Tears before bed time

I ordered 10Kg of Crystacast plaster to try in the machine. Open3dp's favourite plaster is Hydroperm but I failed to find that in the UK. This plaster looks the business and indeed feels it but there are some initial problems.

It is not spreading properly and the surface is full of tears. I think that perhaps the particles are too fine or perhaps not dry enough? When I watch the powder being pushed along by the roller instead of seeing a tumbling mass like a breaking wave I see actual "waves" being formed. This is I suspect due to the non Newtonian behaviour of the powder, when compressed it acts like a solid, forms the strange lumps then tears up the surface underneath. I sieved the powder, perhaps that was my mistake?

Secret ingredients

Mark from open3Dp suggested that I add about 10% glycerine to my mix to help it bond, he also suggested I try plaster instead of the starch based powder I have as a longer term solution.

I ordered some plaster that I hope will be suitable and also tried the glycerine. Firstly I noticed that it seemed to print a little darker, perhaps the cartridges prefer the modified viscosity, I don't know. After the print I left it for an hour and there was a definite improvement however it was obvious that I would not get the model out in one piece. So I lifted it out along with a lot of powder and popped it in the oven for a while to try out. The parts are definitely stronger than before but interestingly the layers do not seem to have stuck to each other.

Two chain links, unfortunately non of the vertically orientated links survived due to a lack of strength because of the lack of a bond between the layers.

And here we have the layers literally splitting apart. Seems a bit odd, is my saturation too low? Is the layer thickness too great? Not sure.

Printer poop

Here is a test mix of 30% binder by volume. Much darker than the print and has much more strength, not that it could have much less.

Homebrew

This afternoon I mixed a batch of homemade binder, this is from a recipe from open3Dp called XF1. It is basically distilled water along with some IPA and a little food colouring. It runs through the print head fine:

But alas it does not bind the power, it is no different to the distilled water that was previously installed. I'm thinking the binder was developed to work with an open source powder or I have to leave it longer or change the saturation settings or similar. Frustrating but a good test of the machine.

Almost done

So I am all but done bringing the Z402 back to life, only minor issues now. To recap this is what I have had to do so far:

1. Found mispositioned ribbon cable that prevented the internal PC from booting.
4. Removed printer assembly to clean "car wash" then removed gantry to thoroughly clean.
3. Cleared blockage to input of binder filter.
4. Replaced o-rings on priming pump so it works.
5. Found damage to print head ribbon cable and repaired
6. Slow axis then failed, found problem with control relays, temporarily bi-passed.
7. Repaired binder low-level sensor, this was just a broken wire and an annoying constant error message :).
8. Replaced roller scraper, I used a piece of polypropylene from a plastic folder copying the old scraper and using a soldering iron to make neat holes (scrape off the kerf with knife).
9. Generally cleaned the heck out of it.

Yet to do:

1. Replace the relays on the interface board
2. Make or obtain binder
3. Make new gantry "curtains"
4. Work out why the laptop doesn't work.
5. Possibly replace some binder hoses.
6. Light bulb
7. Replace missing fixings

My god, it actually works!

There, I fixed it

So after some intermittent powder spreading, the slow axis (the gantry) stopped working all together. I spent the day trying to draw out the circuit for the servo stage which includes an h-bridge chip and some relays. The relays are used to interlock the axis so it cannot move when the lid is open and also to apply a brake by shorting the motor terminals. I felt sure it would be the relays but after a lot of playing still did not know how to prove it for sure, they get power but did they switch, and did I dare apply my own power to test the switching? So I decided to bi-pass, don't try this at home kids:

I have tacked the wires on to the output pins of the h-bridge and pushed them into the red connector at the bottom, I even got the polarity right (by accident). And .....

It works!!  So I will order replacement relays and sort out some binder, the original Zcorp binder is crazy expensive so I will make my own. I also have 6 BC-20 ink cartridges on there way to me so I am covered there too.

"binder"

The part printed!! Well sort of.

I waited a while and proceeded with removal which involves a vacuum cleaner and a lot of care. Ooops, there goes part of the model, hang on a minute..... It looks like the bottle of binder actually just contains water! I can see that the parts printed OK but they have zero structural integrity!!

Then I came to re-spread the powder and the old issue of intermittent spreading reared its head. Not sure what it causing this, it never happens during the build and it can happen even when the feed bin is lowered and no powder is being pushed. I have added grease to the linear bearings too. Hard to diagnose the issue and no useful error messages. Perhaps the roller mech is having problems and the gantry stops because of that or perhaps there is just too much friction in the gantry but that makes no sense as it always works fine when printing. Mysterious!

