Revisiting an old hobby

[JohnD]

11 minutes ago, AndyHull said:

Hold my beer, I'll have that working in no time.

It's been in a 'finds' cabinet on my wall for the best part of 30 years Andy. As far as I can remember it is double cased, has a fusee movement and  is hallmarked 1808 and I suspect the hands are solid gold..... I really aught to take it out some time and have another look inside and take some photographs....One of the coins attached to it is a young head Victoria silver four pence, probably a Maundy piece......

I got this working but I doubt that the pocket watch would be as 'easy' a task...

Edited April 10, 2020 by JohnD

[AndyHull]

That's not a watch, that's a 3d jigsaw. 

[JohnD]

1 minute ago, AndyHull said:

That's not a watch, that's a 3d jigsaw. 

That's how it arrived (from Israel) Andy, fortunately I was able to utilise a few bits and rebuild it using a 7546 donor, looks like this now, shame about the dial...

Still it was like this....

[AndyHull]

Coming in the mail, but not actually a watch. A bunch of EB88xx and BFG parts and of course a scattering of other random stuff. 

Some of these job lots are a treasure trove of hard to find parts.

 

[AndyHull]

One of my other hobbies is electronics, and today's fun combines both electronics and timing... and hazardous radioactive materials, so what's not to like.

First a bit of back story. A couple of years back, I was active on a microcontroller forum for the STM32F103 microcontroller. One of the things I was involved in was getting the real time clock part of this chip to play nice in the Arduino environment, as I wanted to use the chip to control a sunset and sunrise camera rig for time-lapse photography.

As a result, I have a bit of experience in coaxing crystal oscillators to oscill-ate. In particular 32768 Hz watch crystals. 

While messing around with this (and an oscilloscope project using the same microcontroller and a cheap LCD display), ebay in its rather spooky way, suggested a job lot of radiation dosimeters, in its usual snowstorm of random suggestions.

They were naturally not working. 

ebay seems to have a spooky understanding of my psychology.  Naturally I couldn't resist, and as a result I ended up with a pile of SOR/T ex military dosimeters. I messed about with these for a while, and got a bunch of them to do something, then as often happens, my attention drifted off to the next project.

Yesterday I found the box with the dosimeters while looking for something entirely unrelated and I thought it would be fun to dissect one of the dead ones to see what makes it tick or more exactly to see why it didn't tick.

This was a somewhat medieval process as they are designed to be idiot proof and therefore are sealed against both the elements and idiots. There is however no such thing as idiot proof, as I soon demonstrated using a screwdriver, a sharp knife, a pair of wire cutters and some colourful language. 

Here is what lurks under the plastic skin and liberal dose of silicone sealant in a SOR/T  dosimeter (minus the ferrite radio antenna which I carefully removed to make things easier to work with).

The other side of the PCB..  The MITSUBISHI MICROCOMPUTERS 3825 Group SINGLE-CHIP 8-BIT CMOS  microcontroller, an EEROM, reed switch and some other gubbins. 

Note there are also two "tin cans" and an epoxy blob on the board. I'm not sure which of these contains the gamma sensor, and which contain the radios. This thing can speak to the mothership from up to 1km to let it know how much the wearer is getting cooked apparently .

That big black rectangle on the right is a super capacitor that keeps things running while you change the battery, and the reed relay shuts it up while you are torturing it, if you place a magnet next to it.  Presumably this acualy allows the programming device some level of control, but I'm not sure what exactly.

So what has this got to do with watches? I here you cry in a slightly bored voice.

Well it turns out, from reading the datasheet for the microcontroller, and after some prodding around the thing that has failed on this dosimeter (and probably sever others from the pile, with the same symptoms) is the crystal oscillator.

Furthermore the datasheet mentions that the microcontroller on this device has a low power 3.3V mode that uses allows it to run with a watch crystal.

That is what the gold rectangle just below the push button is.   A no doubt very expensive military grade, gold plated watch crystal.

So... I poked it some more. Specifically I poked the crystal with a hot soldering iron and my multimeter, and got some very odd readings, but at the same time, it burbled into sporadic life. So I un-soldered the crystal and rummaged about through my parts bins looking for the bag of watch crystals I bought when I was playing with the STM32F103 chips.

Naturally they were nowhere to be found. However what I did find was an  old LCD watch module in the watch repair junk stash. So this was hacked apart, and the crystal extracted. I swapped it in, and things still didn't work.

So not the crystal then, so if not that, then it must be the load capacitors, or the microcontroller.

