Story of the Orbitron Heated bed design
This story is about the challenges I faced and things I learned during the design of the heated bed for the Orbitron Core XY 3D printer. As usual nothing is straight forward even in the case of reuse proved design approach.
Edit: My intention with this article is not to praise the solution with the cheap polyimide heater I found over ebay. Quality wise is way worse compared to the keenovo heater, but outperforms it performance wise.
From my personal view a good design would be a custom made polyimide heater with properly placed temperature sensor and integrated thermal fuse covered over the bottom side with silicon based insulator.
When I started designing my
small printer, the Orbitron, I liked the idea of three point leveling as done for
the Voron v0. I took over that design concept but with a different mounting
method. With all aluminum parts and belted Z. About this maybe in a later
story, now I would like to focus on the heated bed design and its performance.
First I planned to
design my own heater solution as I have designed it for my other printer. A 24V
/ 350W heater polyimide flexible foil design. Designed for 250x250mm heated
bed. You can download the design files here:
I asked for a price estimate for a 120x120 polyimide foil heater. Best sample price I got was over 100$ plus shipping and import charges, kind of expensive
for a 120x120mm heated bed, so I decided to take the shortcut and take a ready-made
solution.
Voron V0 heater has the same
shape and size as my design, LDO is offering Voron V0 kits and Jason from LDO
was more than kind to support this project and sent me a sample of Keenovo
heater (24V / 60W) and a set of magnetic PEI sheet. I was really happy with the
idea, Keenovo has a long history of making very good quality heaters.
All good, I machined the aluminum base plate for the heated bed out of a 6mm thick casted and precision milled aluminum
plate, installed the heater and after I completed the Orbitron assembly and all
the calibrations and PID tunings I started testing.
I was used to the
performance of the heaters designed by me, I had no previous experience using Keenovo
heaters, but immediately noticed that something is wrong with my heated bed performance.
Sensor was indicating like 60 degrees and I could keep my hand on top of the
plate without issues it was just warm…for sure you cannot do that with my
250mm heated bed. So started to investigate.
First test I’ve done was to measure the set vs actual temperature. Using my thermal camera immediately noticed that surface temperature is completely off by over 10°C. To double check and confirm this I placed a thermal sensor over the top of the PEI sheet attached with a drop of thermal paste and Kapton tape as you can see in the picture. Connected the second sensor to the Chamber temperature sensor input. You can see in the chart the PEI sheet surface temperature is completely off by about 10°C and way behind in time, it took about eight minutes to reach 70°C (red line - PEI sheet surface temperature, blue line the temperature indicated by the sensor in the Keenovo heater).
First, I was not sure what causing this issue. Could be from the heater or the magnetic sheet isolates to
much. Temperature difference between the PEI sheet surface and aluminum plate
was insignificant so the obvious conclusion is the heater or the temperature sensor inside the heater. It looks like the sensor in the Keenovo heater reads the bottom side of the heater is not placed close to the aluminum plate as it should be this causing the temperature reading to be completely off by a large amount.
Next, I drilled a hole 3mm hole in the side of the heated be aluminum plate and inserted a temperature sensor (NTC 100K ohm B3950 Thermistor Cartridge). Redone the wiring to use this sensor as reference for the temperature controller. Redone PID calibration and retested the heating behavior (Blue and Magenta).
It can be seen that with the
sensor embedded in the aluminum plate there is an insignificant difference between what this new sensor reads and the actual temperature measured on the PEI sheet surface. Thermal camera measurement confirmed the
results are correct.
Still the performance is far form being good. It heats up very slowly and not able to heat up over 85°C. Klipper is not even able to set temperature up to 90°C and keep the temperature there, gives an error and shuts down the printer.
Made some quick calculations and reached a conclusion that the actual heating power reaching the aluminum plate is just about 40-45W the rest is wasted over the bottom side. As you can see in the thermal picture the bottom side of the heater reaches over 100°C (105°C max in the thermal video) when the heated bed is set to 80°C. This indicates a bad thermal contact between the heating element and the aluminum plate which facilitates significant heat loss over the bottom side of the heater.
Just as a reference this is an example of an insulated etched foil silicon rubber heating element with a very good thermal contact towards the heated element. This
design example is for low voltage heaters.
As a consequence, the Keenovo
heater solution, it’s a dead end for me. Hope they will redesign it in the
future to have a much better performance. Time to go back to the drawing board
and pay the 100$+.
Searching for alternative
solutions I found a promising solution with a 100x100mm polyimide flexible foil
heater with 24V / 100W (12V / 25W) power rating. Almost what I wanted to
design, not a perfect match but for 11$ worth a
try. One side is laminated with a high temperature resistant adhesive tape 3M
486MP.
Removed the Keenovo heater
and installed this polyimide foil heater, after curing and PID tunning went
back to testing.
I think the result speaks
for itself! Its still far from being perfect but there is a tremendous improvement.
The temperature is right on spot, we can see the sides have 1-5 degrees
deviation which is not the best but its due to the smaller dimensions of the heater foil. With this polyimide heater foil the heated bed can reach easily 110°C.
Reaches 80°C in about 4 min (with Keenovo heater was 8 min to reach 70°C).
Reaches 110
For safety reasons I have
added a thermal fuse with 130°C activation threshold.