I really wish I had finished my homebrew Zcorp printer, it would probably be less hassle than this ;)

The model consisted of some chain, this is part of one of the links that did not shatter, on the left impregnated with cyno on the right as it came out on the machine.

Here we go...

I know this is a blog not twitter but I am currently watching the Zcorp print and I think it will finish. There was definitely a problem with my laptop or the usb to serial converter, possibly because I had to force it to be COM1. Now using an old Dell and all seems well.

Now, I suppose I should read the manual because I am not sure what I do next :)

Every cloud

It looked like both of the projects I have on the go at the moment had stalled, the Zcorp because of a lack of an ink cartridge and the laser scanner because of a lack of component availability. So I settled to have a weekend doing all of the things I should have been doing this past week. I emptied a ton of rubbish from my car, emptied the bins and even started to look at the washing up. While considering which box I could use to tidy up all the Zcorp bits into I noticed a small cardboard box, could it be, no don't get your hopes up, YES another cartridge, an original zcorp spare even! . . .

OK I got a little excited taking off the protective cap!

Success! Well another good step, this is printed in binder and it seems to have worked. I have a few powder spreading issues, I definitely need to replace the scraper on top of the roller/spreader but it's better, definitely better. But then .....

Nooooo! This has happened a few times, might be the PC end not the machine, will try again and then with another PC.

Better than nothing

The results of replacing the repaired ribbon cable are promising. Perhaps I killed the cartridge running it with the broken cable and so only some jets are firing? I will try again with the proper cart and binder.

This is a layer from the middle of a demo object consisting of a chain and a little disk stating the build time. But it should not be so stripy!

Micro surgery

So here is how one repairs a flex PCB, well at least I think it is repaired, the proof will be in trying it on the 3D printer. Photos are from my phone peering into the eye piece of a stereo microscope.

Well, there is one very obvious break! The line above was also broken at the point where the Kapton covering ends. I made two light cuts through the tape and peeled it back and trimmed it. Then I scraped any remaining adhesive from the tracks.

Now the tracks were tinned with a lead/tin/silver solder, it is the thinnest I have at 0.5mm and a lower melting point than most.

To make the repair I took a single strand of tinned copper wire from a piece of multicore hook up wire and flattened it in smooth pliers, then I tinned it and soldered it in place using a flux pen to ensure the solder flowed even if it took me a while. I flattened the wire because I was worried that it might sit higher than the little nipples that form the contacts. So that is two done, finished! Yeah right.

On higher mag I could see that this poor line had a further two breaks!

Much clearer when the Kapton is peeled back

This time I used a thinner wire from the multicore that is found in mice leads. It is really nice wire for many projects so save those mice cables! Here you can see how I made a little handle using the wire, if the wire and tracks are tinned you just need to press down with the iron and they bond, extra flux does help. The wire was trimmed by placing the scalpel on it before wiggling it a little, remember this is super thin stuff!

The other repair was on a bend and does not look so pretty. This joint and the others where carefully shaved with a scalpel to reduce their height, probably not needed but did not hurt.

All done! Well not quite, when I found someone to assist I did electrical tests from the other end of the ribbon to this end and found two more breaks, again always at the boundary of the Kapton, this is simply where stress concentrates. I hope it works after all this effort. If it does I can start looking for an old printer to ravage.

Seek and ye shall find.

I had what I thought was a good idea, put a normal ink cartridge in the 3D printer, one with ink in and then see if it would print, that would rule out binder issues and perhaps I would get to see some printing. The printer's cartridges are based on Canon BC-20's and they simply decapitate them, remove the ink and wadding and glue a plastic screw top vial that has had its bottom cut off over the hole in the bottom of the cartridge. These vials have a screw cap which is compatible with the machine hose lines. I was given a bag of the vials so just hooked the lines to an empty one, it certainly proved that the binder pumps OK.  But alas still no joy, now I was getting errors, too many deprimes! This seems to be when the fluidic connection between the reservoir and nozzle is lost, seemed unlikely it was caused by the fresh out of the package ink cartridge. A guy on the forum suggested it might be the flex pcb type ribbon cable that runs to the cartridge so I took it off and had a look. And it seems it is definitely damaged.

