PeoPoly heater problem, analysis and solution

#1

After examining the wave motion on my and on the forum members’ prints, I could see a correlation between the waves and the temperature fluctuations.

The temperature of the PeoPoly heater is controlled by a two-position controller:

This control behavior leads to strong temperature fluctuations during control. The following figure shows the typical control behavior of the two-position controller (switches abruptly between two states):

unstet3
(Source)

The waves in the print image (in the direction of the Z axis) have the same cycle (and appearance) as the temperature fluctuations. The cause will be primarily the temperature-dependent the polymerization and the expansion of the cured resin and the entire Z-structure (see coefficients of thermal expansion).

The solution of the problem:

Elimination of temperature fluctuations by using a PID controller on the fan heater. This controller type is very adaptable and ideally suited for this application.

I chose the following experiment setup on my Moai 200 to test the PID controller:

PID-Heater%20V1

This heater (900W) I have controlled with this PID controller (with thermocouple and Solid State Relay). During the heating up and the normal control operation I recorded the chamber temperature of both heating variants. Here are the results:

Heating%20comparison%20V1

The temperature fluctuations of the PID controller during the print are <0.3 ° C. The heating phase is only about 4 minutes. Faults (opening the door, ventilating the room, sudden onset of winter, etc.) are compensated within a very short time.

Here is a sample print Raven Skull Fixed bottom by Grevan printed with the new heater (Nex Resin - Gray (HiTemp) at 30 ° C, Z height = 60um, FEP vat, total height = 70mm):

Another advantage of the PID controller with 230VAC PTC heating element is that it can do without a powerful switching power supply, thus causing less electromagnetic interference and also has a longer life expectancy.

Unfortunately, I have not found a ready-made plug-and-play version of such a small heater with integrated PID controller, which is why you have to build such a heater probably yourself or knows someone who can do that and has the appropriate expertise!

@peopoly: The optimal solution would be a in the Moai firmware integrated PID controller for heating.

6 Likes

The Peopoly Heater problem
Hardware assembly stabiliy
Print Quality - Z wobble
Layer lines, again
Striping on prints
#2

I have four heaters with PID controllers and have gone through 5-6 different ones that all got stuck on from the relay fusing from having too fine a control enabled. I have settled on ±1.5C. For this reason, and stopped buying any PID controller where I could not set the thresholds, and set wider thresholds on the ones that I have. The thermal mass of the printer smooths it out.

If someone really wants ±1.0C or finer control, I would get a controller with a solid state relay. They do cost a lot more though.

0 Likes

#3

love your work. now we have found the problem, lets see if Peopoly can come up with a solution.

I am not great with DIY electronics, so building a heater myself is not something I would do, but purchasing an off the shelf solution would be.

I would like to see a vat direct heating solution rather then wasting energy on heating the entire build volume, which is the cause of this wave artifact is that the printed part expands and contracts.

A direct vat heating solution would somewhat mitigate this as it only keeps the resin at higher temps by heating the metal vat.

The overall ambient temperature of the build volume would still fluctuate, but I imagine it would be a much smoother and wider curve.

0 Likes

#4

can you measure the gap between each way and see if it matches the time of temperature fluctuation.

There are hundreds of these heaters out there and the problem is not universal so we are getting several backs to test. It is possible the temperature flucation has effect on print but since when you switched heater, you switched not just the controller but also power supply and heating block. There are 3 moving parts, not just one so let’s see if we can narrow down further.

1 Like

#5

Thanks @johnchen . Heating the vat frame (Alu) directly is a good approach. This is e.g. made in the form2 and the wiper distributes the heat in the resin evenly. The Moai does not (yet) have a wiper … // I’m working on it, later

The hot air generator was only a test of the influence of temperature fluctuations on the print result. At the moment I am working on an infrared heater, which heats the glass plate evenly from below. That is a bit more complicated and there is still much to do…

1 Like

#6

There is still an unclear point in your test, and that is is the peopoly heater really not perturbing the galvos.
Would it be easy for you to use the Peopoly heater with the original PSU and use an arduino/tinkerforge or another similar board for the PID funtction ? Or maybe just use your current setup but adjust the PID to create a big temperature swing and see if you get lines ?

Most of my experience with SLA is using Formlabs machines but my guts tell me a 3°C temperature swing shouldn’t have such an impact on curing parameters, especially since the resin itself has quite a bit more thermal capacity than the air in the printer so it might not actually change that much.

0 Likes

#7

Thank you John for your request.

In my short contributions, I have summarized only the most important information about my experiments. Only if I’m pretty sure (>90%) that I’m not mistaken, I publish my findings.

For your questions and concerns:

There is still an unclear point in your test, and that is the peopoly heater really perturbing the galvos.

Before I installed the PID controller, I had a 230V heater with two-position controller (fluctuation: about 5 ° C, period: 85 s). This has switched much faster than the PeoPoly heater and caused a corresponding pattern:

The wave pattern only became visible in higher layers and stronger after the building board was out of the resin and the support structure was exposed. The heat capacity of the aluminum plate and the liquid resin will probably have smoothed the temperature fluctuations:

The carriers react very quickly to temperature fluctuations as they have a low volume and a large surface!

