James Cooper - Peach Printer Software Engineer Chimes in

[MartinPoets]

Just found this from James Cooper, the guy that is writing software for Peachy Printer. Lots of interesting tidbits.

James here. I'm the guy writing the software that turns the Blender model into audio to drive the Peachy.
First off, don't worry -- this isn't becoming an iOS or Android-only thing. In fact, that app has nothing to do with the main development of the printer itself -- it's something nice that the guys at 52apps did on their own because they are interested in the printer. We aren't spending our time on mobile apps when we know what you really want is to see a better print.I'm working on the Blender and Python code for the printer right now. My main focus is on getting better prints. Our original prints were done without any correction for projections and nonlinearity and without proper dampening, so things were warped and sharp transitions had noticable ringing. Over the last week I've already addressed much of this. We would love to show you the results, but the printer is currently in pieces on my desk while I'm testing and developing my code. As soon as the code is done, we will make some prints and show you the results. At the same time, Rylan is working on designing better galvanometers that will move faster and more precisely while others are working on taking this from a handmade project and turning it into something that can be manufactured.Please understand that it takes time to research and build things. We are doing all we can to provide what you are asking for and produce videos of better prints. As Rylan has said, we weren't originally planning to do major development during the campaign, but since you are all excited to see more prints, we are attempting to do exactly that. But these things take time to develop -- weeks are a short time in the worlds of hardware and software development. Up until now, we have also been doing this in our spare time. I work full-time as a software developer for another company. I spent my entire Thanksgiving weekend working on the printer and I still have some work left to do. So we are doing everything we can to show you what the printer is already capable of, but please understand that it takes time for us to respond. We can't show you better prints until the work has been done to make them better.
Even when I do get this work done, there will be more to do. With several months of work, better manufacturing, and feedback from beta testers, we will be further refining the printer to get better performance while keeping the costs low. This is why we started a Kickstarter campaign and it's why we are offering a limited number of beta test spots many months before our target release. What you've seen so far and what we hope to show you soon is still only the beginning of a long process of refinement. With some hard effort from us, enabled by your support, we plan to have something much better than what you see today. But this is a project in development; you are supporting development, not buying a finished product. Kickstarter is not a store: the final product we ship you will be a reward for supporting us.
Thanks again for your patience. We are as excited as you are and we are doing everything we can to deliver more videos for you before the campaign is over.

: I wouldn't say it's at the stage of "final tweaking". It's a proof of concept. The printer you see in the video is the sole prototype of the last revision -- the one that works well enough to demo. What we're doing now is improving it further. Part of that is designing better galvanometers. Part of that is designing software to drive it. All of it takes time and money. Thanks to this Kickstarter campaign, we'll have the money -- now we just need the time.We will be doing test prints likely this week -- Rylan should have another revision built then and I should have the software written to drive it. Once it's ready, we will show you what it's capable of right now. I'm also trying to organize with Rylan and Nathan to show you what we've done over the last week.You're right to think that, by the point of "final tweaking", we should have multiple printers all making test prints. That's exactly where we plan to be by the time we are releasing units to beta testers, which is currently planned for December. But that's not where we are now. Right now we have one prototype with some known issues that we're addressing. Given the overwhelming response to the campaign, we're also madly trying to line up manufacturers, source different parts, and take advantage of volume purchases to get higher quality components for the same price. I can't speak personally for how much of a difference this is making to the plan, but I can say that it will result in a better printer when it's done. Until then, we are continuing to work on R&D and we are making progress towards better prints.I agree that there is some inconsistency in the messaging, but that's largely because of how quickly this is moving. What was one man in a shop making a prototype has suddenly turned into thousands of backers and lots of contributors all wanting to help out. Now Rylan has myself and at least two others working part time on improving the printer and other contributors are providing all sorts of ideas of risks and opportunities. It's a lot in a short period of time and it has resulted in some fast changes in what we plan to do next.If you are apprehensive about backing this, we understand. After all, it is in development and there are still risks. That's part of being a Kickstarter project. What we have is an idea, a prototype, some plans, and a dream. There is still lots of work to be done and many improvements to be made. That's why we're on Kickstarter looking for funding. That's even more true if you plan to be a beta tester. Unlike beta tests you might be familiar with from AAA video games, we won't be shipping you a gold version just to make sure the servers don't melt under the load. Instead, you'll be getting a rough version that's closer to what we expect to deliver, but is also expected to have problems that will be found during testing and will result in further improvements to the design. There's still a long time between the beta tests and final production and lots of design changes can still be made.I hope this helps to better illustrate where the project is at and where it's going. This isn't a final product, but an active work in development. Your money is investing in us so that we can design and produce a printer. When we're done, you'll be rewarded with a printer of your own before the rest of the world.

