacements. So, I hit up my two competing parts vendors, since I don't have a Fry's Electronics here (pining for Fountain Valley), DigiKey and Mouser.

Digikey 495-2125-ND / EPCOS B57540G0104F000 100k/1% B25/B85=4066K
Mouser ***NOT STOCKED***

I found a spreadsheet on Thingiverse that does the calculations for Marlin. That is done so now to put the numbers into Marlin as a separate data type and I will be ready for the inevitable fail.

Also, on a side note, I remember reading a couple of posting chiming that you can't measure the resistance of a thermistor. I am sorry that I did not reply to the commentors but you can measure the resistance of the thermistor with an ohmmeter. There is no reason why not and there should be no issues with self-heating.

Now to buy a few.

Congrats to Gina Häußge and the OctoPrint/OctoPi team. Have my RPi attached to the printer. This is sweet and the setup with the OctoPi image was easy.

Is this an all-V-slot design?

Also, I have seen several people not use the relay; is there not a mosfet heater driver as a standard part somewhere?

Quote
RRuser
Is this an all-V-slot design?

Also, I have seen several people not use the relay; is there not a mosfet heater driver as a standard part somewhere?

1) All guide rails on the axes are 20mm X 20mm V-Slot.

2) The heatbed can easily draw 35 amps running. The RUMBA board that I have has the PSMN7R0-60YS mosfet as the heatbed device. The unit is rated at 60 amps and 117 watts dissipation. I did not feel comfortable with the design of board as the heatsink, the traces used, and the connectors to be passing this much juice through it safely, especially with the thermal cycling that can happen. The MakerFarm kit came with an external relay to be driven by the heatbed output but since it is mechanical, it has to be run bang-bang. The fan noise is bad enough and the clicking was annoying to the point that I could see the influence in the prints. I had a bunch of IRF4905's laying around. I put one on a small heatsink (a Celeron processor heatsink), added the 10k / 1/2w gate turnoff resistor and used it as a switch. Now I can run PID operation on the heatbed, not worrying about the RUMBA board failing.

I was running a print resolution test when I found out not all of it printed out. What it ended up is that Slic3r was not processing the areas. Apparently, this "feature" is going to be a limiting factor in the accuracy of prints. Now to build another calibration block to find out what Slic3r's limits are and go from there. I am going to have to possibly figure out a workaround, if possible, so I can really task the printer and see what it is capable of doing.

I was running into the issue of "fat corners". If I try to print a sharp cornered cube, the corners would bulge out from excess plastic caused by pressure delay in the hot end. I have found that by rounding the corners of the cube with a 0.8 mm radius, the transition in axes is rounded enough to get rid of the bump and leave clean corners. This, I am sure is system dependent so it will take some fiddling to work on your system.

This is a sweet piece of code:

// polyhole by nophead
module polyhole(h, d) {
n=max(round(2*d),3);
cylinder(h=h, r=(d/2)/cos(180/n), $fn=n);
}

Measured an M5 bolt with the calipers, plugged in the sizes and printed out with no compensation. The bolt went freely in. This will have to go into my code repertoire for when I need to make free holes. I figure if I knock of a few tenths off, maybe 0.4 mm off the diameter, the bolt will bite in with no issues.

How is the flex/stiffness of this larger 12" frame?

Quote
RRuser
How is the flex/stiffness of this larger 12" frame?

The biggest issue with flex that I have seen has been with the two cross pieces that join the Y axis rails together on each end. They are made out of wood and have bending issues when the stepper executes fast moves on the Y axis. The Z axis does not seem to have any issues that I can see at the moment.

As much as one would love one of these to scream when it comes to printing, it is just not possible because of the belts and their "give" For printing, I am running 15 - 30 mm/sec. This seems to be the happiest place for it between speed, and resonance/skipping issues.

I am trying the hardest thing yet I have had the printer do. I am doing a Koch patterned tealight off of Thingiverse ( [www.thingiverse.com] ) and it the spiral vase mode on Slic3r. Since the perimeter is one thickness wide, this will show up any deficiencies in the printer.

Quote
itchytweed
@milkypostman: Good ideas. Nice thing about aluminum...it heats up and cools off real quick. May have to give that a try.

Also, today I made a huge mod. Dropped the heat bed relay. I was finding that when the bed heater came on (bang-bang style), the first and second layers were messing up big time and getting patterns in them. So, I had some IRF4905 MOSFETs from a different project. I took one of them, and with the aid of a 10k, 1/2 watt resistor and a small heatsink, made a high current switch. The RUMBA board does low side switching and this is perfect for this. The IRF4905 turns on when the gate lead is grounded which makes this such an easy mod. Put the gate lead to the heat bed negative terminal and the middle leg of the device went to the positive side of the heat bed and the right leg went to +12. The 10k resistor is mounted between the outside legs and keeps the device off when not called for. So, turned on PID mode and retuned. So nice not to hear a relay clicking on and off. The next big question is, does the print clean up or not? I am currently using two supplies, one 50 amp supply just for the heat bed and a bank of batteries to run the rest (they were there, so use them). What a day.

I'm not pro with circuits(i.e. I've dinked around with arduinos but only ever low current/voltage stuff). But one of the heat bed relay terminals on my printer(also a 12" prusa from makerfarm) popped out of the pcb and caused a nasty spark. I unplugged it all and have been looking for a solution better than the relay. I found a blog post [www.grepstar.org] where he replaces his relay terminals with bolts and ring connectors, he says that because he only prints PLA the temperature isn't an issue. I print ABS, and so I'm worried his solution won't work well when there is more current drawn. So I've been looking for a good solution so I can fix this and start printing again, and apparently lots of people have issues with the relay so I'd love to just get rid of it.

