Velbon rail to Lego connection

[Dubbin]

This is a bit of a 'work in progress' :
The Velbon super mag rail moves roughly 4mm on one complete turn of the winder. The Winder knob has 6 faces ; turning the knob one face moves the slider 2/3mm and turning it half a face 1/3mm.
On my rail the 'half a face' turn is a bit iffy because the gears inside the main rail are a bit rough.
I think I have found a way to improve things using the cross rail which doesn't have the bevel gearing, by adding a Lego worm drive gearbox.

a] Tease off the rubber capping that it stuck on the knob (just work round edge slowly peeling the skirt back) .Revealing the hard plastic knob underneath. [The skirt will fit back on afterwards.] EDIT - A BETTER WAY is to cut off the end off the rubber capping to expose the hole in the moulding underneath.

b] In the centre of the hard knob moulding their is a hole. Lego make a 'stick grenade' shaped pin (part 32054) that can be made to fit very snuggly into this hole. The great thing about this pin is that it fits Lego axle shafting in its end and so can be connected to other Lego parts.
-I used some 2.79mm id 0.9mm wall silicone tubing over the pin to make it a good fit.
b2] If you dont have one of these 'stick grenade' parts ordinary lego shaft/axle can be packed out to fit into the hole (the hole is roughly 5.8mm diameter at the top tappering down to ~5.4mm at the bottom).

c] Drill a hole in the rubber cap and fit it back on.

d] Add a lego worm drive to the modified knob ,
To stop the worm drive box rotating is a bit tricky but a lego brick studs register quite nicely with a groove in the underside of the cross rail casting

d] Align the cross rail in the same direction as the main rail, the main rail is then used for coarse adjustment and the cross rail and worm can be used for fine movement

A Lego technic worm gearbox is 24:1 so one turn of input the shaft is 0.16mm and a quarter of a turn would be 0.04mm.


Here is the mark1

[ChrisR]

Ingenious! The knobs are a bit stiff though, if you do the plastic nuts up on the rails?

You probably already discovered that if you remove the silver grub screw the knob and threaded shaft comes out, giving access to a circular step on the back of the knob. It would then be fairly easy to mount a disc on that with degrees marked - or divide by 400 for 10 microns. I was thinking of about the diameter of the whole rail width so degrees would be just visible. It would hit your lego plate though.

You didn't say the knob comes off - or does it?

[Dubbin]

Yes, the rails do seem to stiffen a lot on mine, as the nuts are tightened.
I hadn't realised that the cross rail winder could be removed, thanks :-)
The knob on the main rail can be removed:
1] Unscrew the lid
2] Take the screws out of one end of the rails (arrowed below), they might need tapping with a hammer or freezing to break the thread lock glue, but this seems easier to break than the glue in the cross slide. Then wind the rail assembly out of the housing by turning the focussing knob.
3] Pull out a small plastic split locking collar (circled below) that fits around the shaft that goes from the knob to the bevel gear.
4] The knob shaft, which has a square cross section, can then pushed out of the centre of the bevel gear by applying pressure with a screwdriver. It all goes back together quite happily.




My main rail had a curious drag even when the nuts were totally slack, this seems to be down to the very thick grease: I cleaned this out and re lubricated with a Maplin's ptfe and silicone spray, the winder is much free-er BUT I can feel every tooth of the bevel gears now, so I don't really recommend it!
I am wondering about dispensing with the main rail and putting a quick release plate on top of the cross rail, to do the 3/8 to 1/4 conversion. They seem a bit pricey though

[ChrisR]

ptfe and silicone spray

Keep that stuff well away from your lenses. Fifty feet should be enough

Yes I had the same thought about using the short rail alone.
It looks as though the 3/8" screw could be removed if the "E" clip were taken off, then a 1/4" cap screw put in its place, but I tried . The 3/8" one doesn't come out unless you dismantle the whole thing.
For our mag-rail-less readers, here it is:
Step behind the knob, referred to in post above, green; leadscrew retaining screw, blue.

