Martin_H's Tantillus build.

I know I am late to the party, but for the past week I've started collecting parts for a Tantillus build. Two trips to the hardware store and scavenging parts from old scanners has yielded most of the mechanical parts. The printed parts and laser cut plywood were also pretty easy to find.

Here's a video from my YouTube channel: [youtu.be]

I still need to purchase the electronics, hot end, fishing line, and steppers. For the electronics I was thinking about using a Sanguinololu with a serial LCD to simplify wiring. I know this will require code changes to the firmware, but I've done a lot of Arduino programing so that won't be an issue. Buying steppers is usually pretty straightforward, and in this case I just need to be sure they're not too long.

Shopping for hot ends looks pretty complicated, so I am going to look before I leap to understand what's available.

My goal in building this printer is to aid my robot building hobby. In particular building custom gears and brackets.

Edited 2 time(s). Last edit at 11/22/2015 02:40PM by Martin_H.

Yesterday I cut all the rods to length using a Dremel with a cut off wheel, and ground the end flat against a belt sander. It was fairly easy to get a clean cut once I got the technique mastered. For the X, Y, and Z rods I used 8 mm smooth rod from an old scanner. It's straight, finely honed, and should work well. For the drive rods I used 5/16 drill rod from the hardware store. It's not quite as smooth or as straight as the scanner rod.

Today I test fit them in the bearings in the frame. When I spun the I could visually see a small amount of runout. To attempt to measure it I used dial calipers to measure the rod diameter when stopped, and when spinning. The difference in diameter was as low as 0.02 mm for the straightest rod, and as high as 0.06 mm for the worst. The printed bushings seem to have more play than the runout, so I am hoping this amount is within tolerance. I plan to drill the holes next.

I think you'll be very happy with the finished machine. I never measured the runout on my rods...though I'd have to think about whether you'll see any significant artifacts in the prints. If you're willing to wait a couple days for the delivery, Enco has the 5/16" drill rod on sale this month for $4.38 apiece (3 foot length). You can always replace them later if needed, but you'll end up restringing each axis again.

I think you want the printed bushings quite tight at first. On my first machine, I had to print them a number of times before I got a good, snug fit. Later versions were much easier to print, as printing on another Tantillus machine makes a very precise bearing.

98% of what I needed to know was in this forum...but it seems as if you've got it under control. Keep us updated.

Thanks.

I'm at a bit of a stopping point because I need the braided line and an extruder to complete the X, Y axis. So I did a few minor tasks like installing nuts and set screws in the pullies, test fit them, and cleaning out the support material from the extruder. The last task was a bit unnerving, but I watched the animation of the extruder assembly on the web archive and it made sense as I went along.

Meanwhile I ordered the spectra fishing line from Amazon, and emailed Colin at Maker Farm to learn about hot end options. I ended up ordering an E3Dlite Bowden kit which contains all the required parts. It sounds like a decent hot end at an attractive price, but it uses 1.75 mm filament. I know that Sublime designed the Tantillus for 3 mm, so I am departing a bit from the specs. I figure I can hob the extruder bolt with a slightly narrower channel and the extruder would be compatible with narrow filament. The kit comes with the PTFE tubing, but it might have a narrow OD than tubing for 3 mm, this might effect how well the nuts grab and retain the tubing, but this seems solvable.

One interesting tidbit is that Colin prefers direct drive extruders which is why he is uses them on the kits he sells. But he felt that 1.75 mm would work well in the Tantillus as it works fine in other Bowden printers. Overall I can see the tradeoff in Bowden versus direct drive. A Bowden reduces the mass you are moving around so you print faster, but the tube likely introduces some play and lag in extrusion and retraction. A direct drive has a tight coupling but higher mass. Basically a comprise either way.

Even though I don't have the fishing line, I decided to test fit all the X, Y axis parts, and couldn't get them to fit together. I noticed that the centers of the X and Y bearings on the carriage were a few millimeters father apart than the rod centers. I re-read the instructions and saw that the carriage middle is supposed to be 3mm thick (equal to two #8 washers). But my plywood carriage middle was 5mm plywood, just like all the other laser cut parts. The result is that the carriage is too thick. With that part as a template, I used my scroll saw to cut another from 3 mm hobby ply, and everything fit like a glove. I swear I could almost cut the whole case with that thing as it is so fast and accurate.