The offending load capacitors and the crystal from the "Hong Kong special" LCD watch module are on the green cutting mat above the now working dosimeter. 

I poked the capacitors with the hot soldering iron, which resulted in some interesting chirps and nonsense on the display, so I removed the load capacitors and it sprang into life.

Quote

Wikipedia: For example, a crystal intended for a 6 pF load has its specified parallel resonant frequency when a 6.0 pF capacitor is placed across it. Without the load capacitance, the resonant frequency is higher.

Now realistically I should replace them, otherwise the oscillator may be out of spec, and I do have some, but they are in the same bag as the aforementioned  32768 Hz watch crystals, so for the time being, it can do without.

Next on the agenda will be reprogramming the EEROM, to give me full control of its brains, but that is a game for another day.  Back to fixing watches for now.

One final thing.. I now know that the background radiation in this neck of the woods is about 0.004 mSv/day which equates to 1.46mS/year which is slightly below average, but this thing is only measuring gamma radiation and has a fairly large margin of error. None the less it is a reasonable ball park figure. Apparently all of those luminous watch hands are not actually cooking me after all.

 

If you want to read the datasheet for the SOR/T -> click here <-

If you want a fully functional, calibrated one then -> take a look here <- Warning, being military grade, they are not cheap.

Edited April 13, 2020 by AndyHull

[AndyHull]

Four more resurrected dosimeters. They had a variety of faults.

Two more with dodgy load capacitors.

One with the LCD cable coming adrift from the glass, which I glued back down with a tiny amount of "GS Precision cement".

This surprisingly worked first time. I'll need to remember that trick.

One had a crystal which was shorted to the case, presumably due to an impact. I desoldered the can of the crystal from the PCB and it instantly sprang to life. I removed the fancy gold plated, but duff crystal and substituted the one from the LCD watch module, which works perfectly.

And finally I decided to conduct the experiment which absolutely has to be conducted when you have a bunch of dosimeters you don't know what to do with, and a bunch of mildly radioactive watches.

I placed the "Beaver" face down on its sensor. (Insert off colour joke here).

I'll keep you posted if anything exciting happens, but it is a very old Beaver, so this could take some time.  

EDIT: One little bit of weirdness I also observed while messing about with my multimeter in Hz mode.
The anglepoise magnifier I use, has a built in compact fluorescent lamp, which emits a hefty amount of electrical interference at around 35kHz, which bizarrely can be injected into the module's clock pin, when the crystal is removed, by using the multimeter as an antenna of sorts.   

This causes it to spring in to life, all be it with a very flickery display, presumably due to the rather low level of the signal, and the other RF junk that would be picked up by the multimeter leads at the same time. If you switch off the lamp, there is still  the residual 50Hz mains hum, but this is too slow to drive the device correctly.  It does however cause it to produce some random gibberish on the display and a few strangled chirps from the piezo speaker/beeper.

Edited April 13, 2020 by AndyHull

[AndyHull]

The original thread for these devices was archived about a year ago when the https://www.stm32duino.com/ site was in flux.

You can find it here, along with the related discussion of pin diode based radiation detectors. -> Click Here <-

The archived thread contains a lot of interesting information for anyone who would like more details about solid state radiation detectors.

[AndyHull]

Exploded Rotary. This looks to be a high end Chinese standard auto-winder caliber, which is slightly disappointing considering Rotary are a "Swiss" brand. However since I only paid a few pennies for it, I'm not that bothered.

https://www.amazon.co.uk/Rotary-Leather-Automatic-Skeleton-GS03862/dp/B00FS2JJW0

If I had paid full RRP, and it fell apart like this, things might have been a little different.

The source of the digits is obvious, and the long thin spring is clearly from the autowinder (which still kinda works, but only in one direction), but the little spring with the hole in it has me scratching my head. It does manual wind, so the broken winder is not a show stopper.

The next question is, do I feel brave enough to try to re-fix the dial furniture, and if so, how the heck does the crown come out of this dratted thing.

If I manage to not screw up the dial, I'm fairly certain I should have another scrap movement with the needed autowinder spring. 

It is running remarkably well considering all the junk I had to remove from the works.

Edited April 19, 2020 by AndyHull

[AndyHull]

More fun with old junk.

A few years back, while walking the dog through the Stirling University campus, I noticed a bunch of old microscope "stuff" in a pile of junk beside the dumpsters.

Free toys!!!

Naturally I grabbed what I could, and took it home.

Don't judge me, you would have done the same.

From the junk I managed to make the fully working stereo microscope in the picture above, which I use regularly.

I have an Olympus too, which I bought about 15 years ago.

These are great, but they are a little awkward to use for watch work because they are actually a little too powerful, great for examining jewel holes, gear teeth and so forth, but because the subject is pretty close to the lens, actually working on the watch is a little cramped.

Also in this pile of junk was a random selection of parts for "The Grey Five Forty" specimen microscopes, which should look a little like this.

The main bodies for two of them are there, as are two of the stands, but no lights.

I can easily arrange for lighting, but the more important parts that were missing from both were the metal housings for the eyepieces. However life moves on, and the parts were consigned to a storage box until I rediscovered them a few months back.

I opened up the box and thought, I could 3d print those, so after a little bit of wrestling with FreeCad, I came up with a design for the lower half of the eyepiece assembly, which I printed. The model is sized from the only existing eyepiece holder body that was in the junk pile.

The project then went on hold again (real life keeps getting in the way), but I picked it up again this evening.

I purchased a couple of random Chinese eyepieces in the hope that I could marry them up with the printed housing, and tonight I got a chance to do a quick test fit. The good news is that it works. 

I chose 30mm x5 eye pieces that would also work with one of the existing microscopes, since I had no idea what the originals were, and I didn't want to waste too much money on the project. It looks like they were a good choice.

It works extremely well. The slightly wobbly hand held camera image above doesn't really do it justice. The view through the eyepiece is in reality pin sharp, very bright, and there is a good depth of focus. I can see for example that the large centre jewel has a tiny spec of dirt in it, and there is a small fiber on the bridge next to the jewel below. 

Now I need to print the right hand assembly and model the eyepiece holder tube, and I will be able to lash it all together finally. 

I  also intend to 3d print an adapter to allow me to strap a camera or a phone to the eyepiece, which will eliminate  "shakey cam" pictures like the one above.

Furthermore I may have enough bits to make two complete microscopes (assuming the second set of optics clean up as well as the first ones did).  I will need to make a cover that doubles up as a storage case for it/them too, as they take up quite a bit of space.

Before I do any of that though, I really do need to tidy up that work bench again.

EDIT: It has just occurred to me that the name of the microscope "Grey Five Forty" presumably refers to magnification. There is a lever on the right hand side that swaps in an extra set of lenses, so presumably it goes from 5x to 40x when that lever is operated. I wonder if fitting my x10 eyepieces will change that to x10 and x80.
There is only one way to find out.

It also has a filter adapter on the bottom, so presumably I can fit a polariser too.

Edited April 19, 2020 by AndyHull

[AndyHull]

2 hours ago, AndyHull said:

I wonder if fitting my x10 eyepieces will change that to x10 and x80.

It does indeed. Furthermore I could 3d print an adapter to take the 23.2 mm eyepieces from the other microscope. I should be able to use anything from a 5x wide field 30mm dia up to  a 25x 23.2mm dia. in conjunction with the 5x/40x lever to give a very broad range of magnification.

The working distance between the base of the body of the microscope below the filter adapter ring and the subject is about 85mm and I may be able to adjust this by extending the length of the adapter tubes. 85mm is plenty of room to get my fat fingers, tweezers, screwdrivers or whatever in to the action.

[AndyHull]

Five minutes of FreeCad and an hour or so on the plastic spaghetti machine and I have a prototype eyepiece tube almost completed.

Meanwhile I did a little more digging around and found the second microscope in the attic. The stand was in a box in the shed, so I re-united them. This one is a mixed bag. It has the original eyepiece holders, but they are 23.2mm and pretty bashed up, so no longer light tight, or straight relative to the optical path. I'm not sure if I'll print 30mm replacements, or patch up what is there and print 30mm adapters. 

As you  can see from the image above, the working distance between the bottom of the optics and the subject is pretty good.

The other odd thing I noticed is that the two different heads appear to be identical except that they are are not interchangeable. They have a different ratchet focussing arrangement. Very similar, but sufficiently different to make them non-interchangeable. Presumably even though they are both stamped "The Grey Five Forty" they are slightly different models.

I also have a second stand. I wasn't sure about that. The stands however do appear to be interchangeable. 

I'm going to convert them to LED lights too, so I can run the light from a USB charger or a power bank.

The second set of optics are very dirty, but I don't see any scratches or mould on them, so I'm going to have a crack at cleaning them back to factory fresh. As you can see from my quick game of "spot the hair on the shock spring',  with a little bit of effort it should be possible to get some excellent images from them.

Time for some alcohol, cotton buds and bamboo skewers. 

I'm currently measuring up the camera so I can design a rigid mount that will drop over the top of the eyepieces, which will make it possible to video things too. Hand holding the camera is very hit and miss, particularly as the zoom starts to "hunt", the moment the camera moves, which can cause the lens to crash in to the eyepiece if I am not careful.

Edited April 20, 2020 by AndyHull

[AndyHull]

The 30mm prototype next to a 23.2mm original. I superglued the two parts together, and filled any gaps with 5 minute epoxy.
Alignment and test fitting the two parts involved some large lumps of bluetack. Pro tip. Don't ever put superglue near optics or glass. The two do not play nicely together, hence the test fit and careful marking prior to mating the parts without the lenses. 

It looks a little scruffy, but nothing a little filling, sanding and a few more coats of primer and paint wont fix. I've made a couple of modifications to the dimensions, as the tube fit for the eyepiece was a little snug, and the hole in the top plate of the lower housing could do with being a little larger. I also need to make a base to allow me to fit some captive nuts so it will bolt on to the metal plate of the microscope head.

I'll finish this prototype, and see if I spot any other snags, then print the remaining parts.

[AndyHull]

A lot of tinkering and some more FreeCad time, and I have all of the parts ready to print.

I increased the length of the eyepiece tubes from 32mm to 60mm as this allows for better binocular separation.
I added a flange on the eyepiece tubes so they should mate directly into the bottom sections, and I increased the diameter of the hole to accommodate this flange, and a bunch of other refinements.

It is way to late in the day to start printing anything, so I'll probably start this job in the morning.

[AndyHull]

I'm happy with the eyepiece tube design and dimensions. The fit is snug without being too tight, so I can place the two eyepieces at fractionally different heights if I need to correct for mismatches in focus, and the internal fluting works, so there is no internal light scattering and base flange all look to be correct.

I'll see if I can integrate a suitable pair of M4 nut holders into the base pieces design,  and then print the remaining parts  tomorrow. 

[AndyHull]

Before we actually run out of masks, I thought I'd rattle up a little bit more PPE on the sewing machine while that lens tube was printing.

I ripped up a couple of old pillow cases, a blouse with double buttoned sleeves, and a pair of jeans to give me some material to work with.

Here I am wearing the latest in recycled stripey pillowcase face fungus fashion.

I was tempted to dive in to my wife's patchwork quilting material, but I think the resulting earache might have been worse than the virus, so I stuck with the scraps bin.

While I was ripping things up and sewing them back together I also took the opportunity to make myself a couple of small tool roll style bags from the double buttoned sleeves.

This does of course increase the likely hood of misplacing, not just one small screwdriver, or pair of tweezers, now I can loose them all at the same time.

Stay safe everyone.

Edited April 21, 2020 by AndyHull

[AndyHull]

I printed the 2nd lens tube and have started to print the first of the two bottom housings.

So far so good. I may stick with the red colour of the plastic, or perhaps I should go for the distressed metal look.

 

[AndyHull]

I stuck with the red, and fired a scoot of flat, mat black from a rattle can on to the inner surfaces to keep out any external light ingress (otherwise the view tends to takes on a rather rosy tint).

Red certainly brightens the place up a bit.

I still need to attend to the dirt on the optics, but it works. I'm just waiting for the second base to finish printing and this one is almost done. Then I'll print the second batch of parts.

[AndyHull]

Wow. That actually works really well.

The old rickety broken bits were really cramping its style, but extending the tubes and fitting the wide field 5x lenses has turned it into a really nice microscope.

I did hit a slight 3d printing glitch. That line on the right hand base, caused by the filament catching on the drum which in turn caused a missed step on one axis.

I can probably fix it by splitting and gluing along the missed step line, or more likely I will simply reprint the parts. As it stands, the flaw makes aligning the base with the screw holes tricky, so a reprint would probably be a better option. I need to print another set anyway for the other microscope head, so two more prints wont be a major hardship.

In the unlikely event that someone else tries to do something similar, the mat black paint on the inside of the parts is essential. Without it there is way to much stray light getting in. This was one of the issues with the broken holders. Light was getting in from the base, and interfering with the sharpness of the images. The new tubes are orders of magnitude better than the old broken ones.

Close up watch work with stereo vision. Its a whole new world. It might even help improve my micro soldering skills too.

EDIT: The glass stage that the watch holder is resting on is also "upcycled". It is a small tractor quarter light window that I found lying around while clearing out one of the old sheds nearby. I was going to use a ceramic tile but the tempered glass with the curved corners looks way nicer. I stuck self adhesive felt feet on it and now it looks like a bought one.

Edited April 22, 2020 by AndyHull

[AndyHull]

A slight hiatus in the microscope building today, due to a minor breakage in the kitchen, which necessitated a 3D printer part.

Braun insist that if you break this flimsy little plastic linkage...

 

You need to buy this...

.. which naturally costs almost as much as simply replacing the whole hand blender.

However a user on Thingiverse has done the work for me, so instead I printed this.

Time will tell if it is as robust as the original, but since it cost about 20 pence worth of plastic I can print about 150 of them before I come any where near the cost of the Braun approved alternative.

Back to printing microscope bits tomorrow I hope.

Edited April 23, 2020 by AndyHull

[AndyHull]

I can see that watch mechanism perfectly.

It is pin sharp super bright, and fills the frame, furthermore I can work on things easily and see specks of dust and other debris that I might otherwise have missed.

This camera , however has problems even when I have it in manual focus mode. Currently it adds too much additional magnification, and struggles to focus correctly.

This about as good as it can manage currently.

I have a couple of other cameras which may be more suited to the task.  Time for a bit more experimentation.

[AndyHull]

A little progress on the second microscope today.

I still haven't decided if I'm going to fit those 23mm adapters to it or not.  I may make an adapter for the adapters as it were, so I can swap them in and out.

I'm also going to do some more experimenting with finishing on the PLA plastic to see if I can get a more uniform ans smooth surface. I was going to try car body filler, but none of the local suppliers are currently open, due to the lockdown so that idea may be a non starter.

Perhaps I'll just see if I can get a decent result just using multiple coats of primer and sanding.

[AndyHull]

Sometimes problems 'aint what you think they are.

The corona virus masks proved popular. So much so that we broke the sewing machine making some more. Or to be more exact I thought we had broken the sewing machine. The fact that it wouldn't sew seemed like a good indicator that it might be broken. Furthermore, the top thread kept breaking, which suggested the issue was with the top thread tensioner. I tried a new needle. I tried different material, always with the same result. It would sew for a bit, then PING, the thread would break.

So I stripped the thing down. The old machine we have, gives you access to the tensioner, but this modern plastic monstrosity has all of the interesting parts carefully hidden from view, ostensibly to keep you from fiddling with them, but in my opinion the real reason is so that you can't see how cheaply constructed it is compared with its 1960s cousin (does this sound familiar).

It has an "Automatic" thread tension setting and a manual wheel to set different tensions if "Automatic" is not working for you, but to access the gubbins that makes it all work, you need to remove *ALL* of the covers, and most of the screws from the thing, which I duly did. 

I fiddled with it, but could find nothing wrong, so I re-assembled it, and after another few sessions of chunka-chunka-chunka-PING as the thread broke for the umpteenth time, I stripped it down again.

This time I decided to look at the bottom tensioner and the hook mechanism. That all looked fine too. I re-assembled it.... more chunka-chunka-chukna-PING .. more swearing and head scratching. I removed a small bur from the metal of the lower needle plate.. maybe that was it... but no... chunk-PING-expletive.

So.. I pulled the lower case off again.. and then I noticed. The lower bobbin looked a little squint. I took it out, thinking there must be something I had missed, stuck under it, but nope. Nothing. I looked at the bobbin in my hand.. and that was when I had the forehead slapping moment. There was nothing wrong with the machine. The cheap plastic bobbin was faulty.

The thing had cracked, and one side was now thicker than the other. This was enough to cause it to jam every so often against the top plate. This in turn caused the bottom tension to go off the scale, which in turn jammed up the top thread, which then stuck fast in the top tensioner, and made it look like the top tensioner was jamming. 

On the plus side, the machine is now fully serviced, working better than it ever has, and is spotless. 

Edited May 5, 2020 by AndyHull

[FLwatchguy73]

Now you can add sewing machine technician to your resume. With your skill set, you should be all set to stay busy during any future lockdowns.

[AndyHull]

Each not much larger than a sewing machine bobbin, four more Timex Petites each got a well earned bath. The brown dialled one also got a new crystal.

These are actually quite attractive, if basic little watches. They are also ubiquitous, with hardly a job lot on ebay that doesn't include at least one of these. As a result I have quite a number, in a wide range of colours.

They are also very easy to work on, and it seems almost indestructible. The brown one is from 1979, the lower white one from 1974, the red from 1976 as far as I recall, and I don't think I checked the upper white one, but it is clearly the most recent of the lot. All except that recent one are from Dundee, it was assembled in the Philippines. 

Edited May 8, 2020 by AndyHull

[AndyHull]

Here are a few more.

I'm sure there are even more if I dig around.