sploo

A little more progress... After the previous torque test it dawned on me that the large (Nema 24) stepper motor in the original design is way over specified (I could never make use of the output torque). It turns out that a tiny stepper (commonly used on 3D printers) should be more than enough. This also made the interface plate between the stepper and the harmonic drive gearbox much simpler (previously the bolt holes overlapped, but were just a few mm different). The only problem is that the output shaft is too small for the harmonic drive's wave generator. So... a simple sleeve solved that: Next, the new design for holding the two parts of the friction drive together. The original driver shaft (from the harmonic drive) was 15mm at the interface, but 12mm for the bearings I had. Some new 15mm bearings meant I could make a new - much simpler - 15mm driver shaft, but also this required new bearing blocks: The two main parts now fit together by the use of four threaded rods. The bearing blocks on each end of each axis would be joined together to prevent twisting, but you can see the layout below. By torquing the nuts on the four rods the two axes are pulled together: With none of the planned support to hold the top pair of bearing blocks together, I got the friction drive to manage just under 10Nm before slipping (the photo below is about 7.6Nm): Unsurprisingly, attempts to torque down the nuts on the four threaded rods started to split the MDF I'd bodged together for the bearing blocks, so I couldn't match the previous 15Nm result. I think this design would work (and it's pretty compact and simple). I don't think that having to torque up four nuts is that much of a problem. However, the cost of steel, alum or even engineering plastic sheet of the required size for the bearing blocks turns out to be expensive. Some quality plywood (and making the blocks wider so the rod holes are further away from the ends) would likely take the load without splitting. However, my second design idea for the above was to just use the four threaded rods as alignment, and instead have a 'U' shaped hanger on the inside faces of the top pair of bearing blocks - the hanger would extend down past the smaller drive rod. At the bottom of the 'U' of the hanger there would be a screw with bearing, which could be torqued up in order to push the bearing into the smaller drive rod, thus lifting the whole bottom axle. That design would allow the use of a single adjustment screw (rather than four nuts) and the load of the hanger could be taken across a larger area of the top bearing blocks - meaning they could be made of a softer material (e.g. plywood). The hanger could be made of thinner strips of steel, which I have already. It's more complicated, but I think it's worth giving it a try... when I get time...

Great job.

I thought that light was made up of waves. I mean photons. Waves. Photons. I'll get my coat.

I think that could go deeply into an analysis of many a con-artist, but it would probably take us down a very dark rabbit hole, so I'll stop there

Well, yes, but... Astronomers spend hours standing outside in the dark/cold/wind, vainly hoping to see something that isn't cloud Astrologers spend minutes in the warmth, making up any old nonsense, and earn a good living from the gullible Conclusion: astrologers are smarter 😁

Back in the mists of time, when I were a young (and rather naive) young chap, I got confused between "astronomy" and "astrology" whilst talking with an astronomer. "what's the difference?", I asked "one of them will get you a punch in the mouth", came the reply Seems fair, in hindsight 😀

Updates... I mounted what was the lever in a wood "slider", and drilled a small indentation in the bottom to locate an M12 bolt (the end of the M12 bolt was turned on the lathe to have a point that roughly matches the indentation left by drilling the lever). The M12 bolt is secured in a chunk of wood that has a couple of M12 nuts embedded. Long term I'd do this differently/use metal, but it's just a test: Mounted back on the main body, to push the output of the harmonic drive onto the friction wheel: And with a lid, plus the "skin" covering the friction wheel: First test was promising; I very nearly got the two 5lb weights lifted, but couldn't quite get the lever all the way to the 9 o'clock position before the friction wheel slipped. Checking the side of the main body, it's clear the force of the bolt was bending the flimsy chipboard: Ideally, the top of the bolt cover would be tied to the friction wheel spindle with a single covering skin (under tension it'd stop the flex - basically like a torsion box). But this configuration is just a test, so instead let's clamp a couple of lengths of timber to keep it straight... and success: I checked the weights on a scale, and they come to a total of 4.61kg (just over 10lb). At a 33.3cm (0.333m) distance between the lifting point and the spindle, I make that: 4.61kg * 9.81m/s^2 = 45.22N 45.22N * 0.333m = 15.1Nm I'm calling that good enough. The lever (now a "slider") can be much more compact, and I think it should be possible to redesign the stepper + harmonic drive hanger to slide up and down too (rather than hinging). More when I next get chance to look at it...

