Advices about improvements

I set the bed to 90C and waited 5 min after it hit 90C so that the surface was close to 90C. I then waited for it to cool off. Picture attached. This had the magnets on but not the flexplate with the glass. I will measure the bed dimensions and try to figure out the thermal mass and temp lag bottom to top. I want to get a first order feel as to what the relationship is timewise from bottom to top. At the bottom of the bed we hit 50C in a couple of minutes and 70C in 5 minutes. It would be interesting to be able to generate a profile whereby you set the bed to 70C for some known period of time to achieve a 50C surface.

How much time per print have you found that you can gain with the higher wattage bed heater?

 

The sensor on the bottom of the bed is stable in 3 minutes. Have you verified the surface temp? Tomorrow I will set up a temp gauge on the surface of my bed. I will set the bed temp to 60C. I will then measure the time lag from bed regulation to surface temp tracking it. If it is significant I will in the future make a temp profile for faster lock at the surface.

 

I did some measurements this morning. I discovered that in my system with the flexplate and the microporous glass I get no regulation at the surface. At the bottom of the bed I hit 50C in a couple of minutes using a 70C set with a huge lag till the top surface gets up to 50C.

I glued a thermocouple to the bed. The graphs below show how useless setting the temps are. After I finished the graph, I raised the the meter thermocouple on the bed and set the fans to 100%. The surface temp dropped to 38C.

This has been a real eye opener and I will see if I can move a thermocouple closer to the surface.

What method did you use to verify the top surface was regulating well?

 

The smaller heater is not the problem. With the fans on 100% it regulated the 70C at the bottom of the bed without a problem. It was the surface of the bed were the problem was.

For the R2 bed and a flexplate system the 160w heater is okay. The problem lies in where the temperature is measured. I would love to drill a hole in the middle of the bed right up to the top of the flexplate magnets and insert a thermister. I am afraid of drilling through a heater trace. I wish I had documentation on it.

 

The surface mount thermistor is on the opposite side of a thick PCB from the heating traces. I removed the surface mount thermistor from the bed and drilled a .067 hole in it's place through the bed and the flexlplate system. I then inserted a 3950 thermistor from a hotend through the board, aluminum plate and up to the surface of the flexplate. This was to place the thermistor closer to the surface. This topped me out at 57C when I set the temp to 70C. I had a fairly constant 20C temp lag during heating. Once maxed out at 57C I set the temp of the bed to 50C, the surface of my glass plate stabilized at 42C. This should be constant regardless of the heater used.

I do not see how the glass plate can regulate the temp any better as it is an insulator. Has anyone looked into a system to set the temp regulation at the glass surface.

My bed will not PID auto tune now, so I will have to play with the firmware constants.

 

Regulation to under the glass is not a problem and is rock solid once I moved the thermistor to the underside of the glass. The temperature gradient across the glass from the heater on the underside of the glass to free air on the top of the glass is the problem. In the case of the 4mm glass the temperature differential is 8C from the underside of the glass to the surface at 50C. This value will change with ambient air temperature in addition to the significant temperature lag time. It makes regulation impossible regardless of how regulated the bed is.

 

My auto tune of the PID was fine I had changed it to 60C and had forgotten. The proportional in the PID doubled which is the equivalent of doubling the power to the limit of bed. So now I have to remeasure the lag time. I set the bed to 50C and then put the thermocouple between the flexplate and the magnets it stabilized at 47C the same tool stabilized at the glass surface at 42C. This does not seem too bad now with only a 5C difference in the 4mm glass. The big improvement was drilling the hole and putting a thermistor through the epoxy board heater, the aluminum plate, the flexplate magnets right up to the steel flexplate.
If I did not have the glass and was using the thin flexplate surface I would guess that surface would be well regulated.

 

I set up two R2's with the flexplate system and a microporous glass bed. One R2 with the standard bed and the other with the a hole drilled in the bead and the thermistor pushed up through the bead and the magnet structure of the flexplate system. The R2 has the thermistor on the not heating element side of a thick PCB so it is insulated from the actual bed temp.

I ran the exact same print on both units and was amazed at the difference. You can see that the R2 with the unmodified bed thinks it is up to temp long before it actually is because the thermistor is thermally isolated from the aluminum bed. The R2 with the thermistor at the top of the bed is slower because it is reading the temp of the aluminum bed.

 

I'm doing progress with the original heatbed.
I disassembled the surface from the heater and measure some overpower situation to check how fast and how stable this can reach the goal temperature.
Anyway, I have to test it with the SKR turbo board. I'm replacing the entire core from ROBO, but keeping some peripherals.
Soon I will have news about thermal isolation and replaceable surfaces.

 

I would not go more than 10% over voltage. I bought some E3D J5 24v hotends for $8 and used the aluminum block, heaters and thermistor, just crimping on new connectors. The hot end heater got up to temp real fast and the PID showed a low proportional number. I think it was probably a 12v heater because it only lasted 2 months before I observed wild temp swings and replace it. It would be a shame to burn out the bed heater.

Drilling a .067 hole where the current SMD thermistor is now and inserting a bead thermistor into the aluminum plate, then scaping the traces to the thermistor and soldering the bead to them allow for much better regulation and a higher top end temp on the plate. A drop of glue is then put on the surface of the plate. I did not find any problem with the existing heater. The PCB the SMD thermistor is on is an insulator between the bed and the measuring device ( thermistor ). Bypassing the insulator is more important than adding more heat.

 

I'm thinking of doing manually trough a variable voltage source.
And compare with the original behavior.
You meant could be better replace the current smd by an external thermistor?

 

Yes. I scraped off the existing SMD thermistor then drilled an .067 hole through the PCB and aluminum bed. I then inserted a 3950 bead thermistor through the PCB and into the aluminum bed. I scraped the PCB traces and soldered the bead thermistor to the PCB. I then put a drop of epoxy on the bed where the hole is. Before the epoxy fully hardened I scraped it flush.

Now my thermistor is measuring the bed, not measuring it through an insulator. This leads to more accurate regulation.

I have three beds and they ranged from 3 ohms to 4 ohms, which is a pretty big variation.