Dorsal


Ventral

[Dubbin]

Thanks for the warning about the spray -is it the silicone ?
I have managed to get the crossrail apart: The smooth slide rails (running bars) 'just' unscrew: I used the molegrip like pliers in the photo (magnusson viper grips?) to grip the head, they have much finer teeth than molegrips and if I made a mistake with them, it was not adjusting them tight enough initially (using the circled adjuster); the rail alloy can be torn easily.
The running bar's thread (left end, in the picture below) is loctited into the body, I tried some isopropyl alcohol and that didn't seem to do anything. I then rested the the rail vertically on a wooden table and hit the bar's head (the end with the two pips in) with a large hammer, once, which broke the thread glue and allowed the bar to be unscrewed.



The slider works perfectly when re assembled!

The 3/8 capscrew thread is 14mm long and the cap end is 14mm external diameter. the whole thing is 22mmm long.

[Dubbin]

Woops, I should have said that I gripped the head of the rail (with the two pips in) and not the rail itself, with the grippers! This has damaged the head cosmeticly.

[Dubbin]

ChrisR- Rather than using a disk with marks, how about a gear with a clicker?


The gear has 40 teeth and the spring came out of an old kitchen timer (it is trapped between the quick release plate and the cross rail)
It makes a very loud click! OK it's only a 0.1mm step, but it is simple :-)

[ChrisR]

So, it would be possible to replace the 3/8" screw with something like a 1/4 Whit caphead screw, with a 14mm diameter washer?

When I've used printed discs before, I've found that using eg
60mm diameter, One degree is just about OK to rotate reliably, as long as you do't change direction. That's around 4/360 = 11 microns. That would be about half a mm at the rim. If you search for such things here by Ed French "elf" you'll find some fancy Vernier ones.
You can draw them in Photoshop quite easily, though the numbering is harder.
I've used them on a Manfrotto 454, which has a leadscrew pitch of 1.2mm. (M6 thread)It's not a great rail though; the Velbon is somewhat better at what it tries to do.
I bought the Velbon rail franky because I had an Amazon voucher which was about to expire, and they were cheaper there than "Used" on Ebay.
I haven't used it in anger yet but expect it to be reasonable in terms of rigidity (until it wears/distorts) with magnifications up to about 10x, which ties up with a minimum step around that 11 microns.
A clicky wheel arrangement would speed things up perhaps, but have found it OK to do eyeballed rotations as long as there's a lever/handle on the knob. A clothes peg is just about right (The Manfrotto takes a hex wrench Allen™ key)

Actually the length of the Velbon cross-slide makes it suitable for conversion to a micrometer drive (push, against a strong spring). There are often cheap micrometer heads on ebay. Even cheap digital options.

I'm not clear which part was loctite-fitted? Do you mean the thread on the two running bars? That wouldn't hammer out!

Also I can't tell what the "nut" part of the leadscrew it, is the traveller threaded?

[Dubbin]

I think that 1/4BSW will fit in OK (with an O-ring on the protruding thread, to hold it in?). I have ordered some (£5 for 10- ebay).

Yes my post was not clear! :
The threading on the running bar is at the opposite end to the head with the two pips in (The nearest end to the lump hammer, in the photo above) The manufacturers have covered over the thread ends, after assembly, with glued on plastic disks.
I struck the head of the running bar only to break the thread glue (not to drive it out). A kinder way to break the glue's bond might be to cool and shrink the exposed running bar with water pipe freezer spray(?).

I now also realise that it would be better not to peel off the rubber winder caps because despite them going back on ok they now have a bit of a squidgey feel to them. - Simply cut off the extreme end of the knob cover with a stanley knife, to expose the end of the 'pipe' in the hard plastic the moulding, underneath.

I will investigate micrometer heads, thanks.

[Dubbin]

ChrisR, I hadn't twigged that ten microns is only 0.01mm, eeek! -I am still at the reversed wide angle lens stage
However, one turn on a double worm drive (576:1) ought to be about 0.007mm
This is what it looks like;
a] made using lego parts as lego intended.
b] If you are happy to glue the two worm drives together -much nicer:



There are about 10 turns of backlash in the gearing, but it feels quite nice:-)

[Dubbin]

A bit of an update;
The cut-down ¼ cap screw does work well and only falls out when the rail is wound right back; however a quick release plate with a 3/8 thread fitting underneath, might be handier.
The rail is now controlled by a Lego Mindstorms programmable brick (an ‘NXT’) using the motor that comes in the set. The motor is quite powerful and has an optical encoder on its output that can resolve to 1 degree.
One motor output of the NXT (it has three) is wired into a relay; when the relay is powered by the NXT its contacts close and join two wires of the shutter release cable to fire the shutter. Though as the NXT brick can make ‘beeps’, one can simply fire the shutter manually, when it beeps.