I don't know; what camera are you using? All my recent experience (as in - the last decade) has been with Canon DSLRs so I'm not too familiar with other systems. If you can set the camera to manual exposure mode, say ISO 100 and 1/50th, then without a lens attached it should react (i.e. show some brightness) on the LCD screen when pointed at a light source. The same should happen if you then connect the T-ring. If that's OK then check you can release the shutter (i.e. take a shot) - it'll just be a blur of light, but proves the camera will fire with the T-ring on.

If you're getting a completely black image on the camera's liveview (regardless of exposure settings) even when the scope is pointed at a daylight sky then could you see what happens if you just connect the micro 4/3 T-ring to the camera (without any other parts)? It's possible that the camera's mount is trying to talk to the T-ring (as if it were a lens attached), and getting confused because it's not getting a response. I don't know what camera you're using there, but if you put just a teleconverter on a Canon DSLR camera then sometimes it's necessary to cover the pins, as otherwise the camera complains (because it's expecting to talk to a teleconverter + lens combo).

I'd assume (with only experience and layman's understanding, rather than any actual evidence) that the sensor temperature shouldn't make much of a difference with light flat frames. The signal-to-noise ratio in a flat light image should be very good; whereas dark and bias frames are capturing the behaviour of the sensor + electronics in terms of noise in the system (which will be much more temperature sensitive). That's not to say temperature won't affect a light flat frame, but I'd assume that only really extreme temperatures should cause problems.

I got a little time to remake the MDF lever in steel, as originally planned. With no welder (or welding skills) it'll be pinned together with 3mm machine screws: To fix the 12mm rod for the bearing, I drilled to 12mm on one side of the lever, then tapped the other side to an M12 thread. I put a thread on the end on some 12mm rod, and used a saw to create a slot at the other end: New lever test fit: As is so often the case, once you remove one weak point, you find the next 😉. While it was a little better than the previous version, the threaded rod on which the lever pivots is now deforming the hole in the plywood cover that provides support. I expect it's probably doing the same to the black chipboard too (though hidden behind the large washers). I could make supports using steel, but on inspection of the rod I think the forces are starting to bend it, so I don't think it's going to be a viable solution. To be fair - you don't actually need that much force to bend a length of M8 rod. It occurs to me that the lever could instead be mounted vertically, and slide upwards (pushed by a threaded rod). It's not the ideal size/shape in its current design, but I plan to look into that when I next get time.

Yep: https://en.wikipedia.org/wiki/Mars_Climate_Orbiter#Encounter_with_Mars I was working in the space industry at the time. The fact that non-SI units had been used anywhere (for that application) was just crazy.

Oh yeah - that's another good one; plus the differences in the hex sizes on BSW and BSF (whereby one spanner fits the other, but offset by one size).

Having rebuilt and restored numerous old bits of machinery, the above is particularly pertinent (and amusing). The old British Startrite table saws are great, but you'll find that the thread used for the rails on which the fence sits is a 1/2" 12 TPI (which was a curveball, as I'm more used to UNC... which just happens to be 13 TPI at that thread diameter - one of the few sizes where BSW and UNC differ). The side rails for the sliding table models are in metric, the grub screws that adjust the table insert plate are BA, and the threads in the fence are either BSW or metric, depending on the vintage. The next machine I rebuilt was an Austrian made planer. Everything was in metric. Oh, and don't even get me started on threads in Stanley hand planes... such as the not-available-anymore #12-20...

I've heard it claimed that the plastic gears in these little lathes is actually a "good" thing; in the sense that it's a cheaper weak link than having to replace the motor or control board; which I suppose is a fair point. Still this one didn't cost me much (second hand), and I couldn't have done what I've done without it so I'm not going to be too harsh on it.