I wrote the US based person who laser cut them for me (they're not on this forum), that I was fairly sure this part was defective, and they might want to recheck it before cutting another set for someone else. I sent them links to the git hub repo and instructions on the Web archive for them to look at. I can see how this could happen because all the other parts are supposed to be 5 mm plywood.

I drilled a test hole through some scrap rod to learn the technique. Since I have a drill press and metal working vice this turned out to be straightforward. I used a center punch to start the hole, put a drop of oil to lubricate the bit, and slowly drilled into the rod. I paused a few times to get all the chips out because recutting them creates heat and is bad for the bit. So here's hoping that I can repeat my performance four times on the actual drive rods.

Finally I partially assembled the Z axis, and while doing so noticed that the holes in the lower bed are off center. Looking at the X, Y stage it seems clear that the work envelope is shifted towards the left hand front corner, and I should mount the larger portion that direction.

So far so good, but I imagine harder bits are yet to come.

I took a day off for Thanksgiving, but today the fishing line arrived, so I decided to drill the drive rods holes. To do it I measured and marked with a sharpie at 22 mm from one end and 42 mm from the other. I then clamped the rod to the workbench, put my center punch on top of the sharpie mark, and punched the rod in both locations. The drilling was straightforward and fast because the 1/16th inch bit was brand new, and I repeated the steps for the second rod. This left a burr on each exit hole, and the entry holes were pretty sharp as well. So I put a counter sink bit into my drill press, and beveled the edges of each hole to avoid cutting the line. I reinstalled the rods and the holes lined up perfectly with the X, Y end stops. One minor detail is that the holes are not completely coplanar which would look nicer, but that shouldn't be an issue for performance. So when the hot end arrives I will be ready to string the X, Y axis.

This took maybe fourty minutes and I still felt like doing a bit more work, so I decided to work on the hobbed bolt. The first issue is that my local hardware store only had 5/16th bolts in 4 and 4.5 inches, not the required 4.25. I bought the 4.5 inches because I figured I could make it work. I can certainly cut the end off the bolt to make it the right length, but the threads stop a mm short and when installed there's play in the bolt. To make it tight I just need a small washer on the end of the bolt to act as a spacer, but the washers with a 5/16th hole were too big and rubbed against the fixed part of the bearing. No problem, I'll solve this by taking a smaller washer and drilling out the hole to 5/16ths.

However the washer proved difficult to hold securely for drilling. My metal working vice only held it at two contact points and couldn't stop it from spinning. The edge also got a bit burred up from the rubbing. I then tried clamping it to a piece of wood, but again it wasn't secure, slipped, and spun free. I was getting frustrated when the obvious solution of vice grips came to me. The curved jaws trapped the washer at many points and drilling was easy. I wish I thought of that first. I know from personal experience that a burred edge on a piece of metal is not safe as it can be as sharp as a razor, so I deburred the washer.

I put the extruder together and the washer removed all play from the bolt. Hobbing it is going to be a learning experience, but I got started by marking the bolt where the extruder intake aligned with it using a sharpie. I took the bolt out and measured it at 14 mm from the head which matches the measured drawing on github. I put it in the drill press chuck, turned it on, and used a rat tail file to make really slow progress. I decided to knock off work after it was 1/2 mm deep and I have quite a bit more to go.

I decided to do two simple tasks to warm up. First, I cut the 1/4" threaded rod to the 170 mm length using my Dremel with a cutoff wheel, and sanded the end flat using a belt sander. Second, I cut the extruder bolt to length as well, and sanded its end flat also. While I've gotten pretty good at aiming the shower of sparks away from me, I noticed that I still had a fine layer of grit on my face from some backwash. Good thing I wear safety goggles.

Now I decided to tackle the harder task of hobbing the extruder bolt. I had already watched this video several times:

[youtu.be]

I put the bolt back in the drill press and used a rat tail file to cut an indentation 3 mm wide by 1.5 mm deep. I cooled off the bolt, put an M4 tap in the chuck, put the bolt in some bearing, and started cutting. The bolt popped out of the bearings and I realized it was no accident that he had the bolt head in the right hand bearing. I repositioned the bolt, and once the tap started cutting the bolt began rotating. I paused a few times to clean out the cuttings, measure the indentation, and add more cutting oil.

