Close



Page 2 of 9 FirstFirst 1234 ... LastLast
Results 11 to 20 of 83
  1. #11
    Engineer-in-Training
    Join Date
    Feb 2015
    Posts
    371
    I only now just started reading on the PID for the bed. I have done the autotune on the hotend. It was my assumption that you could not use the PID loop when using the heat bed relay. Is this the case? A quick google of Heat Bed PID made it sound possible. My assumption is that they are not using a relay however.

    Also, I have read that document many times and I get something new from it every time. I think I have convinced myself even more to look into a 6mm lead screw with a single start. Can't find them in a 12" maker farm length however. 320mm is about the longest I have seen on amazon with a backlash nut.

  2. #12
    Technologist
    Join Date
    Apr 2015
    Location
    Lakeport, CA.
    Posts
    174
    Thx..., an excellent article! I find myself wishing for that sort of diagnostic help a lot.

    So if I understand it, per the article, the PID is an algorithmic function of the firmware..., and mentions it as a tuneable (in Marlin anyway). This is what I was trying to get at when I was speaking of tweaking heat cycling/hysteresis. In which case, the youtube vid that Chadd shared, would suggest that that the table expansion/contraction effects (if they're present in a particular machine) could be significantly improved with firmware tuning alone..., no hardware changes.

    The article also mentions testing with PLA and blue tape as a means of isolating the problem, i.e., determining if a problem exists sans heating effects.

    Quote Originally Posted by super fly View Post
    PID=Proportional Intregal Derivative
    It's a feedback controller based on current change, future change, and past change. Each term is weighted to come up with an output (this is where PID tuning comes into play). I use one to controller my smoker temperature.

    Also, here is a pretty lengthy write up on z banding. To be honest I haven't completely read through yet as I don't have my printer yet. I saved the link just in case I needed it later.
    https://www.evernote.com/shard/s211/...c268dcdd2d687e

  3. #13
    Technologist
    Join Date
    Apr 2015
    Location
    Lakeport, CA.
    Posts
    174
    Hmm. If the PID still drives the heat bed in an on/off manner (but with what is in effect a tuneable duty-cycle), then I could imagine that a mechanical relay may not like that for very long. In which case, switching to an SSR might be the answer. OTOH, if the PID drives the heat bed in an analog fashion..., yah..., that's a different thing. I guess my question is: is the default MakerFarm firmware using the "PID" algorithm, or something else?

    As for the lead screw, yah, that feels like the right solution (or at least a big part of it). In my case, the machine is good enough as-is for what I'm doing..., though heat bed heat uniformity is becoming a bit of a problem...

    Quote Originally Posted by tsteever View Post
    I only now just started reading on the PID for the bed. I have done the autotune on the hotend. It was my assumption that you could not use the PID loop when using the heat bed relay. Is this the case? A quick google of Heat Bed PID made it sound possible. My assumption is that they are not using a relay however.

    Also, I have read that document many times and I get something new from it every time. I think I have convinced myself even more to look into a 6mm lead screw with a single start. Can't find them in a 12" maker farm length however. 320mm is about the longest I have seen on amazon with a backlash nut.

  4. #14
    Yes to use PID you will need a solid state relay. The way the Marlin firmware works is using PWM I believe (like I said I don't even have my printer yet lol); meaning it would vary the frequency of the pulses to the relay in order to maintain the temperature of the bed. The mechanical really just can't turn on and off quickly enough for PWM to work. You would also probably need a heatsink for the relay more than likely if you are running it on 12v. In 24v maybe or 110v especially the problem goes away because there isn't as much current draw; P=VI. Of course you would then need a heat bed that would work on 24v or 110VAC.
    Last edited by super fly; 08-28-2015 at 11:19 AM.

  5. #15
    Engineer-in-Training
    Join Date
    Feb 2015
    Posts
    211
    Quote Originally Posted by lakester View Post
    Cool demo of the problem (thx!).

    2 questions:

    1. What does "PID" refer to? (Sorry..., been puzzling on that...)
    2. What is the table structure of the Mendel90 in the vids? What are its similarities to the MakerFarm wood+glass table?

