large diameter heated-bed design

Nobody else seems to be having problems with bed-heating performance but my recent construction is a complete fail!
Comments are welcome, help is even better! I am particularly interested in other people using wide silicone heaters.

I have a large delta-style build inspired by DC42's build, but the bed design is different.
I thought I had a decent thermal design, but I am getting horrible behavior, like 75 on the glass with the temperature set to 120. WTF? Other people don't seem to have this issue.

I have calibrated my thermistor reading against a seperate thermocouple, thermistor is reporting correct temperatures.
The picture shows my construction and everything is to scale.
The top glass/plate is 350mm OD.
The heater is 300mm OD.
That center inner insulator is about 30mm.

Here are my observations:
1) The original Kossel Mini heater (PCB, 120W) worked fine (in the Kossel Mini), or at least well enough to do ABS, so what has changed with this new design?
2) The only real changes are using a silicone heater and making the glass/metal plates wider than the heater.
The original PCB heater was about 3.4 W/in^2. The new silicone one is about 3.3 so 500 watts is barely enough.

Tests:
1) No glass bed, thermal-couple just pushed hard against the center of the metal plate, set heater to 100 deg, measures 95 deg. Metal plate seems to be heating up ok. I set the PID D-parameter to very small so the heater would go full 500W heat-up. It would seem that with glass installed, the insulation-quality of the glass is too high and heat goes downward through the insulation.

Possible issues:
1) Heater has poor thermal contact with the metal plate (gap, adhesive, silicone itself).
2) Heater section directly below the thermistor gets extra hot without contacting something to draw the heat out (like the metal plate). Thermistor locally overheats.
3) Glass and metal plates are wider than the heater. Unheated outer section of the glass/metal plate sucks out too much heat. (should actually not be a problem by my estimates because the the metal plate is insulated both top and bottom surfaces. Only the edge is exposed.)

Fixes: A lot more insulation underneath to ensure more heat is pushed upward through the glass instead of lost through the insulation.
Move thermocouple to outer unheated ring of the metal plate. Will this cause the heater itself to reach undesired high temperatures and cook my insulation off?

How about fixes: thick, flat aluminum plate and no glass, thermocouple or thermistor placed wherever it is convenient?

My printer has a 12" square 1/4" thick cast aluminum tooling plate bed covered with 5 mil kapton tape and a 450W heater and that's it. No glass on top, no insulation on the underside, no glue, no hairspray, or other slop required. The bed reaches 105C ABS print temperature in about 5 minutes and prints stick. I use PID control to regulate the temperature, and after an initial overshoot of about 5C, it settles to the set temperature and doesn't move. The surface is flat enough to print edge to edge.

Here's a recent print- 10" diameter ABS:



Here's the bed flatness/level, X and Y axis sag all measured:

https://vimeo.com/142670595

It really doesn't have to be so complicated...

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

@shadowphile, it's very difficult to measure the temperature on top of glass using a thermocouple because of the insulating properties of the glass, the tiny contact area between the thermocouple and the glass, and the heat carried away by the thermocouple leads. My advice is to set the bed temperature about 10C higher than you actually want and not worry about it. I print on 4mm thick float glass plates and I use indicated temperatures of 70C for printing PLA and 120C for ABS.

I have thicker insulation than you do (8mm cork), however the main benefit is to avoid heating the electronics under the bed.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I don't have much experience with this setup so take this with a grain of salt

I would not put the thermistor in direct contact with the heater. The reason being that you want the temperature of the aluminum bed reported, not that of the heater (the latter getting really hot really fast, I would guess).
With that in mind, I 'd test with the thermistor on a corner of the aluminum bed (not touching the heater).

It might help using aluminum foil between heater and cardboard to reflect IR rays partially.
-Olaf

I use cheap thermal paste between bed and 2mm heat spreader and thermal paste between spreader and glass. The bed gets into the 100c range pretty good.

---

My Personal Blog. Build blog.
[engineerd3d.ddns.net]

Modicum V1 sold on e-bay user jaguarking11

I agree that a metal plate would solve all heat transfer issues, just haven't been ready to discard my investment in glass and metal and kapton are easy to scratch. Kapton is expensive, especially in large sheets.
However, too many people seem to get a glass setup to work for me to abandon yet. That is why I am confused..what are other (simpler) glass bed designs working ok? There is a major flaw in my design, the over-design is mostly a (failed) attempt to get the performance I need. All I can think is that the heater is not making a good contact with the metal spreader plate. With no contact, the metal plate acts like a reflective IR barrier! A black coating of paint might fix that.

Quote
mandrav
I don't have much experience with this setup so take this with a grain of salt

I would not put the thermistor in direct contact with the heater. The reason being that you want the temperature of the aluminum bed reported, not that of the heater (the latter getting really hot really fast, I would guess).
With that in mind, I 'd test with the thermistor on a corner of the aluminum bed (not touching the heater).

That is why there is a small disk of cork (sandwiching a foil layer) under the thermistor. Didn't help much. I tried moving the sensor to a spot just outside of the heater (plate extends out by about 1" bigger radius) and got a huge temp in the middle. I just don't have good heat transfer.

Quote
o_lampe
It might help using aluminum foil between heater and cardboard to reflect IR rays partially.
-Olaf

There is a layer embedded in the cork, outside of the cardboard.

Sorry about al the previous posts, should have consolidated all my answers in one post.

Quote
jaguarking11
I use cheap thermal paste between bed and 2mm heat spreader and thermal paste between spreader and glass. The bed gets into the 100c range pretty good.