This is the part that contacts with the cartridge. Now let's get all CSI on its ass, zoom in on sector D5.

And there is our problem, one of the tracks has broken and even peeled back, now give me a hard copy of that. But BladeRunner references aside the cable and I have a date under the stereo microscope with some very thin wire and a very small soldering iron. I'd better lay off the caffeine now! Fingers crossed this is the issue because I am so close now to printing something.

Progress slowing

Well I got the printer back together and it turned on again without incident, tried adding some powder and it spreads but not every time, sometimes the gantry just stopped. But I thought I would press on so after adding a little grease to the rails I set it to print. No joy, I got an error concerning the print cartridge and then the software running on the machine crashed. Today I put in another cartridge which I am sure is knackered and it started to go through the motions but no juice came out, then I got an error related to the serial buffer which may be because I am using a USB to serial converter but I have no idea. I think I have one normal ink cartridge that will fit so I may try this just to check the signals are getting through OK. I also have an old PC with a parallel port which I can try. Starting to lose some confidence I can get this going but must push through that, do things step by step and get in to the details

Spread em! The cover over the print section is removed here, as are the curtains the gantry pushes through.

Looking like a real machine

 So this is the initial reason for the clean up, the "car wash" was really encrusted with powder and binder. But look a few minutes in hot water, a little tooth brush action and then a little more with a tooth pick and it is good to go, quite a transformation.

At the bottom you can see the squeegy that wipes the print head and above that a little box that makes a seal with it to stop it from drying out. That rocks in to place when the carriage hits the little button at the bottom right. If you need to clean this thing then you just need to disconnect the ribbon cable and the fluid lines from the printer assembly (the upright bit on the gantry) and then lift it off the machine. The car wash is held by a single screw but mine needed prising because it was glued in place with binder.

That looks a lot better doesn't it. The chute for the excess powder is removed in this picture for cleaning too and next to the hole you can also see I replaced the velcro strip. I think this is for cleaning the powder spreading roller. You can see at the back of the machine the "curtains" though which the gantry enters the build area, this keeps powder from the back of the machine, at least it did! It seems to be the braid that is used to protect cables laid flat with some soft plastic sheet inside. I should be able to source something similar to replace it.

And the Gantry all cleaned up. The main gantry went in the shower for cleaning, then I blew it dry with dry compressed air and oiled the bearings etc. Top left is the printer assembly, this just had to bee wiped carefully, not something to dunk really. I cleaned the linear rail, wiped the limit switch (on back of carriage) and not much else, will grease the ways at some point. Top right is the funky plastic thing that passes through the curtains, pointlessly complex design if you ask me. Reassembly should happen tomorrow night, maybe...

Clean up on isle 3

I decided to clean the "car wash" the part that wipes and stows the print head when not printing. The manual says remove and clean under the tap. I tried to remove it but it seemed that the entire printer section would need to be removed too. The manual does not say how to do this so I started unscrewing things and convinced myself that four bolts holding two plastic upright brackets would need to be removed. Turns out this was not the case, the print assembly can be lifted out of said brackets indeed you can see them in the photos below. But once out I could really just how dirty everything else was, not just coated in powder but hardened powder. I decided to remove the gantry, this was not very easy to do but was possible. I will try to document it when I put it back together to help others.

So here is the gantry with the print section removed, you can see a slot that it prints through and also the slot where the powder roller/spreader sits.

And another view, on the left is the powder chamber which is currently lowered and on the right the build area and the power chute which is connected to a vacuum system. I don't like the idea of the vacuum system because it is noisy, it would be better with a gravity system with just a slight negative air pressure IMHO, but what do I know at this stage.

After removing a lot of screws, some of which were a nightmare to access (because I did it in the wrong order) I was able to remove the gantry, a few more screws and the white plastic part also came off, this part is shaped to go through a slot in the back of the machine formed from two pieces of plastic cable braid that act as a curtain and prevent powder from going into the back of the machine where the electrics reside.

And the underside complete with caked on powder. The whole lot will get a good wash and the bearings an oil. Should look good as new.

O-boy

It was a long shot and one which I would normally expect to fail. I took a trip to B&Q to see if they had any suitable O-rings. To my surprise and delight they had quite a few destined for plumbing applications. Plumbing is a bit old fashioned I suppose so the thickness's were imperial. 1/16th seemed to match and even the diameters seemed like they might work. I basically bought everything and spares, a cost only a fraction of the proper O-ring kit and even less than a new pump (400usd from zcorp).