Heating performance with 80°C
setpoint
Heating performance with 110°C
setpoint
The thermal videos proves the improved performance, but still far from being perfect. This
is mainly because the heater element is smaller compared to the aluminum plate. About 13mm
of the aluminum plate sides are not heater causing it to be cooler compared to the center area.
Seems the best option
would still be to design a custom designed polyimide foil heater and spend the
100$+ for the samples to get the best result. I might do that in the near future.
Tips and trick for installing a polyimide heater foil
Proper installing requires time, do it only when you have 5-6 hours of spare time. It
does not need your full attention but you have to watch over the curing
process. If you just attach the heater, it will cure itself in about 24h
at room temperature, we do not want that because any air pockets trapped inside cannot be correctly eliminated anymore!
First clean the aluminum
plate very well with some IPA. Try not to touch the surface after it with bare
hands. You have to install it carefully not to leave air bubbles between the
foil and the aluminum plate. Don’t worry if it happens, (it happened to me as
well), with a needle or a cutter with a small tip punch a hole in the polyimide foil between the copper tracks and squeeze out all the air bubbles. Very important not
to leave any air pockets! By heat they will expand and delaminate your heater
on a larger area.
After this step you need to
cure the bonding between the foil and the aluminum plate. Do not try to heat it up and print immediately!
Best is to have accelerated
curing by applying temperature rather than leaving it at room temperature for
24h. As you apply heat any remaining air bubbles will expand and will be easier to
spot them and squeeze them out before the adhesive has cured completely.
The curing process I applied:
· Set temperature to ~40°C degrees, check for
air pockets if you see some get rid of them with the method described above.
Leave it at 40°C for about two hours.
·
Increase temperature to 60°C, check for air
bubbles – normally you shall not have any. Let it for about 4h.
·
Increase temperature to 80 degrees, let it for about 2 hours.
·
You’re done, let It cool down and perform PID tunning.
Note: Do not worry if you still have some tiny air bubbles left after curing (smaller than1mm), will not cause any trouble.
Power demand
I have seen some suggest to
limit their heater power to 60-70%. To be honest I do not understand
this idea, if you do not want to use full power buy a lower powered heater, there are better methods to prevent bed warping or fire hazard.
The printer board drives the heater with a PWM (Pulse Width Modulation) signal. This means when
the PWM is ON full power is applied and when its OFF than zero power is
applied. As example if you set 70% max power limitation, the heater will be on
full power for 70% of time and off (zero power) for 30% of time, therefore the
average power delivered to the heater will be 70% in case of my polyimide foil
would be 70W.
This fact does not allow you
to use a lower powered power supply!
As example using this 100W heater the 150W power supply of the Voron v0 default option will not be enough even if you limit the heater power. The PWM cycle time is pretty low, in a few HZ range, the power supplies must be able to provide full peak power of the printer not average. Therefore, this heater need minim of 200W power supply, on my Orbitron I have a 350W MeanWhell power supply.
LRS-350-24
24V/14.6A MEAN WELL
The hater element used in most heaters used in 3D printers is made out of
copper which has a positive temperature coefficient of its resistance.
Meaning as it heats up its
resistance increases, which reduces the heating power. At higher temperatures
more heat is lost via convection and radiation so actually more power is needed to keep temperature. That why the heating graph has a logarithmic shape and it tops at a certain
temperature, where the heat losses are equaled with the actual heating power.
To get a better feeling look
at the next chart I have made during my 250x250mm heater design. You can see
there is about 100W (~30%) drop of the heating power between 20-100°C.
Warping due to fast temperature change
To have a good heater
performance we need about 0.7-1W/cm2 of heating power. Aluminum is a
very good thermal conducting element. To have a warping effect we need to have
a big temperature difference between different areas of the aluminum plate.
Like much higher temperature in the middle compared to the sides. A good heated
bed design should avoid such issues.
Examples which can lead
to warping, remember it’s all about physics, warping will happen if there is a
stress inside the aluminum plate regardless of it cause.
Overpowered heater or heater element much smaller than the aluminum plate, leaving large area over the edge which is
not heated. Remember there is also a distance between the heater element and
the silicon heater edge. Look at this really bad heated bed design example.
This picture shows a 400x400mm heated bed with large unheated area at the edge area.
A good result we have if the
distance between the heater element (not the silicone heater edge but the
heater element inside) and the edge of the aluminum plate is not more than the
thickness of the aluminum plate. In this regards the cheap heater I used is also not
the best design for this purpose.
Thickness of the aluminum plate matters as well. Thicker plate has lower thermal resistance making it less likely to have temperature differences inside but its thermal mass will be increased which means it takes longer to heat and cool down. A good compromise for heated beds up to 250-300mm is 6-8mm casted and precision milled aluminum plate (core XY printer types). I personally prefer 6mm thickness made out of casted precision milled 5000 series aluminum alloy plate for shorter heating time.
The aluminum plate needs to
be manufactured out of piece of casted and precision milled aluminum plate
(like mic6), not laminated. I have seen many heated beds on sale using cheaper precision
milled but laminated aluminum plates. The end user cannot even identify how it was manufactured and will experience weird warping issues.
Large air bubbles between
heater and aluminum plate. Air is a good thermal insulator leading to hot sport
on the heater and cold spots on the aluminum plate.
Last conclusion, using
casted and precision milled aluminum plate (6-8mm thickness), heated with no more than 1W/cm2 power, and have very small
gap between the heater element and the aluminum plate edge will give you a very
good long-lasting performance.