Monday, January 19, 2015

Post have been a little scarce.

Posts have been a little scarce as of late. I am still making clocks, but all of my spare time has been put into my new business called Airframes by Rosco, making wooden sunglasses and spectacles!
Here's a little video of making the 3D model for a pair at 64x speed.
See you soon!

Saturday, July 6, 2013

Epi blog

The biggest problem that I find with any project that I do in my spare time is that it gets dragged out over long periods. The problems that causes is things like forgetting what you were thinking last time you had a 'play' or weather or not you had fully thought out what you had done, then coming back only to continue expanding onto something that may not be correct.
Or maybe thats why they call them prototypes. :)
Well, I had a couple of small problems with this prototype, that could have been disastrous, but all worked out fine, call it luck, or call it something that I did actually work out and forgot that I had worked it out.

The cut of all the parts went well. No mishaps. The first little problem I found was with the internal gear. It just seemed to have a bit of 'grinding' in one place, which I think is a combination of the screw mapping on the cnc machine not being perfect (it will never be) and a bit of sag in the bearings. I use high speed bearings, because they have a bit more 'play' in them for when the bearings heat up and expand.
You can see where the black gear meshes with the teeth on the internal gear. That was easily fixed with a bit of a san down of the small gear or pinion by .3 of a mm. Smooth as silk then!
The next problem came in the form of the 2 gears that you can also see here in the pic above. The clear and the black behind it. The clear gear is 36 teeth and the black one behind it is 33, with the same diameter...impossible you might say, but it just needed a rejig of the tooth shape and width to get it to work.

So I got it running, and quickly found out that it was doing a 1 hour revolution of the minute hand in more than 15 minutes...This was a problem.
Oh, I also found out at the same time that I had made the chassis too short for a 90 BPM pendulum... :(
Thats why you can see it here on aluminium blocks.
This is where I get my comment at the start of this blog from.
A...had I forgotten that I was allowing for a shorter pendulum
B...I had miscalculated the pinion to 24 leaves on the escapement wheel and continued to keep designing the clock. It should have been 6.

Well, after a good sleep I realised, that I could change the pinion on the escapement to be 6 leaves instead of 24...I had never made a 6 leaf pinion, 8 was my smallest. And since it was running a little less than 4 times too fast I could shorten the pendulum to be what I calculated to be about 93 or 94 bpm!!!

You can see the old bearing hole for the 24 tooth pinion just below the escapement wheel 2 above, and the behind shot showing the new 6 leaf pinion.
BTW, the positioning of the hole for where the new rod with the pinion and esc wheel would have to be PERFECT. Was quite a trick, but it meshes very smoothly.

The next problem came in the form of not having enough fall for the weight for the clock to run 24 hours!!! When I designed this I forgot that the previous clock called Mini, had its barrel and maintaining power above not below. The reason for that was so that there was a rod coming through the clock at the 5 and 7 hour positions, so that the weight could be lifted all the way up to the 10 and 2 hour positions. On this clock, because the rod with the barrel and maintaining power were below, the below supports had to be wider now out at the 4 and 8 hour positions which is directly below the 2 and 10 hour positions...which meant the weights could only be lifted up till they hit that support at the 4 and 8 hour positions.
The chassis of Mini is just above. You can see the wider supports above for the weights to pull up to. Above Mini you can see Epi. Its weight supports are out wider but not as high.

The only solution was to hope that the weights could be short and drill new holes for a rod out a little wider at the 3 and 9 hour positions. You can see above. This worked fine, but I quickly realised that this gear train really isn't all that more efficient than a traditional gear train (as I was lead to believe) and the proof of that is that here I was really comparing apples with apples. The other clock called Mini would run fine, but this clock would only run for 10 to 20 seconds with the same weight!!! Same parts directly lifted from the design called Mini, using the same materials... It is maybe slightly more efficient. What saved me here was something I have been keeping up my sleeve.