    1. To put it simple PID uses PWM to control your bed temp so it turns the bed on and off very quickly to maintain the temp vs turning the bed on 100% for 5 seconds at time to maintain the temp. The reason you have to have a Solid state relay is because a mechanical relay can't handle the quick on and off switching. In that video link what you are seeing is the bed distorting when it turns on @100% power and then returning to its normal position after the power is off to the bed. The second video shows the same thing with PID, the bed switches on and off quickly to maintain the temp so it doesn't stay on/off long enough to show the change in distortion.

    http://techteach.no/kybsim/pwm_control/index.php

    2. The bed designs are pretty well identical.

  6. #16
    Engineer-in-Training
    Join Date
    Feb 2015
    Posts
    211
    Quote Originally Posted by lakester View Post
    Thx..., an excellent article! I find myself wishing for that sort of diagnostic help a lot.

    So if I understand it, per the article, the PID is an algorithmic function of the firmware..., and mentions it as a tuneable (in Marlin anyway). This is what I was trying to get at when I was speaking of tweaking heat cycling/hysteresis. In which case, the youtube vid that Chadd shared, would suggest that that the table expansion/contraction effects (if they're present in a particular machine) could be significantly improved with firmware tuning alone..., no hardware changes.

    The article also mentions testing with PLA and blue tape as a means of isolating the problem, i.e., determining if a problem exists sans heating effects.

    As mentioned you have to switch from the mechanical relay to a solid state relay to use PID, but you wouldn't have to change anything other than that.

  7. #17
    Engineer-in-Training
    Join Date
    Feb 2015
    Posts
    371
    Where does one get a solid state relay for use on the maker farm 12" heated beds? I'd switch if I could use a PID loop instead of the bang/bang programming we are now using.

  8. #18
    Staff Engineer printbus's Avatar
    Join Date
    May 2014
    Location
    Highlands Ranch, Colorado USA
    Posts
    1,437
    Add printbus on Thingiverse
    PWM (Pulse Width Modulation) will vary the ratio of on time to off time in a pulse train, not the frequency of the pulses.

    Regarding Solid State Relays, note that for use with the existing 12V DC heater, you'd have to be sure to buy one meant for controlling DC outputs, not AC. A DC SSR will likely use a MOSFET. An AC SSR will likely use a triac. SSRs are really nothing but the MOSFET or triac with driver circuitry in a standalone package. As an example of one, here's the one often recommended for use with the PWM heater output on Smoothieboard - http://shop.uberclock.com/products/solid-state-relay-dc-to-dc.
    Last edited by printbus; 08-28-2015 at 12:56 PM.

  9. #19
    Technologist
    Join Date
    Apr 2015
    Location
    Lakeport, CA.
    Posts
    174
    Do the common arduino boards (aka RAMPS, RUMBA, etc) provide circuitry that facilitate the generation of a PWM signal?

    Probably breaks down to two questions:

    1. Is there a performance hit on the cpu to generate a PWM signal for the heat pad (apart from some minor math computing the heating values)?
    2. In addition to the SSR, do you have to relocate the HB "signal" wiring to a different spot on the controller board that does support PWM, if the current location does not?

    Have been googling a bit, and I can't tell for sure if there is any "helper" hardware on the common boards..., or whether the PWM is entirely software implemented...

  10. #20
    Technologist
    Join Date
    Apr 2015
    Location
    Lakeport, CA.
    Posts
    174
    Hmm..., think I have the answer.

    Looks like it is a soft pwm...., a quick scan of the code doesn't reveal anything driving specialized pwm hardware. Freq is around 7hz..., I guess that, masked in with the other soft pwms (e.g., for the HE), that amounts to writing to a couple of ports a few times a second, without especially critical timing requirements. So, I'm guess'n it doesn't hurt performance much.

    Cool.

    I still think a PLA on blue tape test is worth doing, to see if it's worth going to any trouble in this regard at all.

Page 2 of 9 FirstFirst 1234 ... LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •