Starting at the beginning with Mirror Screws

[gary]

How much more hair will I lose?
Steve O

Just be thankful that you still have some to lose... some of us don't.

[Panrock]

It's a good thing I've still got some hair in stock, because I've just tried a first run up of the 'new' 120-line mirror screw and it's horribly unbalanced - quite unlike before.

First thing... check there really are 120 slats there rather than 119 or 121 etc... yes, this is OK.

Next thing... grab all the slats bodily and squeeze (try) and pull them apart (only slight movement is possible) and try again... no real difference.

It's dangerous to run up to full speed in present condition so what I shall do next is feed in an unlocked square wave into the line-of-light and run the thing at vastly reduced speed. Then view the pattern and see what slats need tapping to or fro. Once I get a reasonably straight display... ergo... the screw should be balanced.

Read all about it, warts 'n all... only on this channel!

Steve O

[Panrock]

Getting there!

The reason for the imbalance was the grinding profile of the mirror surfaces of the slats, which is not quite parallel to the rear faces end-to-end. I have now cancelled out almost all of this with a counterweight (an ounce or two) fixed to the upper slat clamping disc. I'm confident I can remove the remainder of the imbalance soon. To make room for the counterweight, I've moved the frame sync triggering to the bottom clamping disc.

I have now seen my first 120-line colour test pattern, but at the moment it's pretty awful. There's still some tapping about of the slats required to provide a reasonably undistorted picture suitable for final correction with Karen's box.

I can use for this the special slat angle setting tool I adapted from a long jawed mole grip.

It looks like the picture is going to be bright enough... always a concern with the remote line of light and tiny scanning spot at 120-lines.

Now's the time to order that lenticular film Karen recommended for the lines of light.

Steve O

[Panrock]

Having "hacked the slats" to get a roughly correct overall positioning, the next step is to (try to) precisely set the angle of each one using a laser diode module reflecting off each adjacent pair of slats, with the beam travelling across the workshop. At the target end, there is a chart bearing two lines. The gap between these indicates 6 degrees of difference, which is what you get with laser light reflecting off two slats set at 3 degrees with respect to each other (120 slats, each displaced by 3 degrees = 360 degrees of course).

Since the two lines on the chart are over a foot apart, this should make quite a sensitive method of getting the settings right.

Doing the above has got me thinking.... would it be possible to shine the mirror screw image onto a screen by using a modulated vertical line laser? Or three (RGB) lasers for colour? A standard line-of-light won't do, because the light diverges sideways out of it.

You can convert a standard laser to a line by shining it through a cylindrical rod. Even at 120-lines, the projected picture could be quite large not too far away from the screw. Something to try in future years perhaps?

Steve O

[Dmitrij]

In the manufacture of screw I resaw machine held a 2 line difference in the 11 and a quarter degrees on the side of the stack of plates. At the turn of the screw I combined the first line on the plate with the second line on the next plate,it can be seen in the photo. Screw adjustment took 10 minutes. Sorry for my English.

[Panrock]

Dimitrij, your English is fine. It is better than my Russian. I did a Russian course here in England to GCSE level, which I greatly enjoyed. I used to visit Russia often. However i have not been back for 14 years now, and my Russian язык has declined.

The way you have set the angles on your mirror screw is logical and elegant.

In my case I am using index pins in holes cut using a laser cutter, operating from CAD (computer aided design) software. My problem is that, for a picture of a given shape, the tolerances at 120-lines are 16 times more critical than at 30-lines!

My current method is reasonably accurate but it is not accurate enough. So I now am reflecting a laser off each pair of slats and observing the spacing of the two spots on a chart the other side of the workshop. Then micro-adjust the angle.

In general, I am not satisfied with the standard of work I have achieved so far. So I have decided to strip it all down and start again!

When the mirror screw goes together again, there will be various improvements. One thing I will do is to micro-adjust each slat angle using the laser, one at a time, as I build up the screw.

Steve O

[Panrock]

And so I have started... it is a slow process and I will only be able to add and micro-position a few slats every day.

Presumably - because since grinding the slats are no longer exact rectangles, I have found I can't always set them correctly with the brass indexing pegs in position. So I am now positioning them (using the laser) sitting 'loose' and then pouring into the lined-up index holes "Chemical Metal" - a metallised hard plastic filler. After ten minutes, the slats are locked with 'virtual' indexing pegs, and are then sitting exactly at the correct mutual angles. The required tolerance is ±1 minute of arc!

Peter Yanczer used to make and sell Mirror Screws. I wonder what indexing method he used? His 120-line demonstration picture was pretty blurry, but then he didn't have the luxury (as I do) of an electronic 120-line source and had to use his own mechanical method. A true pioneer, and greatly missed.

Steve O

[AncientBrit]

Steve, I admire your dedication.
You've certainly set the bar high on this project.
As you said earlier in this thread use of Karen's hardware will ease the task slightly but even so you must be working at the limits of mechanical engineering.

Well done, I look forward to your posts,

Kind regards,

Graham

[Panrock]

Graham, thanks as ever for the positive vibes.

One further refinement that has occurred to me. I shall 'decouple' each group of 20 slats... meaning there will be six groups whose positions can be independently adjusted at the end. This should help correct any cumulative errors building up in each group, if such there be, caused by laser chart calibration errors.

Steve O

[Panrock]

Just to reassure all you blokes and blokettes out there that I haven't been twiddling my thumbs here since the last post, here's a blurry view (taken in the dark) of my laser aiming chart.

The two red spots have arrived there across the workshop after the beam from my 'house' laser diode has bounced off two adjacent slat edges together, which then splits the beam into two to go their separate ways - 6 degrees apart.

After carefully measuring the travel distance and doing a spot of trig', I was able to mark up the distance between the lines on the aiming chart to within a fraction of a millimetre, so I'm hopeful this will make a sensitive and accurate way of setting up the (3 degree) slat angles.

The "proof will be in the pudding" though. I won't have finished setting all the slats until mid-December.

I had a nasty moment last night lying in bed when suddenly the horrible thought hit me... "will the bounce difference angle be 6 degrees irrespective of the angle that the laser beam is incoming to the slats?" After getting up, and some fiddling with pen and paper, I came to the conclusion that "yes, it would". Phew! Otherwise I'd have had to strip everything down for a third time and start again...

No, I still don't really understand mirror screws!

Steve O

[Panrock]

The long-winded laser setting and locking compound pouring process (at just three slats a day) proved to be forty wasted days! There was a standard error in all the settings. I still don't know why. The end result was the mirror screw had to be taken to pieces once again.

At my wit's end, I then reassembled the screw over two days without any indexing at all. The slats were merely positioned roughly and the stack was clamped end-to-end. The whole assembly was then spun up and angular corrections made 'to order' with the setting tool I had made previously - an adapted long nose mole grip with a locknut added, intended to 'click' the slat edges into their correct positions. Since their grinding and polishing, the slats are now slightly tapered in shape, which further complicated the process.

Anyway, after a fair bit of fiddling around and hours of running the screw (which seems to further improve the positions), the slat positions became not too bad. The last thing was then to connect up Karen's brilliant Timing Corrector box and spend an hour or two adjusting the final position of each line. Karen's box generates its own test signal for this process.

SUCCESS! I spun the screw up this afternoon and have watched extracts from the 1946 technicolor musical "Till The Clouds Roll By". The pictures still require slight improvement but already they are great to watch... lineless, sharp, contrasty, colourful and 'limpidly' clear. The only thing that spoiled it was failure of the sound channel. To be fixed tomorrow.

Going to 120-lines has definitely been worth it! The 12-inch colour picture can be viewed like any 'normal' television. It's good enough to satisfy non-enthusiast viewers. It's far better than its 60-line predecessor.

I aim to bring this to the Convention, but because the facilities tend to be cramped for a display that will require more than four metres of clear floor space, I'll have to ask if somewhere special can be found where the it can stand.

Karen has put in a lot of work - and ingenuity - into her TC 'box', all done for nothing in return, and it has proved essential.

This achievement of 120-line colour mechanical television should therefore be seen as half Karen's - half mine - a combined effort. Oh, and we mustn't forget Steve A, whose original electronic design still lurks within... and seems well up to the job of providing the necessary bandwidth.

Steve O

[Harry Dalek]

Hi Steve sounds good any chance of a screen shot from an up coming test /

[Panrock]

Harry - good on you mate and thanks. BTW Happy New Year to all here.

Yes, I fully intend to get some screenshots up - and maybe a video too (if I can lock the frame bar). Only one problem...

Yesterday I was enjoying watching cartoons on it, followed by Esther Williams swimming around in about 1945, in "pin-sharp definition" and "living, vibrant color" - with sound too - then it all went 'phut'.

Well actually not the bits I made, but the WC-01 converter lost the ability to lock its output to the frame pulses. This is the second time it has done this! The first time involved me taking a 100-mile cross-country trip (yes, England is smaller than Australia) to an electronics firm who could fit the replacement surface-mount opamp required. Too fiddly for me to do. All very kindly organised by ppppenguin, who is also now a recent member here, I understand.

So now there is no sync lock. No mirror screw watching either, until I get the WC-01 back, which this time I expect to send back to the USA.

C'est la vie....

Steve O

[Steve Anderson]

Oh, and we mustn't forget Steve A, whose original electronic design still lurks within... and seems well up to the job of providing the necessary bandwidth.

Steve O

Steve, as I recall the active stages have a bandwidth in the region of 300-350kHz. But long cables to the LEDs may reduce this. By long I mean over 10 metres. I did do some tests with a 10m multicore cable and it really didn't make that much difference. Should 10m plus be required I suggest using CAT5 or CAT6 network cable which is easy and cheap to obtain. Use the pairs as pairs if you see what I mean. There's four pairs in the cable, you'll only be needing three I presume.

Steve A.

[Panrock]

Hi Steve,

It's been such a long time since I made up the harness and I've now forgotten... but the cable on there already looks like it could be CAT5. Indeed, I think you may have originally specified this.

Anyway, the length is much less than 10 metres and the picture is very sharp, more constrained by LED width than by the bandwidth of your excellent circuit.

Steve O

PS. In connection with the LED line-of-lights, I'm using some special micro-lenticular film at the front (suggested by Karen), which merges the LED images into continuous thin lines while not diffusing them in the other direction. Splendid stuff!