anon4321

I have a version that looks like it should work. I pushed the fan shroud up about 3-4 mm Haven't printed it yet due to strange issues with my printer and the first few layers separating. I think it's temperature related.

>> Can't you just raise the mount for the fanduct? The problem is that where the fan duct attaches is bounded by curves and other geometry like the loop. This occurs both at where the fan duct attaches to the main body as well as where the "tabs" on the duct meet the swooping curves of the duct itself. Additionally, the interface of the duct assembly to the main body has a clip on one side and the inverted obtuse L like shape on the other. There are no good right angles, parallels or straights in the are to cut out so the whole thing can be moved up. The situation is the same on the duct itself. Here is the best pic I could get from youmagine:

I guess it is a matter of opinion. The issue is that the first run wizard seems to be OK with allowing you to select the self-built bed option and then asking you if you want to update the firmware. This sort of implies that it is sort of supported. However, if you update the FW, the printer is dead. I was playing around with various versions of Cura and I think in 14.01, if you have picked the self built bed and attempt to update the firmware, you are messaged that Cura doesn't come with the firmware for the self installed bed and the firmware installation is prevented. Obviously, we are a little on our own with mods like self-built heated beds so expecting Cura to come with firmware for a self-built bed is not realistic. However, the 14.01 check was nice and would prevent people for falling into this trap.

I've run into an issue. I thought I was done but when I virtually assembled everything, the fan duct is too low. In other words, the nozzle, block and heat break plus the shorter heatsink is too short. There is only a .1mm clearance between the lowest parts of the duct and what I calculate as the tip of the nozzle. The fan duct is only curves and I can't figure out how to rework it so that it doesn't drag on the build surface or print. Once you add in a lot of rounds and off axis geometry, DesignSpark Mechanical is a real pain and often refuses to do some operations.

I'm making progress with converting Nick's V5 to V6. I'm working on redoing the shroud portion now and after that it should be ready in the original form. However, I'm looking to cut it so it goes from 3 to 4 pieces so that the main body is easier to print.

Another simpler host is octoprint It doesn't have as many bells and whistles especially in the area of viewer toolpaths. If you want to make your printer wireless, get a RaspPi and use the octopi image.

I HAVE SUCCEEDED in making a design that is completely unprintable!!!!! Now I understand why the geometry is so complex in Nick's design.... Sigh, back to the drawing board.

Those levelers look very nice.

Yeah, there is that moment where you hold your breath and power it up and hope you don't see smoke! BTW, since your are using an SSR capable of switching on and off quickly, you can set the bed heating algorithm into PID instead of the default bang-bang mode. Then once you get the full bed assembled, you can use the PID autotune to derive the P, I and D values. In this mode, the power to the bed is more closely modulated which you will see as the LED flashing quickly. To do so, you need to recompile the firmware. Get the firmware here https://github.com/Ultimaker/Marlin using the download zip button on the middle-bottom right and unpack. Get the Arduino 1.0.5 IDE here: http://arduino.cc/en/Main/Software and install In the location you unpacked the firmware source, open marlin.pde. In windows .pde is associated to the Arduino IDE. In Tools -> Board, select "Arduino Mega 2560 or Mega ADK" In the tabs, find configuration.h Find this line: #define TEMP_SENSOR_BED 0 change to: #define TEMP_SENSOR_BED 1 Find this line: //#define PIDTEMPBED Changed to: #define PIDTEMPBED With the printer connected and powered on and Cura closed, click File -> Upload and the IDE will build then flash the new firmware. Once your bed is completely assembled, you need to tune the PID parameters for the bed "system". You can play around with this without your Alu surface but you will need to redo this as the values will change with the thermal mass of the Alu. To run the PID autotune, Open Cura Click File -> Preferences Select the Pronterface UI for the printing window type and click OK. Now go to File -> Print. Once the connection is established, use the text box at the bottom right to send this command: M303 E-1 S70 C8 (Note the S value is the target temp, adjust it to be the temp you believe you will use the most). The bed will go through several heating cycles and output three values as the PID is "dialed-in" Record the last set as the output instructs. Now to "burn" those values into the FW, find these lines: #define DEFAULT_bedKp 10.00 #define DEFAULT_bedKi .023 #define DEFAULT_bedKd 305.4 Replace the values you recorded from the autotune run. Make sure Cura is closed and click File -> Upload in the Arduino IDE. Enjoy the blinky goodness of the hyperaccurate and almost certainly overkill of running the bed in PID mode. More info about PID and autotune here: http://reprap.org/wiki/PID_Tuning

You can insulate the diode. It probably isn't really used as it is for blocking inductive kickback (http://www.falstad.com/circuit/e-inductkick.html) and the bed isn't inductive or only parasitically inductive. Note that without the LED, the polarity of the connection to the bed doesn't matter. However, it will when the LED is added, Also note that the PWM output for the fan on the UM print head is actually ~19.5 volts. The UM1 does use a 12V fan which is apparently OK with the higher voltage. However, in my experience, other 12V fans do NOT tolerate the higher voltage and immediately go "POOF".