Does the solution you have with the MOSFETs completely replace the relay? And if so is there anything I should be aware of as far as pros/cons? And could you perhaps(if not its cool) snap a pic of the setup with the mosfet so I can make sure I don't fry anything if I attempt this?

Thanks alot

This thread is wicked and I'd love to hear how the Koch tealight went with all your mods on the printer.

Edited 1 time(s). Last edit at 06/09/2015 10:20AM by dennmat.

Attached is the circuit diagram of my heat bed relay eliminator using the IRF4905 mosfet. No circuit board was used just attached resistor to leads, soldered on wires and wire tied to the heatsink for support (dead bug assembly). Not pretty but real functional. When the HB-OUT terminal is taken to ground by the FET on the RUMBA board, the mosfet starts conducting.

Two other items: 1) this device is static sensitive even though there is some protection inside; 2) There is an internal back diode forward conducting from drain to source that is not indicated on the schematic. If you reverse the drain and source, your heat bed will come on full and not be controlled as it is being powered by the forward conducting back diode.

Edited 2 time(s). Last edit at 06/09/2015 08:18PM by itchytweed.

That's perfect.

Thanks a bunch this is what I'm going to do.

Here is a picture of the Koch tealight.

It is printed using Slic3r in spiral vase mode, 0.4 mm nozzle, 30mm/sec speed, 195 deg C, Inland Gold PLA 1.75 mm (MicroCenter).

@dennmat - here is a pic of the mosfet on the heatsink - dead bug style.

This is mounted under the RUMBA board so that airflow from the fan that cools the stepper drivers makes it way to the heatsink and help out cooling. I can now run the heatbed in true PID mode and as I type this away, I have two yellow flashing lights being real happy at setpoint (bed / nozzle).

@dennmat - thanks for the compliment. I can guarantee you that I am having all sorts of fun with this project that keeps me off the TV......

Update. I am finding that even though the wood that is used is an MDF laminate, it is still subject to moisture expansion / contraction. I am noticing this in the z-axis homing. It seems that every session, I am having to readjust my z axis home to get a good starting bottom layer. I don't know if it is the wood expanding / shrinking or other factors but it does get to be a tad distracting. I can see where auto bed leveling would be a help here but I am going to come up with some way of using a dial indicator to level the bed to the nozzle and use the z adjust to fix the gap.

When you are pulsing 500 watts on and off, interesting things can happen. The interesting thing that happens in my mancave where the printer is located ... flash. I have dimmable LED lighting installed and as the heat bed turns on and off, they slightly flicker at the same rate as the power supply does it job. So, I rewired the supply to the printer. I have an Anderson PowerPole fused distribution strip. The power supply is on its own fuse, the electronics are on their own fuse, the heat bed is on its own fuse. The addition to this setup is an AGM car battery - 27 series - on its own fuse. For the geeky ones out there, this is equivalent to adding a 24,000F (!!) capacitor to the printer. Needless to say, the flicker is now gone as the battery moderates current draw on the power supply, as it should. No more flickering.

Next thing to do - try my hand at 3D scanning and printing. Going to try Insight3D and my regular camera. May not be automatic but it looks promising. Wonder how it will handle 18MP images...

Wow that print came out awesome, you've got almost no resonance that I can see on the x or y.

And good idea mounting in there. I wasn't sure where to put it, but I might build a little enclosure for it all now, with some mounted fans from some old half together pcs I have laying around.

Bookmarked this thread, will be checking up on it once and a while.

It seems the printer has developed an interesting issue.

The Y-axis rails have taken to pivoting about the two support pegs that cross under the rails. This allows the Y-axis bars to skew out of square with the X-axis. As a temporary fix, I have two adjustable strings running from the front corners of the Y-axis bars to one of the outside Z-axis stepper anchor screws. Not the best solution to the issue but that will require fashioning up two angle brackets or right angle wood triangles to minimize the bending between the crosspiece wood and the rear stabs that run parallel to the Y-axis bars.

I am also thinking of elevating it on some 20x40 v-slot. This will allow me to anchor it to the table using some clamps. This will prevent the printer from walking around the table and put all the energy into axis moving and not a tabletop field trip. The walking is no fault of the printer, just its living conditions.

There was a question somewhere as to how much power the printer pulls from the supply.

I managed to find my AC/DC clamp-on ammeter. On a cold start with both the bed and nozzle heating and the rest idle, it was 34 amps at 14 volts (476 watts). So 40-50 amps sizing on the power supply is reasonable for this printer.

Relay controls the bed? I would imagine a non-relay option would be a first upgrade. Or is the cost a big jump?

Quote
RRuser
Relay controls the bed? I would imagine a non-relay option would be a first upgrade. Or is the cost a big jump?

Look above a few posts. Schematic and picture of what I used to eliminate the heat bed relay and run PID mode on the heat bed. All it took was a 10k resistor, a free heatsink I had in a junk box, and the IRF4905 mosfet ( $3.14 currently at Mouser). Far more efficient and way cheaper than an SSR.

Had a nasty mess to clean up. Apparently I did not have the connector between the heater head and heat sink sufficiently gorilla'd down. Paper towel made cleanup easy....as long as parts were hot.

Had one of those nasty jobs to deal with - cleaning a plugged nozzle. Have it cleaned out and running again. Kudos to a small butane torch, some needle nose pliers and a strand from 12 ga copper wire. It measured out to be 0.4 mm in diameter. Cleaned out the nozzle and will try it out on full tomorrow.

*Update* Back to being a printin' fool. I kept off the silicone cover from the Hexagon. I have found by doing that it is easier to clean the outside of the nozzle before a print job and I am not collecting a ton of cobwebs/strings on it. The only thing that I am going to have to do is Kapton tape the thermistor into its hole.

Edited 1 time(s). Last edit at 06/21/2015 05:12PM by itchytweed.