What I did was to shorten the length of the crutch to lessen the arc of the pendulum, which in turn needs less power to run the clock, which meant I could use less weight, which meant that the weights could be shorter and I could get the clock to run for 30 hours (6 hours grace). whew what a mouthful...
Normally shortening the length of the crutch won't shorten the arc of the pendulum, but with this design it does. It is now swinging at less than half its original swing and its running with lots of power!

So for now, Epi is running fantastically! She just needed a little bit of thinking to get her going, instead of rejigging the design to make a beta version, which would have been a lot more work.
I am very happy with the outcome.
Hope you enjoyed the insight.

Wednesday, July 3, 2013

Epi Vid

Well, Epi is completed and running.
Here is a video I put together of it. (not the best)
I will post a new blog and some still pictures soon.

Thursday, May 9, 2013

Strutt Epicyclic..."Epi"

So the Epicyclic clock made by William Strutt looks a bit like this...
And the gear train looks a bit like this...
The gear train essentially does its 'calculation' in place, instead of stepping 1 by 1 through gear ratios.
I'm sure there is a better way of explaining that.
One of the interresting parts to this gear train is that it uses an inverted gear, or a gear on the inside of a wheel, not on the outside. I have never tried one of these before.
So I thought this a great opportunity to sport the new Grasshopper Escapement with it.
Here are a couple of pics of the design steps.
Here the teeth count is just way too large, so I needed to reduce them.
This is where it was going...The pendulum hinging up so high doesn't work too well for various reasons.
Here's another style, but I wasn't liking it at all.
I finally decided to go with the configuration that was used on Mini. It works well. So I flipped it upside-down and used similar legs to mini, with the added vertical legson the sides. I think they look better.
You can see the 'guts' of it a bit better here. I am to the belief that it is a more efficient gear train, so I am hoping to be able to get a 2 day run out of it at least, but we will see. Mini runs for 30 hours only.

Here is the final design...I think. I'll just need to spend a bit more time looking at it, to make sure it will work.

Holidays are great for getting this sort of thing done. Stay tuned for more updates!

Wednesday, May 8, 2013

3 Mini's and a new Grasshopper!

Well I can't believe that it is over a year since I have posted a blog.
There are now 3 versions of Mini that are running, 2 at home and one at work.
2 of them use the Graham Escapement and there has been a lot of development into a new version of the Grasshopper Escapement. Also a lot of work over the last year into making my clocks more reliable. Essentially I wanted to get them to a point where they don't need any lubrication, and I am pleased to say that I think i have it licked!
Here is a picture of the new Grasshopper...
It has a very particular ballance of parts, and has taken 10 minor version changes to make all of the parts work right!
This new design is based from a design that I found on the net, where the designers have strived to produce the best possible version on the escapement based on Harrison's design principles. Here is a picture of the geometry before it gets made into somthing that is real.

I will post again soon with my new clock design, "Epi" which is short for Epicyclic.
This uses my new design of Grasshopper Esc and is a totally new type of gear train.


Saturday, February 25, 2012

The Mini is ALIVE!

The Mini is running and doesn't want to stop...which is a good sign that the design is good. You will notice the vertical 30cm ruler on its right, leaning against the speaker.

The weights are a little large, but the only things that I could muster at the time. I will be able to make their diameter a lot less by casting solid lead, rather than using lead shot. (Casting gets rid of the air between the small lead balls) The weights drop 10mm per hour, so a 30 hr run gives a fall of 30cm! It is also sporting my new gear tooth design that is proving to be superior to any I have created or purchased/copied.

Here are a couple of construction pics...

Here is a picture before going to bed...

Here is in the morning...