It's close. The diode should be used to protect the relay on the load side, reverse biased. With an electromechanical relay, you would need something to protect the MOSFET from inductive kickback from the relay's coil. SSRs don't have this problem. Also, I don't have access to the schematics but I believe the UM1 board has that protection diode on the board on all three heater connections. See D1, D2 and D3 in my sample pic - Not that it matters electrically but you are low side switching which means more of the positive side is always connected. I prefer to use high side switching as in the attached pic because it interrupts the + side earlier. This means in the off state, most of the circuit is at ground level. With low side switching, most of the circuit is at the PS voltage. This would lead to a higher chance of a problem if some part of the circuit is grounded by accident even in the off state.

Ugh, this is a lot harder than I thought. NIck's design has a lot of nice but complex geometry....

https://www.sparkfun.com/categories/new_products

Yes, from a product perspective, definitely needs to be fixed ASAP!!! If you mistakenly update your FW, the printer is basically dead. You need to pull the arduino out and reflash with custom built FW. It's the pulling it out part that is a PITA. Others might just give up and generate a call to support which isn't good either...

Jonny, was hoping you had a design I could steal! I think I'll try modding Mr. Foley's design.

Hey Jonny, have you worked on a V6 mount yet? If not, I have a model of the heatsink reversed from the engineering docs and I'm checking the top against a V6 as soon as a sample is done printing. Meant to say I could pass you the model if you needed it for your mount design. I think tomorrow I will print a hotend from the model to see if the dimensions match the real thing.

I don't know if there is much you can do. You can try adding support through Meshmixer which uses a different technique and can analyze the model to reorient it for optimized support.

Wow, awesome prints. Next time you do something that has that kind of extreme, well, bridging, please video it ! I'm very jealous of your UM2! Good stuff!

The steppers are skipping steps and aren't really completing the moves. To eliminate one variable, can you go back to the firmware from 14.01 and see if the printer does better during the leveling wizard? Does the head move easily when the power is off? If not, you might want to try some sew machine oil on the 8mm rods.

You also want stranded and flexible. If you are near an R/C store see if you can get 14 gauge wet noodle or superworm. http://wsdeans.com/products/wire/index.html http://www.acerracing.com/ball_bearings/Superworm_Silicone_Wire_14_Gauge_10_feet-1464-75.html

Those gauges sound a little overkill. I based my selection on the ampacity column here: http://en.wikipedia.org/wiki/American_wire_gauge Assuming you are still using at 12V with the 12V 16A Brick power supply, I think 14 gauge should be plenty.

PS: Did youmagine stop posting new designs here?

I have made some minor updates that prevent the cable from working its way out. Also, I added the source file.

Actually, now that I look at it. I don't think the diode is connected optimally. It really should be across the bed leads (in reverse polarity, the marked side going to the more +). The way I have it protects the relay but not the PS.. It really should have the marked side connected to the lead going to the - on the load side of the relay and the unmarked side connected to your P3/P4 (-). Like this -

To absolve myself, please note that YOU DO THIS AT YOUR OWN RISK AND I CANNOT BE HELD LIABLE FOR ANY DAMAGE, INJURY, FIRE AND/OR DEATH. The big issue with the SSR is that both the input and load sides are polarized. You should make all connections EXCEPT those to the relay and then verify that polarity using a meter and that the relay connections will be positive to positive and negative to negative. The hardest one is the HEAT BED terminal on the controller. The outer most post is + and the post next to the other connectors is -. The diode is required if the load is inductive and protects the SSR load side from inductive kick back voltage that occurs when the load is switched off. The heat bed is mostly resistive and probably only marginally inductive depending on its construction. I used the diode as cheap insurance for the $30 relay. Again, verify ALL POLARITIES before connecting to the SSR but this is how mine is connected: Here is the SSR in IRL. Note that my PS also had two positive and two negative leads/pins due to the current. The two + are tired together at the SSR terminal. The two - and one lead to the bed are tied together by the wire nut. The remaining lead from the bed is attached to the other load terminal. The diode is reversed biased across the SSR so that it doesn't conduct when the polarity is correct ! The input side simply goes straight to the heat bed terminal with, of course, the correct polarity.