Jump to content

anon4321

Dormant
  • Posts

    914
  • Joined

  • Last visited

  • Days Won

    1

Posts posted by anon4321

  1. PID is implemented by varying the pulse width time (PWM). So PID and PWM are related. Bangbang is a slow on and off cycle and is determined by the temperature falling below the desired temp by a certain amount causing the relay to switch on until it is above the set point by a certain amount. That on and off might take several seconds. PWM switches on and off many times a second trying to supply just enough power so that the temperature is kept very close to the desired temp.

     

  2. I think the drivers were originally done by pololu and then copied many many times. They have distributors in Russia:

    http://www.pololu.com/distributors#russia

    These are compatible with the ones that come with the printer:

    http://www.pololu.com/product/1182/

    You can upgrade to the black version which handle more current or the DRV8825 which handles even more current.

    http://www.pololu.com/product/2128

    http://www.pololu.com/product/2133

    The upgraded versions are both pin compatible with the first one and those in the printer. However, the DRV8825 version defaults to 32 microsteps whereas the original and black version default to 16 microsteps. For the DRV8825, you need to alter the jumpers on the shield so that it is in 16 microstep mode or update the firmware steps/mm.

    However, for any of the three pololu supplied drivers, you need to solder the pins/headers on and they do NOT come with a heatsink. You can use pliers to remove the heatsink off the driver you are replacing and then scrap and clean and then use a thermal adhesive like http://www.arcticsilver.com/ta.htm to attach to the new driver.

    With all that aside, the question is why is the driver shutting down? Did you check that the stepper is free and easily moved after removing the filament? can it feed the filament through the Bowden tube until it gets to the extruder? You should be able to easily turn the large gear by hand (slowly) without the filament and then fairly easily until the filament gets to the extruder.

    Note that you want to turn the current down if the driver is shutting down due to heat. Refer to this to know which direction to turn. It shows the direction for increasing so you want to do the opposite like a 1/16th of a turn and then test. stepperdrivers

     

  3. As a poor speller myself (native English speaker so I have no excuse), it's unfortunate that misspellings in a presentation give the impression of lack of knowledge, intelligence and/or education when that impression can be quite incorrect. In my case, the US education system wasn't the best when I was in school. It's better now but no where near other countries.

    Anyhow, that dragnet feeder is pretty slick. I was thinking of doing something similar where both the idler bearing and knurled bolt are driven but I couldn't workout the geometry that would allow it to be powered by one source while maintaining the ability to adjust it. Seems like you have...

    One thing I did I really liked was to replace my UM1's feeder gear set with a herringbone set. It allows very little backlash. If you have backlash, you might considered a similar solution.

     

  4. Someone wanted to do a layer-by-layer timelapse using a similar command. Someone else suggested the m42 command in the layer comment.

    You can probably leverage the same solution for your use. However, it requires you to tap into the board expansion points and I think you will need to solder a header on to do so.

    https://groups.google.com/forum/#!msg/ultimaker/vZok0JVD430/jVH5kOwj620J

    Daid

    4/12/12

    Actually, you don't need to touch Arduino code. You can change pins with the M42 code. Combine this with G4 (delay) and you can code your camera trigger in pure GCode.

    For example:

    M42 P23 S255 ; Set the trigger camera pin high

    G4 P100 ; Wait 100ms

    M42 P23 S0 ; Set the trigger camera pin low

    Gijs

    Nice!

    do you also know how that corresponds to the 1.5.4 ultimaker pcb? because from the wiki it's not clear to me where I would connect the opto-coupler,

    cheers,

    Daid

    The pin numbers correspond to the numbers shown here:

    http://www.reprap.org/wiki/Ultimaker%27s_v1.5.4_PCB#Pinout_of_Extension_connectors

    I recommend you use one of the PWM or ANALOGUE pins, because the EXP1 and EXP2 are used by the Ultipanel later on, if you get one.

    (The analogue pins can also be used as digital outputs, something you might not guess from the name)

    Gijs

    OK, suppose I would use PWM 8 and 9, what would then be the corresponding M42 pin assignment? just P8 and P9?

