- Slitting saws.
- The makespace slitting saw.
- The bandsaw I used instead.
- All Capulets
Makespace needs better slitting saws. But I avoided using a hacksaw, so at least all of the blocks look the same. They are not as neat as I wanted.
I used the four jaw chuck on the lathe to put a 15mm hole into four blocks, and then I gave those shoulders so I could bolt them onto lengths of extrusion. Those give me a connection that goes from square to round. However, they are not a perfect fit on the round bearings. I think they were all the same size when I made them, but thermal expansion made the aluminium bigger. That’s how fine the differences are. I took them all to the same size. Maybe it’s tiny ridges inside, but I’m sure the pieces were larger when they were hot and inside the lathe, clamps not withstanding.
The shouldering operation gave me a hat shape with an offset circle inside the top part. I drilled through the shoulders and the widest remaining part of the big bit of the hat. I cut through into that, which is the part that should have had a slitting saw, from the side. So then my circle had a little outlet where metal had been removed. I tapped half of that, so I could put a bolt through and cinch it closed. It also meant that the pieces would open up wide enough that even the tight ones would admit the bearings. Then I put the bearings in and bolted them closed and put them onto the bearing rods. Then I ran them up and down and smiled a hell of a lot.
I might end up opening the drilled parts of the shoulders. Right now they are tapped, which means the bolts will tighten into them. If I drilled them out, bolts would go through them instead, and tighten in the extrusion. I’m not yet sure which way up they go. I don’t want to do extra work or waste what I’ve got, but if I have to turn the blocks upside down and/or drill them out, I will.
I’ve got a gantry, and some bearings; those need to be joined. The blocks need drilling, partly tapping, and to have the tender minstrations of a slitting saw applied. I’ve got a day to do that.
I need to assemble it all, including the blocks and a central lead screw. Then the gantry has to be attached to all of that. I don’t have a connector for the lead screw right now. I’ll look into oldham connectors just because I haven’t used those yet, but I’ve ordered another of the aluminium spirals. Last I looked, oldham connectors cost money. It would be fun to make one, though, so I can say I have.
The microswitches need to go onto the frame, and be wired in.
I need to 3D print cable lay, and work out where to put it, including platforms for it to run along if necessary.
My programmer squid (don’t ask) arrives tomorrow, and I need to get her up to speed. I fully expect I will then be left in her dust. I’m OK with that as long as she leaves a trail of comments.
The breadboarded electronics need to be veroboarded.
Today I bought a new HDMI screen and connected everything up. I now have a working Rasbian computer with wireless mouse and keyboard. I’m well chuffed.
I’d like to do a lot more on it, like wire in the USB sockets on the top of the case, which isn’t horribly difficult. Right now, though, ZipPi is a functional, dedicated computer. It may be a Raspberry Pi 3 soon.
ZipPi the Raspberry Pi computer now has a MOSFET and board, and uses two case LEDs that indicate power to the board, and power to the Pi. It boots up, and we have played with it, but it needs a dedicated screen, mouse and keyboard, and a wireless dongle. I might get something with a decent antenna, or I might just get the Pi Dongle again. Those have to be extended out the back so I can close up the aluminium case and have it done. The board also has to be mounted inside its allotted hidden dead space. It works, though. And nothing burned down.
Chunks of cooking chocolate tend to be too sweet to make good hot chocolate. But we can try!
Take a small handful of chocolate chips, and add about a third of a cup of milk. Microwave for a minute and stir until bored of stirring – about ten seconds. Add milk up to the top, and heat for another minute, maybe minute and a half. If necessary, wash the microwave turntable.
Squeeze in red-top Sriracha until it’s no longer far too sweet to drink. Wander around kitchen drinking and wincing, but unable to put down cup of burny hot happiness.
With the help of an engineer at makespace, I debugged the gantry. Running a multimeter across the wires to check the resistance of each motor coil made sure that they were not burned out, but then we got out the oscilloscope and tried to work out why the hell the driver was not driving.
Finally we spotted that the live jumper that should have been 11V was only about 2V by the VMOT pin. We whipped that out, put in a different jumper to some different spots on the breadboard, and got the X motor working again. Then the Z driver got put in (although not attached to the lead screw) and they both worked.