[iplayfast]

I had an idea on how to solve the ringing problem. If you have two mirrors with the laser bouncing off both of them producing a raster like effect, then all you need to do is vary the laser intensity and sync to the top of the screen like old crt monitors do.
This would solve the ringing problem because the motors would be turning at a constant speed.

[rylangrayston]

I had an idea on how to solve the ringing problem. If you have two mirrors with the laser bouncing off both of them producing a raster like effect, then all you need to do is vary the laser intensity and sync to the top of the screen like old crt monitors do.
This would solve the ringing problem because the motors would be turning at a constant speed.

Right this is called rastering
I think the peachy printer will be able to do this but it comes it comes with some huge down falls!

1 The print would become much slower as the laser would spend most of its time in the off state.
2. We would need to turn the laser on and off Very Very fast to get great resolutions with this method

Having said that I still want to make the peachy printer do rastering for the purpose of laser light shows!
Good thinking, if we do this jsut right it could actualy take advantage of the ringing!

Currently I have plans to pulse the laser at a photo diode to see just responsive it is.
It certainly possible for 405 nm lasers to be switched very fast but we need to see how the closed loop
current limiter that comes in the laser module acts when pulsed.


Rylan Grayston

[RomsMaklaet]

Peachy will turn out in the end.
sbobet

[Morten]

I always wondered why you need to turn the laser on and off at all. Why not have a small disk with different sized openings (and a part of it w/out any opening) in front of the laser. Rotate 30 degrees to select a small opening for the laser, 60 for a large opening (infill) and 90 degrees to block it completely. Or something like that.

[Slatye]

I always wondered why you need to turn the laser on and off at all. Why not have a small disk with different sized openings (and a part of it w/out any opening) in front of the laser. Rotate 30 degrees to select a small opening for the laser, 60 for a large opening (infill) and 90 degrees to block it completely. Or something like that.

Probably just because of speed. Even a super-quick servo will take maybe 100ms to get reposition the disc, whereas the electronics to turn the laser on/off can probably operate in microseconds easily. Then there are all the issues with fast mechanical devices - wear and tear, backlash, vibration, electrical noise (digital servos do horrible things to the power supply lines). Finally, if the disc is blocking the laser, either it's absorbing that energy (and warming up) or it's reflecting it (potentially onto something that won't appreciate a lot of UV light, like the bottle of resin sitting next to the printer).

[Morten]

Probably just because of speed. Even a super-quick servo will take maybe 100ms to get reposition the disc, whereas the electronics to turn the laser on/off can probably operate in microseconds easily. Then there are all the issues with fast mechanical devices - wear and tear, backlash, vibration, electrical noise (digital servos do horrible things to the power supply lines). Finally, if the disc is blocking the laser, either it's absorbing that energy (and warming up) or it's reflecting it (potentially onto something that won't appreciate a lot of UV light, like the bottle of resin sitting next to the printer).

Great reply, thanks! I was thinking more of doing it the same way that the mirrors already move, a black disk on a string with several different sized apertures and a blocking part, but can see it would be way to slow to raster with and maybe heat up too much (altho I'm not sure, you would only block the laser every so often).