Things that still suck about the V1 printer - according to the founder

[User_Defined]

It is unfortunate you could not go with the PCB coil method I suggested, it would make for much more secure interconnections (and more consistent coil inductances).

As for the remaining issues, I think that the most significant bug is that the final prints are not accurate, and it pretty much throws anything I would use the printer for out the window.

I'm sure that you guys can find a better calibration method, and as a backer, I hope that you prioritize this as soon as possible.

-UD

[rylangrayston]

It is unfortunate you could not go with the PCB coil method I suggested, it would make for much more secure interconnections (and more consistent coil inductances).

As for the remaining issues, I think that the most significant bug is that the final prints are not accurate, and it pretty much throws anything I would use the printer for out the window.

I'm sure that you guys can find a better calibration method, and as a backer, I hope that you prioritize this as soon as possible.

-UD

Ya that PCB coil idea was really great, I dont think we can get the power ( turns ) we need out of it tho. Our coils have many hundreds of turns in them
and they are rather small. Ether way fundamentally is a perfect idea. I want to use it somewhere. And I can see a good place for it in peachy 2.0 galvo system
As for calibration, it is coming soon, after we implement variable laser power, and acceleration, Calibration is next on the list.
One step we have taken toward better calibration is we have come up with a better UI for entering the calibration points.
You can now move the laser beam around with your mouse and click on various places on grid paper to set calibration points. Its much faster than our previous method and I can now imagine it being reasonable to enter in 100 plus calibration points.

PS UserDefined
quite some time ago you asked a REALLY good question, something that is constantly on my mind and still is.
Why is the peachy printer able to print perfectly smooth columns but dose not achieve the same smoothness on other prints?
This has botherd me ever since I saw that first green column print.
I still don't have a definitive answer for that, I have tested and disproved many theory's so far, and I still have more theory's to test.
When I find out why I will make a thread on this. My apology's for not answering you with a simple, I dont know. I remember composing a rather in depth
answer in my head including all my theory's at the time, but Im sure I never went back and actually typed out a reply to your question.

[User_Defined]

Ya that PCB coil idea was really great, I dont think we can get the power ( turns ) we need out of it tho. Our coils have many hundreds of turns in them
and they are rather small. Ether way fundamentally is a perfect idea. I want to use it somewhere. And I can see a good place for it in peachy 2.0 galvo system
As for calibration, it is coming soon, after we implement variable laser power, and acceleration, Calibration is next on the list.
One step we have taken toward better calibration is we have come up with a better UI for entering the calibration points.
You can now move the laser beam around with your mouse and click on various places on grid paper to set calibration points. Its much faster than our previous method and I can now imagine it being reasonable to enter in 100 plus calibration points.

PS UserDefined
quite some time ago you asked a REALLY good question, something that is constantly on my mind and still is.
Why is the peachy printer able to print perfectly smooth columns but dose not achieve the same smoothness on other prints?
This has botherd me ever since I saw that first green column print.
I still don't have a definitive answer for that, I have tested and disproved many theory's so far, and I still have more theory's to test.
When I find out why I will make a thread on this. My apology's for not answering you with a simple, I dont know. I remember composing a rather in depth
answer in my head including all my theory's at the time, but Im sure I never went back and actually typed out a reply to your question.

No worries about not answering the questions. Naturally I assumed that your team is trying to make the best possible prints and would be pursuing that goal if you could.

The question was more to catalyse some thinking. I found the stairs of the purple rook to look very glassy, and thought it had just been a question of layers.

It might simply be that the glass columns were printed from high up and were straight, so the laser was interacting less with surrounding resin and produced a consistent surface as it went.

It could simply be a question with laser power, spot size, layer numbers, and print time. If you print it slow and take your time it may come out better?

In any case, Im already super impressed with the prints so far, and with better calibration for accurate parts, the peachy will have surpassed my expectations (which were still high) by far.


-UD

[rylangrayston]

It could simply be a question with laser power, spot size, layer numbers, and print time. If you print it slow and take your time it may come out better?