To mount the motor, the side of the rail was rubbed against a file to remove the black paint and to slightly flatten the metal and a beam with joining pins in it was epoxied to the side. I assembled the whole Lego contraption prior to gluing it on to make sure that everything roughly lined up.



Assembled (but with a slightly different style of worm gearbox):


The drive into cross rail’s winding knob is a Lego shaft with some silicone tube slid over it to make a tight fit into the hole in the end of the knob, however I notice that the silicone tube has a bit of give in it.

The brick is programmed using a very graphical ‘drag a block here’ software that comes with the Lego kit, the program is downloaded into the NXT and remains there even when it is switched off.
Here is an example that allows the step angle to be adjusted with the arrow buttons on the front of the NXT:
https://dl.dropboxusercontent.com/u/976 ... 20NXT2.rbt

The stepping sequence is started and stopped by bumping the orange button on the NXT. It beeps once for mirror lockup and again for the photograph.
Once the orange button has been pressed the second time (to end the stepping) the motor should rewind the rail back to its start position.

How bad is the positioning?
The worm gearbox has about a quarter of a turn of input-shaft backlash and the shape of the gear teeth in the worm gearbox mean that roughly every 7 degrees the angle of the winder leaps forward or falls back about 0.7 degrees (which on the Velbon rail is equivalent to ~8 microns). Not a problem at 2:1 but at higher magnifications not good.
However using two worm gear boxes driving the same shaft, with one set a quarter of a tooth off from the other eliminates most of the backlash and reduces the wandering.
This construction is pretty large though:



Here is a time lapse showing the stepping of a protractor stuck onto the double worm output shaft (one degree steps):
https://dl.dropboxusercontent.com/u/976 ... id1fps.wmv

The NXT motor itself usually moves to +/- 1 degree of its target, but if it overshoots and returns back to its mark then the backlash in the worm gear will make the stepping irregular. The ‘Motor Block’ in the Lego software gives options for control and the better option seems to be “Ramp up” (rather than “Continuous”) for very small steps. There is an alternative (free) PID Motor Block written by Hitechnic that seems better still in its positioning, even if it is a bit more violent in its movements. This control block has some interesting features; it can move to a physical stop and set its zero position and unlike the Lego block it doesn't reset its angle back to zero every time it is used, which should mean that if it didn't quite make its target position last time around it should catch up the next time.

If it can be made to work a little bit better it might be a useful at higher magnifications(?)

The relay:
(This works for the Canon 600d and I presume similar models).
The shutter relay is a 5 volt reed relay from Maplins (FX88V), but any five or six volt relay should do.
With this particular relay pins 2 and 6 (or 13 and 6) go to the NXT and pins 14 and 8 (or 1 and 7) are joined to the shutter release cable.
On a canon shutter cable (that has a jack plug) when the white wire is connected to the outer braid the shutter is triggered.
The NXT cable has white and black wires in it which are the motor power lines. These are connected to the coil so that when the ‘motor block’ is run in the program, the relay closes.
To work out which pin is which, the key thing is the dent in the end of the plastic moulding, I have marked it on the red outline below:



I put a diode across the relay coil to stop any back voltage that might be generated by the coil from damaging the NXT and then put a second diode on one wire to protect the first diode from being powered backwards and blown. It doesn’t matter on which coil pin they go as long as the diode ends with the black banding are joined together. These diodes might not be needed on a relay this small, but they are there as an insurance. If the relay doesn’t work, it could be that the direction of the motor in the program is set the wrong way round.
It is a tiny relay so I soldered mine onto a bit of circuit board.

[Dubbin]

The middle threaded rod of the rails can be removed and a finer pitch thread put in its place, I happened to have a 12 inch length of 2 BA threaded bar which has a pitch of ~0.8mm.