The video didn't have any measurements of the size of the hob, but some things I read online said that having the threads extend on the sides and bottom were needed to give gripping power. When I reached this point the indentation was 4 mm wide and about 1.65 mm deep. Touching the threads with my finger nail indicated that they were pretty sharp, so hopefully it will work, but I don't have a way to test it at the present time.

On the Bowden Tube, without causing bad angles, try to get the tube as short as possible. If the tube is short and doesn't twist up, then you really have no problems with Bowden tube extruders. On my Funbots that have a Bowden Tube length between 300-400mm I use a retraction setting of 2.5mm with good results.

Thanks for the suggestion.

Yesterday I put the anti-backlash spring into the into the Z axis. Minor difficulty in that the spring I bought was too long, so I cut it in half. Even though its ID was 7.1 mm it still was binding against the threaded rod. So I opened the coils up a bit and everything seems to work smoothly now.

Today the hot end came and I've hit the first step that frankly seems im-freakin-possible. The e3d hot end require that you use fiberglass sleeving to insulate the thermistor and insert it into a ludicrously small hole as described here:
[wiki.e3d-online.com]

Well if you breath too hard that glass starts to unravel and become impossible to insert into the hole. Moreover one side or the other catches and wrecks the end every time I've tried. I've swapped sides and trimmed the sleeving three times now, so I'm down to 32 mm from the original 35. If the hole was even a 1/2 mm larger I could do it, but I just don't have a clue. If you drop that thermistor it looks like it is all over.

Four times now and over a half hour of trying.

Edited 1 time(s). Last edit at 11/30/2015 08:47PM by Martin_H.

It also looks like the e3d's heat sink won't fit into the carriage bottom. So it looks like the hot end and X, Y carriage are incompatible.

I emailed Colin at Maker Farm, and although his policy is that you can only return unassembled items, but cutting the sleeving was small enough that he might make an exception. But he suggested using this X, Y carriage instead: [www.thingiverse.com]

Here's one that looks even better: [www.thingiverse.com]

Which is a good idea, but I don't have access to a 3D printer to make a new part.

Update: Here's a photo showing the overlap of the fins and the printed part. [ppl.ug]

Edited 4 time(s). Last edit at 12/01/2015 12:19AM by Martin_H.

After sleeping on things I realized that the only problematic part is the bottom of the X, Y carriage, specifically the lump near the screw. It's job is to hold the linear bearings perpendicular to the top pair, as the spacing between the bearings is achieved by the plywood retainer. Such a piece seems fairly easy to hand fabricate, so I am using brass tubing and JB weld to make a part, and see what happens. It only needs to last long enough to print a replacement which should be achievable. Some sanding and paint may even make it presentable.

As for the thermistor. I wrote E3D and their feedback was along the lines of "yes it's hard, but possible, but if you mess up we sell replacement sheathing. Please send us a picture of the heater block so we can see if it is within spec." That's all well and good, but shipping for 50 cents worth of sheathing is counterproductive. I think Amazon sell the same material with Prime shipping, so that's a possibility. In reading online about hot ends the thermistor on the E3D is problematic because it can move and short during printing. But other than that the hot end is well regarded.

Edited 1 time(s). Last edit at 12/01/2015 09:34PM by Martin_H.

I sent E3D a picture of the heater block and glass sleeving. It turns out that I have some of the old stock white glass sleeving and E3D has switched to a more rigid blue sleeving. They're mailing me a new thermistor and sleeving which should be easier to install. That was really good service, so I thanked them for it, and figured I should give them kudos here.

I completed my hand built lower carriage and test fit the E3D in it. Everything fits, so now I need to try it on the X, Y rods. It looks rough, but the brass tube and JB weld are probably stronger than the original 3D printed part. I built a gantry robot in the past using a similar technique, so I think it will work.

The steppers from tony34306 arrived and they look pretty solid. So the only part left to buy are the electronics.

Edited 1 time(s). Last edit at 12/02/2015 10:52PM by Martin_H.

My local hardware store didn't have 4-40 grub screws long enough, so I cut a slot in the top of some 4-40 threaded rod, cut it to length, and ground the end flat. Repeat to make a second grub screw. I then ground a small flat spot on opposite sides of one stepper shaft, put the nuts in the gear, grub screws in the nuts, and tightened the gear on the shaft with a screwdriver. The stepper mounted on the bottom of the extruder with the m3 screws which completed the extruder assembly.

