BrundleFab - a powder thermal fusion printer

Powder and inkjet printing
ezrec
Posts: 112
Joined: Sat Aug 15, 2015 1:31 pm

BrundleFab - a powder thermal fusion printer

Post by ezrec »

Hello, I'm Jason S. McMullan, based in Pittsburgh, Pennsylvania US.

I've developed BrundleFab (http://reprap.org/wiki/BrundleFab) as a test bench for thermal fusion powder printing (see the HP MultiJet Fusion for a commercial device http://www8.hp.com/us/en/commercial-pri ... nting.html)

My current focus is sugar printing, using the following process:

* Apply fresh powder layer
* Ink powder layer with a dark pigment
* Use a high intensity halogen light to heat the pigmented area to the melting point of the powder
* Let melted area cool to solid
* Repeat!

I'm still in materials testing, and the basic technology is working - still need to dial in the fusion pass (too hot and even the unimplemented area fuses, too cold and no fusion occurs) and the cooling phase (too short and the recoater will scrape off the semi molten layer).
ezrec
Posts: 112
Joined: Sat Aug 15, 2015 1:31 pm

Re: BrundleFab - a powder thermal fusion printer

Post by ezrec »

.. and here's my first multi-layer fused piece:

Image

Moving from granulated to powdered sugar now.

Oddly enough, powdered sugar appears to require more energy to get to the fusion point than granulated...

For Plan B, what powder layer height gave the best results? I need to re-tune my ink saturation, I believe...
User avatar
dragonator
Site Admin
Posts: 595
Joined: Fri Aug 14, 2015 4:48 pm
Location: The Nethelands
Contact:

Re: BrundleFab - a powder thermal fusion printer

Post by dragonator »

On Plan B, the last used settings were layer height 0.18mm, with inkjet printing at 250% compared to 96DPI and 3 sweeps per line (3 sweeps per 12 nozzles of width). This much ink and coverage was needed because the cartridges deposited less binder than it deposited ink.

Looks very promising. That is an actual printed part if I ever saw one.
ezrec
Posts: 112
Joined: Sat Aug 15, 2015 1:31 pm

Re: BrundleFab - a powder thermal fusion printer

Post by ezrec »

My next set of experiments are waiting for me to integrate an IR thermal sensor into the layer head.

I'm going to play with multiple passes of the fuser: a low heat pass to pre-heat the powder to a critical temperature, then a fast-but-bright pass to do the actual fusing.

Hopefully the preheat will make the fusing more consistent.
ezrec
Posts: 112
Joined: Sat Aug 15, 2015 1:31 pm

Re: BrundleFab - a powder thermal fusion printer

Post by ezrec »

Fusion Techniques:

Single Pass:

* Running the fuser slowly during warm up, then speeding up once hot, was moderately successful, but gave varying results depending on part bin powder temperature, and ambient temperature.
* A large problem was that if the powder "overcooked" it would adhere to the recoat blade on the repowder pass.
* Very hard to consistently control powder temperature

Multiple Pass:

* Video of the multi-pass experiment: https://www.youtube.com/watch?v=p7PbIBS-NHo
* This seemed to work better - the overall bed temperature rose slowly and consistently, and the final pass was able to fuse the inked areas only.
* I need to replace my fuser housing - it is not tolerant of the multi-pass technique
* I will need to add a IR sensor to measure the part bin bed temperature, and re-pass the fuser over the part bin until it reaces the target fusion temperature.
ezrec
Posts: 112
Joined: Sat Aug 15, 2015 1:31 pm

Re: BrundleFab - a powder thermal fusion printer

Post by ezrec »

Looks like some convergent evolution has occurred - I have reinvented High Speed Sintering:

https://www.sheffield.ac.uk/news/nr/gre ... g-1.392999
User avatar
dragonator
Site Admin
Posts: 595
Joined: Fri Aug 14, 2015 4:48 pm
Location: The Nethelands
Contact:

Re: BrundleFab - a powder thermal fusion printer

Post by dragonator »

The Multi-Pass technique looks great. I can't believe the strength you are getting and how unsintered the surrounding powder is. My only experiment with Nylon sintering only left me with weak wire frames. Your sugar has surprising strength for the fact that it is sugar. The smoking heat shielding is a bit awkward, but that kind of stuff is bound to happen when you use equipment out of their intended use, and luckily, it should be an easy fix.

I already saw the similarity and you beat me to showing it, but indeed, high speed sintering is almost identical to what you are doing. I have had a project similar to yours planed for nylon, but my project list is massive and never got around to ever doing the experiments. Yours is not yet really 'high speed' sintering, but the foundation is mostly there.
ezrec
Posts: 112
Joined: Sat Aug 15, 2015 1:31 pm

Re: BrundleFab - a powder thermal fusion printer

Post by ezrec »

If I invested in 15 ink shields, I think I could get to 'high speed' :P

In all seriousness, one of the key reasons my process is so slow (aside from the fact that I only am using one slow ink head) is due to unheated print bed, and that I have no thermal control over the powder, nor bed temperature monitoring.

Part of my next round of improvements, aside from the metal layer head, is to use a IR thermopile to measure the bed's surface temperature. Then, I can have BrundleFab dynamically adjust the fuser speed to get the powder to the right (sub critical) temperature after inking, then perform a fusion pass to cross the threshold to sintered.

Another speed-up would be to use a wide-format ink head (see http://www.memjet.com/products/wide-format) but MemJet's products are a wee bit out of my hobbyist price range.
User avatar
dragonator
Site Admin
Posts: 595
Joined: Fri Aug 14, 2015 4:48 pm
Location: The Nethelands
Contact:

Re: BrundleFab - a powder thermal fusion printer

Post by dragonator »

Heating the printbed would indeed help the speed. If you only need to push the temperature past the melting point by a bit, there is no need for 4-5 cycles (and no more smoking heat shields). The rising heat passing the printhead might be an issue, but that is a fixable problem.

While I totally agree that the memjet would make for an awesome printer, there are still plenty of printheads between the Memjet and the C6602A (luckily as well). My sight for the next 3DP is still set on the HP45, mostly because it is a massive step forward and I already know pinouts, voltages, layout and timing diagrams (in theory, but proven). The only issue is that I need to figure out how to control it. I have around 40 contacts that are active, and 300 printhead nozzles @ 600DPI. The nozzles are capable of 22kHz, full continuous cascade. I need a way of managing and controlling 6.6 MILION droplets of ink per second, all while moving and controlling a printer. I am still not sure what kind of controller could handle that, and just how streamlined the firmware would have to be.
ezrec
Posts: 112
Joined: Sat Aug 15, 2015 1:31 pm

Re: BrundleFab - a powder thermal fusion printer

Post by ezrec »

For a pen controller, I would suggest using a system that can initiate a DMA to a serial-to-parallel shifting IC - or a bunch of them.

For example:

Use 5 74HC164s attached to 5 of the DATA pins of a PC's parallel port. Use the 6th pin as a gating latch on a set of 8-bit buffers from the output of the 75HC164s.

The PC Parallel port can handle 2.5MB/sec (DMA'd) so:

2.5MB / 8 bytes per pen 40-bit control = 312.5K pen signals / second.

The 75HC164s can handle up to 150Mhz easily, so if you have a controller with a parallel interface you can DMA to (ie an external 8-bit memory bus) then you may be able to reach up to 18.75M pen signals / second.

Well within your sights, if you move from Arduino to something like the a slowed down Raspberry PI or other embedded ARM system.

https://www.fairchildsemi.com/datasheet ... VHC164.pdf
Post Reply