But did they fit I hear you ask? Well yes, in fact the 8 piece 1/6th mixed o-ring set by Plumb sure had enough rings for two pumps. The pumps contain three O-rings, one on the non return valve which seemed fine, one on the piston (which was as loose as can be) and another that seals the assembly. I replaced the piston O-ring and it made a good seal but could still move in the groove to provide a valve action and it then started pumping but it also leaked so I replaced the body O-ring and all seems good. Fluid is flowing into the cartridge now, I am almost ready to print. Decided to try and clean a bit, may yet regret that.

Getting blood from a stone

Well not quite but getting the binder from its bottle at the bottom of the machine all the way to the print head is proving a challenge. I've flushed most of the lines and found a blockage at the input to a filter and now it gets all the way to its little header tank but it shows no sign of going any further. The primer pump is not working, this is a solenoid operated metering pump by Valcor scientific (an SV560) I stripped it down and I think the o-rings are past it. Nothing in my kits will fit nor can be made to fit but looking on the yahoo group reveals a kit from the pump manufacturer is only 33usd, that's OK but it might take a while to come :( boo hoo but a chance to have a clean I guess.

So far so good

I picked up the Zcorp this morning, a hot drive to Stoke and back and an even hotter struggle with the machine out of the transit van I hired and up the drive. The seller was extremely helpful and loaded me with lots of extras I did not expect like bags for the powder vacuum system, powder and even a cool old school American vacuum cleaner to use when cleaning the machine and another few dozen bags for that too. He also showed me how to make my own print heads from an off the shelf Canon one, a plastic vial and some epoxy. He told me the problem with the machine was that it did not power up, this is normally a nice kind of problem to solve because it is probably some element of the power supply. He had tried a few things (mainly replacing everything he had a spare for) and suggested a few more things to try.

When I took the back of the Zcorp I was a little surprised by how much electronics actually faced me. I guess I forgot that printing is a fairly complicated task made to look easy by custom electronics normally, it also has a lot of gizmos to control and was built in the 90's. I checked the transformer that he had just replaced and it was working and there was power to the PC power supply that runs a standard PC motherboard but it was not turning on. He suggested it might be the power switch but that worked fine too. I ended up taking a large ISA board out of the motherboard so I could get access to the pins that normally connect to the controls on the front of a normal PC. These had a little interface board plugged in which I also removed. Shorting the relevant pins (after finding the manual for the MB online) caused it to spring in to life, the PC part at least.

But why wasn't the switch doing its thing? The little interface board was connected by a ribbon cable to another board that seems to manage the power supplies and connects to the power switch. But the connection to this board was odd, there was one less pin on the board than there was holes on the ribbon connector, you could therefore put it in two possible positions. Some continuity testing made me think it had been put in the wrong position and I was right. Moving it by a hole connected the power switch to the same pins I shorted. Now it booted, did a funny dance with the bins and started beeping. Enough for one night I think!

It is pretty odd, the pins on the right are the ones the ribbon connects to, the pins in the middle are not even used. Not very clear at all.

I love the smell of ratchet straps in the morning (thank got for Lidl and there regular sales of them)

A small world

I love the internet. I posted to the diy_3d_printing_and_fabrication Yahoo group about my recent Zcorp acquisition and was almost immediately offered help to get it working plus a link to another Yahoo group about this exact printer the zcorp402. I was just accepted as a member and now have the service manuals which I found in their files section. Service manuals are like the holy grail when it comes to repair as you would expect.

New toy: Zcorp 3D printer

I'm wondering if my old Ebay addiction is coming back to haunt me. I just won this little beauty, a Zcorp 3D printer. It is based on inkjetting a binder into starch powder to build up a 3D model. I actually started trying to make one of these myself once: my attempts at a 3D printer but lost momentum. This one is not working but at £235 it is worth it just for the mechanics, if I get it working then that is great, otherwise I can probably put my own electronics on board. Printing in starch/plaster powder is not a patch on the resin based printer I just supported on Kickstarter except that the parts are perfectly supported by the powder. This should allow me to print CT data of bees that I collected some years ago directly. That was the original idea behind building my own actually.

There are two vats, one the build and one the powder, the gantry has a print head and a powder spreader. The powder vat piston goes up and the spreaders spreads the powder onto the build area, then it prints the binder and the build piston lowers and the process repeats. Simples!