    Daid

    Yes, P8 and P9 would do the trick then. Small note about the PWM outputs, those can be PWM driven, not sure if the Arduino will do that or not, just be sure to use S0 and S255, no values in between to enable/disable the output.

     

  5. No problem Daid. I wasn't complaining. Betas have bugs. I was just giving you more information in case you were still trying to determine why it wasn't working.

    With the latest firmware built from the latest source, the beta Cura and the changes mentioned by Dim3nsioneer, the volumetric stuff seems to work.

    However, I've yet to print a part with a lot of retracts so we will see how that goes.

     

  6. Odd, as there are zero changes in that part.

     

    Daid,

    I don't know if this helps but....

    The "Install default firmware,,," menu item just opens a progress dialog that just sits are "Reading firmware...."

    However, "Install custom firmware" works fine.

    I chose Ultimaker with custom heated bed and I wonder if that trips up the "default" firmware upload. Is there even a valid "default" firmware for the UM1 + custom bed solution?

     

  7. I tried the raft but couldn't separate the print from the raft. Any suggestions? UM1 at the defaults but with a heated bed. Temps were 220/70. Trying a second time @ 220/50.

    Second attempt had the same problem. The raft and the part are basically one can can't be separated.

    This is a UM1 using ProtoParadigm's red PLA.

     

  8. The problem with linear regulation is that when not at 100% or 0%, the regulating part must dissipate a lot of power which is usually in the form of heat. For example, I have a bed that is 10 amps @ 24v which is 240W (power = voltage * current or P=VI). If that where linearly regulated at 50%, the bed would get 120W and the regulating part would have to do something with the other 120W.

    Have you ever unscrewed a 120W blub that was on? Very hot...

    So often the regulation is done in full on or full off model switching quickly enough that something (in this case the bed) is averaging the on and off times and smoothing the change. So if a MOSFET turns the bed on for 1/10th of a second and then turns it of for a 1/10th of a second, the on/off time is happening 5 times a second. Because the temperature of the bed can't change immediately when heating or cooling, the average power consumed is the ratio of on time to off time or in this example 50%. From the MOSFETs perspective, it is either on so the voltage across it is 0 or off so the current through it is 0. If you look at the P=VI formula, if V is zero or I is zero, the power consumed or dissipated is 0.

    Well, the caveat is that the MOSFET is perfect and in the real world, nothing is perfect. The MOSFET takes a small amount of time to transition from off to on and on to off. During this time, it is sort of in a linear mode and so dissipating power as heat. However, again like the bed, the MOSFETs temperature can't change instantaneously so it's temperature is based on the time it spends between full on or off to the time it is full on or off. This ratio is very small.

    The SSR I'm using has a turn on of 1ms and a turn off time of .5 ms. So as long as the switching frequency is low so that the SSR spends more time on or off than in the middle, it won't need to deal with much heat.

    Then there is also the fact that real SSRs, unlike the theoretical model of a switch, have a small amount of resistance. P as equals I squared times R. But MOSFETs tend to have a very low value of R keeping the power they need to dissipate low too.

    Welcome to electronics 101 ;-)

     

  9. PID seems to be more accurate or at least more granular so that the temperature is more closely controlled to the set point and doesn't overshoot after settling in. I believe it also regulates better when the conditions change such as when the fan comes on.

    For the bed, I believe that earlier in the thread there was a discussion that regulating the temperature extremely close to the set point isn't required and that bangbang is more than sufficient.

    I went with the SSR just because I like solid state and dislike clicking....

    With both that SSR and SSRs and MOSFETs in general, you need to be aware of the turn on/off time. That SSR has a 1ms second turn on (or was it off? well whatever it was the opposite was quicker at .5ms) so the PID on/off sequence can't be faster than 100hz. Just looking at the LED, the PID looks like it is switching at 5-20hz or so so no problem there.

    In either case, the current is only on or off. Neither mechanical nor SS regulates the temperature linearly such as with say 65% current. It's always PWM.

     

×
×
  • Create New...