I hand turned the stepper and the bolt and idler turn, so the mechanism seems to be working. There's mild binding in part of the rotation, but I suspect this is due to minor irregularities in the gears,and that will go away after a short run-in period. I should be able to mount the rest of the steppers either today or tomorrow, and I imagine that soon all the mechanical parts will be finished. But that leaves wiring, dressing the wiring, the electronics, and firmware. None of which are small challenges.

The electronics are an area where I am conflicted. An Arduino Mega, RAMPS, and Pololu drivers are fairly expensive, while Chinese electronics are considerably cheaper. But you get what you pay for, and frankly I don't need to debug flaky boards. A Sanguinololu PCB can be purchased quite cheaply and Digikey should have all the parts. That way could avoid poor quality parts and get a good build quality. Fabrication of a Gen 7 PCB and a Digikey order is also under consideration.

This morning I mounted the extruder, Z axis, Z axis stepper, and shaft coupler. It all went smoothly and this completed mounting all mechanical parts and the printer is deceptively heavy for its size. Next I need to string and tension the drive cables and align the drive gears. That seems like a small but tedious task, and might have been easier before mounting everything else. But if I have the build platform at the bottom and reach in through the sides and front it should be doable.

Attaching the Bowden tube to the extruder will be another area of challenge because the original design uses a #10 nut which is a bit too big for a 1.75 mm tube. The E3D comes with plastic bushings that grip the tubing, so we'll see if one will work with this extruder. If it doesn't a bit of improvisation will be required to secure that end of the tubing.

Edited 1 time(s). Last edit at 12/06/2015 09:51AM by Martin_H.

I purchased a Sanguinololu PCB because I know I can source genuine components from Digikey to avoid ersatz capacitors and IC's. Since I don't plan to use the USB feature, or the heated bed MOSFET, I will omit those components. I plan to use 0.1" pitch screw terminals rather than headers for the stepper connectors. I also have a ton of headers, shunts, 1k, 4.7k, 10k, and 0.1 uF components in inventory.

One comment on the 'tedious task' of stringing the cables... It's much easier to get access to the interior when the extruder and/or the Z carriage is removed. Both are relatively simple to remove/install as modules since they bolt to the case. Stringing isn't really that bad, if you get a couple long nosed tools to help. I used a couple medical forceps which really made it easy. Each loose end could simply be clamped in the forceps and hung over the side of the case.

I'm sure anyone reading the thread would enjoy a photo or two when convenient.

Thanks for the suggestion. The light in my work room isn't great, but here are some pictures.

From the front showing some parts unused up to now:



A side shot showing the extruder:



An above shot showing the drive rods and X, Y carriage:



Edited 1 time(s). Last edit at 12/08/2015 06:12PM by Martin_H.

That looks really good...neat and tidy. Since it's not a printed case, stringing the axes should be significantly easier. Thanks for the photos!!

Love the wood finish of the case.

Thanks. I used plywood because in my experience it is more durable than acrylic, and I liked the burnished look of the laser cutting, so I decided to highlight it with varnish rather than hide it under paint or leave it natural.

I strung the X axis and I think I made every mistake possible and a few impossible ones as well! The obvious things to watch out for are making sure both cables try move the axis the same way, but I still messed it up. You also want to make sure that you have enough wraps around the drive bar so that the carriage can move the full amount, doh. You also want them to wrap around the idler bar so they try to move it the same way.

I also started tightening the bolt and it didn't go into the nut, so I was spreading the X axis bar end rather than tightening it. I didn't notice at first and the plastic at then end got a little white with stress before I stopped. So first thing I will do is print a spare in case it breaks.

Update: Looking at the X axis I didn't like the way it was strung, so I redid it better. It was quicker the second time, so I did one side of the Y axis. Having the machine inverted with the Z axis at the top made it fairly easy. I had stop because it was time to make dinner, but I will pick it up later.

Edited 1 time(s). Last edit at 12/13/2015 07:10PM by Martin_H.

I've never tried to string the printer inverted. I'm surprised you've got enough access through the case holes - and wouldn't it be impossible to string the back X-axis rod?

Quote
robster34
I've never tried to string the printer inverted. I'm surprised you've got enough access through the case holes - and wouldn't it be impossible to string the back X-axis rod?