Theoretically, with a temperature change of 3 ° C, a change in length of approximately 12 μm results for only one carrier (assumed: 80 10-6 m / m / K). The geometric structure of the support structures also has a great influence on the layer displacements due to leverage.

For undisturbed printing results, a constant ambient temperature is a condition !!!

Would it be easy for you to use the Peopoly heater with the original power supply and use an Arduino / Tinkerforge or other similar board for the PID function? Or just use your current setup, but do you adjust the PID to create a large temperature swing and see if you get any lines?

At the moment I can not print and I do not have much time, because I’m 99% involved in other projects.

Also, I can not imagine how the simple 24VDC heater from PeoPoly could electrically affect the galvos. Do you have a theory?

0 Likes

#8

Hello Mark, aka @peopoly!

I would like to measure the grooves exactly and compare with the switching cycles of the heater, but I miss the exact total printing time, as this is unfortunately deleted after a print of the firmware from the display :confused: (firmware change request).

However, I have not the slightest doubt (see above) that the temperature fluctuations are the cause of the shift shifts. There are so many pictures of users who prove this. See especially The Peopoly Heater problem from @Whazaa:

Each groove corresponds to approx. 12 minutes. The piece shown has 25 waves and will therefore have been printed in about 5 hours. With a bit of phatasia you can even see the sawtooth from the temperature curve. Is not it?

0 Likes

#9

My only reason to talk about these issues is that Peopoly themselves have been saying, after user reports, that they are investigating faulty heaters (namely the heater’s PSU I think) for having an influence on the galvo and causing these isues. Now from what I’ve been reading these past few week on the forum it would’t surprised me if that wasn’t an issue at all and they are leading a meaningless investigation. just like the “enhanced” Z axis.

From what you wrote and the evidence you provide you definitely seem to have put your finger on the issue !

0 Likes

#10

Hello John and everyone else,

It may be that the power supply of the heater or the temperature controller has a defect. But then he does not heat or he heat constantly.

I used this power supply to power an electric motor (<4A). It has turned off after some time and after a few minutes without power it then worked again. Then I connected a resistive load and let a current of 4.7 A flow for about 5 minutes. This has mastered the power supply without any problems. Also, the oscilloscope showed no abnormalities on the 24V line (without load: 24.2V, with load 23.5V).

Well, the power adapter may be cheap, but it is quite suitable for use with the heater. The voltage on the running heater showed on the Oszi spikes of 2.3V with a frequency of about 500 kHz (probably from the fan). A 1000μF capacitor in parallel with the heater completely eliminated the spikes. However, these spikes can not affect the galvos or their voltage signal!

The grounding concept of the Moai is not ideal. The frame and the steel plates, the galvos, their cable shield and the driver heatsinks are more or less at the same potential, namely GND. That should be enough to derive strong electromagnetic interference from outside. Better would be a complete shielding of the electronic modules (CPU, power supply and galvo driver), but not necessarily.

The problem of waves (layer shifts) in print is guaranteed not to be caused by external electromagnetic interference.

The separate power supply of Moai and heater is also fine. And it is also an elegant solution to use only low voltage in the device. This is safer and reduces the protection costs considerably .

One/the problem is the temperature fluctuations inside the printer !!!

In my view, the temperature should be kept as constant as possible (fluctuations <0.5 ° C).

Sure, there are also other causes for a distorted print image. These include the “shaky” construction of the Z-axis on the Moai 130 and the large peeling forces.

There is a clear feature when the temperature is the source of the disturbance:
The distance between the waves depends on the printing time of the layers.

If the Z-axis does not turn “round”, then the distances between the disturbances are constant, no matter if much or little was lasered on the planes.

1 Like

#11

We have received the power supply and the heater from @johnchen initial check of the adapter looks very stable with analyzers showed the pattern to be consistent with a quality adapter. We do not believe this adapter is the cause of his problem with pattern. John also reported that once he turned it off, the issue was gone. This matches what @Whazaa see in his tests, as with @ad_fontec We also could not reproduce John’s results here in southern China where the temperature is hitting 30C in the day time now.

This would sense if the pattern is related to how much temperature is changing within a time frame and how much heat is lost during printing is highly dependent on the environment too. We have sold a lot of these heaters, and while a small percentage of users have problems with either bad adapters or this wavy pattern, most users don’t. When there is a problem that is related to temperature change, what is the solution? I suspect the issue is not how high or low the temperature swings but how fast. This is supported by the fact that we have printed at resin at 35C indoor (no heater) and at 15C with no heater and the results were not like you got. We can test this by putting a thermal layer that reduces heat loss during printing and thus extending the time from peak to valley. Can you guys do this test with a small print? I know several users put a cardboard box over Moai during printing, and one user reported his success of removing wavy z-axis by wrapping his Moai in an aluminum wrap. We don’t need to go that extreme but adding a box over it would slow the heat loss.