I know how to build the Y axis on the unsupported rod, by clamping X section aluminium in front of and behind each of the cylindrical bearings. That’s then two points of contact along each rod, and each of those points can move, so I lock them together with more X section and bolt the gantry to that. That’s the minimal build that my programming friend will need.
I’ve got end stop microswitches, which I’d like to get into place as soon as possible, but cable lay is probably more important. Still, the gantry head moves and the motor on it turns, and that’s nearly the Morse Code machine I’m after for this stage of the build.
Also, am chuffed; the guy who was helping me probably has a spare bench power supply to lend me, so I’m going to be able to do more at home. He wrote down the name of the motor driver I was using, so as well as solving a problem with me, he got something definite out of it. DRV8825, for the record.
I went back with the aluminium ending to my friend, who tried out the cane and decided it was too short. I smugged internally, very briefly, and then measured how much longer it needed to be. Then I ordered the brass.
I’ve cut this a little longer than he thought, because I can take it down pretty easily, and having a cane that is a touch too long isn’t a problem. However, there is a slight problem that the rubber end is starting to perish, so there may have to be some more repair work done.
The cane is now a lot heavier, but I didn’t want to thin down the extention too far. If it’s horribly heavy I can do that, but I rather like the swing of it, so I’m leaving it like this.
A few days ago I helped someone at makespace do some laser cutting. I saw a man looking at the laser cutter stock shelves, and he asked me about the stock and one thing led to another and once we managed to post bail, I explained that volunteers to buy stock and get their money back were what made Makespace run. So, he very kindly ordered what we needed as well as what he needed.
His name is Steven, and he’s a radio astronomer, and he was making the front of a clock. But, as he wasn’t yet trained on the laser cutter, he needed someone to do the standing there getting hypnotised by the moving red light. So I did that. He had the file ready, and I just plugged in the cutting speed and power, and pressed Start.
Red light travels least, as I recall dimly from some source or another, which may be why the lights here are red, or it might be that the red light has other properties in radio telescope land, like keeping night vision working, or it might just be that red lighting was cheap. But anyhow, this is now ‘at the Thorrowgood telescope at the Institute of Astronomy, helping astronomers find their stars’.
Sidereal time is to do with where the bit of the earth you are on is pointing, and nothing to do with whether you’re late for work.
About 1/4 of the balls escaped from one of my linear bearings. I put back the ones I could find, but they are .5mm across, maximum. Too small to measure.
The bearing carriages take M2.6 screws. I hadn’t been aware that it was a size, and indeed it seems to be a wart on the metric standard. Every other metric screw is proportional to other metric screws; as the width increases, the thread size, or pitch, increases, so you can’t tell from a drawing how big a screw is, just what the length to width ratio is.
And then there’s M2.6 – which has the same pitch as M2.5 – the smaller size will fit into the larger holes, just loosely. Nobody uses M2.6 except some model makers and some toy car manufacturers, but given the size of my amazingly dinky linear bearings, it makes a certain sort of sense. So, I sent off for the cheapest batch of M2.6 I could find, and at the same time sent off for locking nuts for the M2 that were holding my bearings in place. (Those should probably be M2.4, but I’m OK with looseness right now. I’ll rebuild that part later anyhow.)
Meanwhile, I decided to move the gantry around without taking it apart, because it’s getting complicated to undo everything and pack it away. I should at least have taken off the lead screw, because what I really had was a great big lever attached to a gantry built of parts held together more by hope and surface tension than by any actual force. So today’s lesson, children, is that locking nuts are your friend and leverage is not. Also, really tiny ball-bearings will stick to and drop off anything, as they desire. They don’t even physics.
The MOSFET I wanted to use doesn’t provide much current at 5V. I’d seen the graph but my eyes had passed over it until Mat read it off, and then I understood what it meant and why it was important. And, y’know, why I should pay attention to the fact that graph is an upward curve. More voltage means more current – and I want to run the Pi at 5V through the USB power supply. It’ll protect it best, and allows for most flexibility. .7A is not enough.
Mat gave me the right sort of MOSFET, although he did insist on breadboarding it himself. I say I’m just teaching silicon to respect me.
Learning out of this: graphs are probably important. I should look at them or something.