Main rail:
The conversion is rather as one might expect except that I used a clear ‘Bic’ biro tube pushed into the middle of the rail’s plastic bevel nut. ‘Bic’ biro tube has a gentle taper and wedges in well. 2BA nuts are then shrink wrapped onto the ends of the Biro tube.
It is not a brilliant conversion but the extra fineness helps for 2:1. Unfortunately there is a limit to how much grip a jammed-in biro tube has, so it might drift if it is used vertically.

Here is the Biro tube pushed into the spinning nut on the main rail:


The threaded rod is fed through the biro tube and nuts are tightened onto each end of it. One end is shrink wrapped, to trap the nut, then the tube is pushed further through and the other end is shrink wrapped:

2BA nuts fit quite well onto the rail’s endplates, without washers.

The cross rail can be converted with a biro tube, in the same way:


However the hole in the winder end of the frame is too large to hold 2ba nuts/washers so this hole was covered with a square Lego plate (arrowed, above) with its four corner pips filed back by about a third and a hole drilled into its middle. This was lightly stuck to the frame with two spots of ‘super glue’.
As the threaded rod in this rail rotates, it has 3 washers under its nuts at each end. The nuts have to be locked to the thread itself, rather than tightened to the frame.
For a new knob a bottle cap with a hole drilled in it can be locked on with a nut.
It should be possible to convert the rails back to their original thread, as very little is glued.

[pulsar123]

Sorry for reviving an old thread! I have questions about Velbon Mag Slider - can either of you (Dubbin or CrisR) comment?

I am in the process of building a fully automated macro rail rig based on Velbon Mag Slider + stepper motor under control of Arduino micro-controller.

The first challenge I am facing is how to reliably connect my stepper motor (which has a plain round metal shaft, 4mm diameter) to the focusing knob of the Velbon rail. I'd like to only rely on metal parts - some kind of a metal coupler which would connect to the motor on one side, and to the metal axis of the focusing knob on the other.

I read carefully this thread, but couldn't see if any of you removed the plastic knob from the metal axis on the slider. I can see that the axis has a square cross-section inside the rail, but it is not clear what is going on in the part of the axis protruding from the rail. (From my measurements, it protrudes by ~5mm - enough to attach a coupler.)

Ideally, I'd like to know if it is possible to remove the plastic knob from the metal axis without damaging much (or anything), and what is the cross-section of the part of the axis which is protruding from the rail. Is it square, or round?

CrisR made an interesting statement here (post #2): "You probably already discovered that if you remove the silver grub screw the knob and threaded shaft comes out, giving access to a circular step on the back of the knob." Can you please explain it in more detail? Where on the knob is the silver grub screw? I don't see any on my rail.

Thanks for any feedback!

[ChrisR]

The grub screw is the chrome one pictured about 9 posts up, arrowed blue.

I can't remember much about it now.
I do remember that the short rail is better than the long one, and that they're both rather tight if you do the nuts up to reduce the "play".
If you can parallel-mount the motor and used a toothed belt (or gears) you would get finer control and more torque.

[pulsar123]

The grub screw is the chrome one pictured about 9 posts up, arrowed blue.

I can't remember much about it now.
I do remember that the short rail is better than the long one, and that they're both rather tight if you do the nuts up to reduce the "play".
If you can parallel-mount the motor and used a toothed belt (or gears) you would get finer control and more torque.

Thanks! I see - your comment was about the shorter rail, right?

I'd really like to use the longer one if at all possible, to justify all the efforts (my most efforts will actually go towards programming Arduino; judging from the existing examples - see here http://www.ryleeisitt.ca/articles/build ... ontroller/ , this can be quite a task, with ~2000 lines of code).

I also prefer to directly connect the motor to the knob. I used this approach in my "dumb" motorized macro rail ( http://photography-on-the.net/forum/sho ... ?t=1415746 ), and it resulted in a very compact and sturdy construction, great for field work.

And I think with my setup I already have fine enough control: 1600 microsteps per one rotation, or 2.5 microns per microstep. It also allows me to rewind reasonably fast: I think I can go up to 100 rpm with my setup - 9 seconds to rewind from one end to another.

It sounds like I'll have to discover myself what is under the plastic knob - hopefully without rendering the whole rail useless. I'll probably do that over the weekend, and will report my results here.