0 Likes

#12

Hi! I also think that the problem is caused by too fast temperature fluctuations. The ambient temp in my workshop is around 22c. When I use the Peopoly heater in my Moai 130 during printing I have noticed that the temp in the build chamber drops to about 28-29c before the heater element kicks in and when it does the temp rises very fast to about 34-35c (so fast that my digital temp meter that shows the ambient temp near the vat on the other side of the heater jumps from eg. 29c directly to 30,5c and then directly to 32c…).

0 Likes

#13

@Peopoly - really strange that you could not replicate the problem. which does seem to indicate temperature fluctuation.

I have since switched to a direct vat heating solution using a belt heater for brewing. This works perfectly on the M200.

This is not temperature controlled but gives a nice constant heat of 30c - 60c adjustable via dial.

I have this setup on an analog timer and have it 1 hour on 15min off. This is just so I don’t over heat the heater. The nice big metal vat holds the heat well, but does not get too hot. so during the off time, the temp drop would be much smoother.

These things are low voltage ( 35w) and designed to be left on for a long time, so I think its a safe and effective solution without too much work.

1 Like

#14

we and @Whazaa believe the effect is due to how quick temperature changes, and NOT peak temperature. This has as much to do with environment temperature and enclosure and the way heater works. @ad_fontec believe the pattern is caused by the printed part expanding / shrinking during rapid temperature that affects the print. This would be consistent with why your conductive heater works better because it does not affect the temperature of printed part that much

1 Like

#15


Heres the belt heater on my Moai 200.
No more waves and prints turn out great!

3 Likes

#16

Congratulations John,

the (silicone) heating tape on the frame of the VAT is really a smart heating solution. Simple, effective and pretty to look at. :+1:

:slightly_smiling_face: A great advantage is that it continuously heats the VAT frame with the set heating power, without the rule fluctuations come into the system.

:slightly_smiling_face: Due to the tilting movement of the VAT, the heated resin on the edge is distributed more or less well throughout the VAT.

:slightly_frowning_face: However, this unregulated heating has the disadvantage that it continuously brings energy into the system. As a result, the temperature continues to increase until equilibrium with the energy released occurs.

:slightly_frowning_face: Another disadvantage, in addition to the non-constant temperature, the final temperature is not reproducible. The final temperature depends on the outdoor temperature and the set heating power.

:slightly_frowning_face: Another disadvantage is that with a heating power of 35W it takes a very long time until a stable temperature is reached.

:smirk: All in all, however, this is a better heating solution than the fan version of PeoPoly, which heats up an unnecessarily large area of ​​the printer and brings high temperature fluctuations into the system!

The following points could greatly improve the heating solution by @Johnchen:

  1. Heating foil on all 4 sides of the VAT frame
  2. Higher total heating capacity (approx.> 200 W)
  3. continuous temperature control (PID controller) with solid state relay

:neutral_face: However, a disadvantage remains systemic, in the middle of the 240mm VAT, it is always a little cooler than the edge. Maybe this is not a drama. :hugs:

:point_right: PS: Very important: Temperature fluctuations are poison for precise machining processes!

3 Likes

#17

Hi Bernhard,

thanks for the reply, I agree with all your points. :slight_smile:

The belt heater solution can definitely be improved, however for an off the shelf solution costing less then $20 from ebay that did not require any complicated setup, this was perfect for me.

I print mostly statues and organic objects so absolute dimensional accuracy is not as critical for my requirements. A bit of sanding if required is no big deal.

But I agree, that for the best accuracy of result, a temperature controlled environment is a must.

I have proposed to @peopoly they could take this idea and improve it as an official product.

For now I am quiet content with the belt heater :slight_smile:

2 Likes

#18

While you still have the existing set of heater, can you run a print and with printer covered by a box or something like this to see if it improves (by trapping the heat) ?

0 Likes

#19

I’m printing now with the original heater setup with the printer inbedded in ”sun protection foil” for car windshields. I will report result.

And I’m back with some confusing results.

In this first picture you can see the lines though they are not as visible as before in my opinion. Please notice that I printed another object at the same time that was lower in height compared to this sword. Thus the space between the lines differ according to how long it took to print each layer. When the other object was finished the space between the lines stayed the same. So far so good…

Then when you look from the side the lines make a wave pattern - if temperature change to the resin / printed object was the cause for the lines they would not go outwards on one side and inwards on the opposite side of the sword. I’m a bit confused and these results suggest that the heater does affect the galvo system so that it shifts just a fraction during heating.

1 Like

#20

Thanks for the pictures of your print results @Whazaa.

The layer shifts, in spite of “winter cap”, confirm my above observations and measurements.

With the whole-body insulation, the system-related control behavior of the Peopoly heater is not changed (two-position controller). This means that the controller causes inside a temperature fluctuation of about 3 ° C (as long as the outside temperature is <33 ° C).

In particular, the supports respond very quickly and sensitive to temperature changes due to change in length. This leads to the shifts in the planes, which become particularly visible when the object is at an angle and the supports become longer.

What can you do against the waves? :thinking:

First, maintain a constant temperature throughout the printing process.

A good solution is a constant heating by @Johnchen or a fan heater with PID controller and SSR.

:point_right: Tip: only heat if absolutely necessary.

1 Like