I forgot that one other difference between this and my original Mini Kossel PCB heat plate was a layer of thermal paste across the whole thing.
This design used the built-in adhesive on the silicone heater; now it is rather torn up. Without an adhesive, the physical support is missing without something underneath the heater to press it onto the plate. My elegant design that didn't need any physical support under the bed is now forfeit. (I will be looking for a thermal-adhesive to fix the problem)

Quote
dc42
@shadowphile, it's very difficult to measure the temperature on top of glass using a thermocouple because of the insulating properties of the glass, the tiny contact area between the thermocouple and the glass, and the heat carried away by the thermocouple leads. My advice is to set the bed temperature about 10C higher than you actually want and not worry about it. I print on 4mm thick float glass plates and I use indicated temperatures of 70C for printing PLA and 120C for ABS.

I have thicker insulation than you do (8mm cork), however the main benefit is to avoid heating the electronics under the bed.

How is your silicone heater contacting/attached to the upper side (metal plate)? I found out this problem because I tried to print ABS for the first time and jacking the command to well above 100 C didn't seem to be enough. My 10mm cube was all warped!

Edited 1 time(s). Last edit at 12/10/2015 03:29PM by shadowphile.

Quote
shadowphile
How is your silicone heater contacting/attached to the upper side (metal plate)? I found out this problem because I tried to print ABS for the first time and jacking the command to well above 100 C didn't seem to be enough. My 10mm cube was all warped!

I stuck the silicone heater to the underside of the 6mm aluminium plate using the adhesive sheet that came already attached to it.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42

Quote
shadowphile
How is your silicone heater contacting/attached to the upper side (metal plate)? I found out this problem because I tried to print ABS for the first time and jacking the command to well above 100 C didn't seem to be enough. My 10mm cube was all warped!

I stuck the silicone heater to the underside of the 6mm aluminium plate using the adhesive sheet that came already attached to it.

That is exactly my original build. I thought a 500W heater would be overkill, although all it should improve is heat-up speed. It was a good deal though.
This is weird.
Somebody mentioned somewhere that the bulk of the heat from the bed was via radiation from the metal plate, not conducted through the glass. They would certainly not be measurable with a contact probe.

I have not tried my Kossel 250mm heater yet, but on my TAZ 5 (300x300mm square glass) the heater was glued to the glass with 3M 90, and that sat on top of an aluminum plate. I think you have the heater glued to the aluminum plates bottom side. Flip that plate over and see how it goes You will lose temperature where the glass is not over the heater, of course.


David

Edited 1 time(s). Last edit at 12/10/2015 10:25PM by voi9viper.

Quote
shadowphile
I agree that a metal plate would solve all heat transfer issues, just haven't been ready to discard my investment in glass and metal and kapton are easy to scratch. Kapton is expensive, especially in large sheets.

I've seen the "easy to scratch" comment before. If you manage to sctratch the metal, you must have gouged through the kapton tape, so you're going to replace the tape. The scratch in the metal doesn't matter if it's covered with kapton tape. I have a couple scratches in my bed from head crashes when setting up the machine. They are of no consequence at all.

If you applied enough force to gouge the aluminum, what would have happened if it were glass? Maybe it would break into tiny shards? I've never heard of aluminum shattering if dropped on the floor...

You don't have to buy kapton tape in large sheets. I use 2" wide tape to cover my 12.5" x 12" bed plate (you can see in the previously posted photo). Since most printing occurs at the center of the bed, that area is most likely to get damaged. When it does, I replace the damaged strip(s) only. I do not use the cheap, 1 mil tape that most people buy- it's too fragile. I use 5 mil tape that is very tough and lasts a long time. I typically replace the central two strips of tape on my printer after a few months of almost daily printing. I bought a roll of the stuff almost 3 years ago and I'm still using it. Sure, the edges of the strips leave tiny lines spaced 2" apart in the bottom of prints. It's OK. I haven't had a single instance where the tiny lines on the bottom of the print mattered.

The bottom of your prints will always look different from the top and sides. It's OK. That's the nature of FDM printing. If you need the bottom to look more like the top, print on a raft, or do some acetone vapor smoothing.

Edited 2 time(s). Last edit at 12/11/2015 03:54AM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Have you considered a thicker sheet of insulation? I find that a sheet of ~1/2"-thick styrofoam blocks almost all of the heat escaping through the bottom. Just be sure to have some insulation between, since Styrofoam will melt around 100C.

Lemme see if I got this right... You're proposing to use a piece of insulation to insulate the insulation that might melt.
Why not just use the insulation you're going to use to insulate the insulation and do without the insulation that's going to melt?

This is just too much craziness!

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

You can skip styrofoam and go fiberglass. That should handle the 100c range of the bed. Just get ready for a respirator and allot of itching.

---

My Personal Blog. Build blog.
[engineerd3d.ddns.net]

Modicum V1 sold on e-bay user jaguarking11

If you have a light bulb in your basement/garage work room that doesn't produce enough light, do you 1) put a bunch of mirrors up to try to make better use of the light it produces, 2) try to capture some of the heat the bulb produces and convert it into electricity that you can use to light a supplemental lamp, 3) Install solar panels to capture some of the light, convert it to electricity and use that to light a supplemental lamp, 4) try to modify the bulb to make it produce more light, 5) replace the bulb with one that throws adequate light? While all of these are potential solutions to the problem, the simplest one that guarantees the end result is number 5.

In the end, this and most other print bed heating problems come down to one thing only: inadequate power going into the heater. You can't insulate your way out of this problem. You need more heat. Just get a higher powered heater and/or beefier power supply and be done with it. The result is guaranteed and the time and effort required are minimal.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]