I used a pair of forceps to reach in from the side and wrap it around the back rod. Once I wrapped one end I used tape to secure the end and work on the other. It was a little tricky, but not that bad.

The reason I inverted the printer is it seemed like the easiest way to see what was going on with the inverted rod end stops. I could see the fishing line threading through the holes, and gravity works in your favor holding the washer and bolt in place while tightening with an Allen wrench. It just seemed the natural way to approach the problem.

During the tightening operations I found I needed better access to the rear X bar at all positions rather than just one. So I removed the Z axis and that gave me unobstructed access. I've followed the instructions and the cables are looking snug, and the drive rods can move the head anywhere in a plane. But the X and Y axis are not quite square. The X is about 2 mm closer to the case on one side than the other, while the Y is about 1 mm out of square. For small object a 1-2% variation won't be noticed, but this isn't something that should be left unaddressed

So I will need to read the build instructions on squaring to see if there are any tips.

By working the axes back and forth with repeated tightening, I was able to even the tension and square the axis to within a half mm. I then moved the steppers into position and tightened up the gear set screws. The X axis is very fluid with no binding, but I have some mild sticking on the Y at a few points in the rotation. I imagine that this is caused by mild concentricity errors in the gears and should run in fairly quickly. I started reinstalling the Z axis, but decided to knock off for the night after getting halfway through.

My Sanguinololu PCB arrived today, so I should be able to start assembly of the electronics by the weekend. I don't have all the parts, but Digikey is usually prompt.

Edited 1 time(s). Last edit at 12/15/2015 09:50PM by Martin_H.

The holidays are slowing me down as I've had to do shopping, attend two holiday parties, and my son's holiday concert. But I found some time to complete installing the Z axis, and now the electronics are required to test out the steppers. So that's the next thing to work on.

You're killing me. Your step-by-step reporting has inspired (forced) me to assemble another Tantillus out of parts I have on hand. I'm sure it'll be months before it's running, though...

Quote
robster34
You're killing me. Your step-by-step reporting has inspired (forced) me to assemble another Tantillus out of parts I have on hand. I'm sure it'll be months before it's running, though...

Great! I'm posting to keep myself motivated as this is an involved project, but it is good to hear I motivated someone else too.

Today I found an hour to work on the electronics. When soldering PCBs I always check twice and solder once. So I had a picture of the assembled Sanguinololu and the instructions from the WIKI. Before installing a component i would read the instructions and check the picture. The passive components (e.g. resistors, headers, sockets, and 0.1 if caps) are all pretty goof proof, but components with an orientation (e.g. LED's, electrolytic capacitors, transistors) all merit extra scrutiny.

With luck I should be able to burn a boot loader on the ATMega1248p before the end of the long weekend. That would allow me to see if the steppers will move things around as expected.

I soldered all the parts i ordered form Digikey on the board, then washed it with isopropyl alcohol to clean off the flux residue. I looked at the bottom with a magnifying glass to look for cold joints and bridges, and it looks pretty good. Before I install the ATMega1248p I will check all the power and ground points to make sure the board is good. I'm using a salvaged 400 watt Dell desktop power supply, so I plugged the ATX four pin power to check the voltages. First problem is that there's no 5 volt power. My mistake, I thought the four pin ATX power header included 5 volt and 12 volt power. It's only 12 volt, so I'll need to order a voltage regulator for the board. In the mean time I can power the board from my FTDI cable when burning the bolt loader and uploading the firmware. I have a Pololu AVR programmer, so I plan to follow the Gen7 instructions to burn a bootloader shortly.

I knew the steppers I ordered from tony34306 on ebay had six wires, and his description claims the common coil pins are 1 and 6. I figured I could use the DMM to check resistance to separate the leads for each coil and verify the common pins. Well separating the leads into two groups was easy, but the resistance between all leads with a coil is too close to be useful. If I had and inductance meter or oscilloscope I could measure it, but I don't have either, so I wil have to trust tony34306.

Update: I had a nagging feeling that if I looked hard enough I would find an lm7805 and 1uF capacitor lurking in my spare parts box. Sure enough I found both and installed them. I powered up the board and there was no smoke! I then measured voltages in key places and found 12 volts where I expected it, and 5 volts where it should be. So I installed the ATMega1248p and I am ready to try burning the bootloader.

Edited 2 time(s). Last edit at 12/27/2015 07:33PM by Martin_H.