-UD

"It could simply be a question with laser power, spot size, layer numbers, and print time. If you print it slow and take your time it may come out better?"

yes longer print times small spot sizes and slower laser seems to help, but more than 100 layers per mm dosent seem to make any difference.
No matter what settings I use there is often some subtle oscillation leaving rings on a print, it seems to happen with a period of about 50 layers.
You can see it in this rocket, pronounced near the tip.

Attachment 6564

more layers per mm probably would eliminate the patters you can see on the tops of each stair here( altho you cant see these with just the human eye) :
Attachment 6565

my latest Theory's are things to do with the way resin flows during the print, as well as how air may be affecting the surface of the resin.
Im coming up with test prints that can help support or disprove what I think is happening. I feel like we are very close to really nailing perfectly glossy prints( as far as the human eye is concerned).

[User_Defined]

"It could simply be a question with laser power, spot size, layer numbers, and print time. If you print it slow and take your time it may come out better?"

yes longer print times small spot sizes and slower laser seems to help, but more than 100 layers per mm dosent seem to make any difference.
No matter what settings I use there is often some subtle oscillation leaving rings on a print, it seems to happen with a period of about 50 layers.
You can see it in this rocket, pronounced near the tip.

Attachment 6564

more layers per mm probably would eliminate the patters you can see on the tops of each stair here( altho you cant see these with just the human eye) :
Attachment 6565

my latest Theory's are things to do with the way resin flows during the print, as well as how air may be affecting the surface of the resin.
Im coming up with test prints that can help support or disprove what I think is happening. I feel like we are very close to really nailing perfectly glossy prints( as far as the human eye is concerned).

If it truly was surface tension on the print layer, one would need to explain why a straight column prints so well.

It has to be something, especially if it is evenly distributed along the print. I'm thinking software or physical/mechanical. I would see if you can line up those errors with a video of the print and see if you can spot them as they are being made.

I would also check for linearity over the print surface. Maybe the column has no variations over its height and doesn't suffer from nonlinear effects as much.

Also see if you can print different shaped columns and see where the boundary is between glassy smooth and beginning to get rough.


-UD

[mike_biddell]

If it truly was surface tension on the print layer, one would need to explain why a straight column prints so well.

It has to be something, especially if it is evenly distributed along the print. I'm thinking software or physical/mechanical. I would see if you can line up those errors with a video of the print and see if you can spot them as they are being made.

I would also check for linearity over the print surface. Maybe the column has no variations over its height and doesn't suffer from nonlinear effects as much.

Also see if you can print different shaped columns and see where the boundary is between glassy smooth and beginning to get rough.


-UD

I think it does explain exactly that. This is quite difficult to explainpeachy.jpg
I hope my cr*p sketch explains it. And you must understand this is only a theory. But it does explain why it does not affect even columns. The effect of the surface tension is to delay Z. The printer is printing Z, but as there is no resin layer on the object, therefore this printing has no effect and these Z values are 'lost'. When wetting occurs, the old Z values are not re-printed and if the object is tapering, the new X,Y values are smaller than the x,y for lost Z values, thus creating a step.
With a cylindrical object, it doesn't matter about the lost Z values, because post wetting, the x, y values are the same and the resin is hardened without a step. I hope that makes sense, it is quite difficult to visualise.

[mike_biddell]

This problem is difficult to resolve apart from printing through the bottom of the tank. But there may be a potential improvement in delaying Z. That is you print nothing for 50 layers of Z and then start printing. Then experiment with Z delay until you get the best print.

On reflection, this wont make any difference, there will always be a point where you are printing with a bulging meniscus and no resin on top of the print and hence you are printing into 'thin air'

I'm rambling on ..... LOL........... heating the resin could significantly reduce the surface tension of the resin and could be worth a try

I've just thought of a serious contender for a solution. As you will know, surface tension is the uneven pull of the molecules into the surface of the fluid, because there are no molecules above to pull equally. Well why not float say 1 cm of ordinary tap water on top of resin, to equalise the pull and neutralise the surface tension. Firing the laser through the water. The water should float on the resin easily.

My theory is that this should neutralise the surface tension in the resin layer and print the missing Z values.

Angles must not be too steep or reflection/refraction could occur in the water layer.