I'll keep you updated!
cheers rosco

Thursday, February 2, 2012

Mini Design

Here's a new design of mine that I am looking forward to making. The reason for its existence is simply that you don't have to hang it on a wall! The clocks that I have designed so far need to mounted on a wall, and the fixings need to be able to support a good 8kg's. Which means for mounting you really need to commit to where it is going to go. This clock you see above is only 600mm's high!...maybe the height of a large screen LCD TV. It will run for 30 hours with the weight falling only 300mm (10mm per hour). It's tricky, but I am confident that it will work. (Especially with my nice new gear tooth design.
I received a reprint of a clock making book written in 1905 for Christmas! (thanks to my lovely wife Kelly) It described some alternately geared gear-trains. I have used that info here. This clock runs at 90 BPM and has a 22inch pendulum. There are many other exotic gear trains and BPM's, but with this, the 30 tooth escapement wheel will always turn 1and1/2 times every minute and make it a lot easier to set the pendulum period with the minute alternating between the top and bottom of the wheels turn. (I was originally going to do 88 BPM and that would have been a real pain)
Here's some preliminary designs drawings...

This one is looking at gear sizes and where to put the escapement (the green lines) I ended up opting with it upside down like on the left most sketch.

This one is looking at skeleton chassis parts for the guts and you can also see on the left, the 2 vertical skinny rectangles...they are the weights(1 on each side for balance)!

Starting to go together...

Here I have realised that a large gear driving the escapement is not going to work, as it will hit other axles. So I then have made it into 2 smaller ones to get the same ratio...Although the escapement will now run anti clockwise.

This picture is getting a lot closer...

These 2 are with no chassis...

Other side...

Will keep you updated. Hope you like it!


Friday, December 9, 2011

Nixie Tube Clock

For Sale  $250!!!
Here is the last of my Nixie Tube clocks for sale.
This is $70 cheaper than all of the others I have sold, but this is not numbered, dated and/or personalised.
Very accurate, totally cool and mows laws. What more could you want!
Email me if you would like it!

This clock is quite safe with the open circuitry, but is not suitable for a position/home where children can get to it.

Tuesday, December 6, 2011

Breakthrough after breakthrough!

Its been a week of break-throughs, I am pleased to announce.
1. A winner gear tooth profile.
2. Double safety measures for the recently made, yet to be named electric clock.

The first being a new gear tooth profile, has got me grinning from ear to ear. I have spent around 3 years now trying to come up with a tooth profile that works Super efficiently. Basically put, its really easy to make a gear tooth profile that works, and they will work in varying degrees of efficiency of energy transfer from the gear tooth to the mating pinion tooth (smaller of the 2 gears).

 The place where the efficiency is most needed is earlier in the drive chain where there are large amounts of pressure on the teeth. And especially where the next 2 teeth start to engage. What tends to happen with a badly designed tooth profile, is that the newly engaging tooth tends to 'crash' into the face of its mating tooth. This causes a drastic decrease in drive pressure and much more often than not, this is when/where your clock will stop because there is not enough torque to turn the escapement wheel. The trick is to get the 2 new mating teeth to smoothly slide together, like on a tangent. This I have managed to do fairly well in the past with certain gear ratios, but I have found that once the ratio changes the gears don't tend to mesh as well.

This occured to me recently on the Gravity Escapement clock (the one that is a bit like a 4 leaf clover) Gear ratio 64-8 meshed beautifully, but I had to change it to 64-16 to get the drive weight down, and it didn't like it at all. Also some escapements accuracy is dependant on the drive pressure being constant...and a drive pressure that fluctuates up and down as teeth engage is not good. I have 2 of the same clocks sporting different teeth profiles and the different accuracy between the clocks is VERY evident.

So after around 5-6 different gear tooth profile designs and studying many others over a period of around 3 years, I am pretty sure I have cracked it! ...No actually I know that I have cracked it!! Sorry, but I am going to have to keep it a secret on how to reproduce it. :) Whats more, I think that they are quite pretty looking to!

Not being an electrical engineer by any account, I have been quite worried about running the clock with no supervision. I have once gone upstairs at home for 15 minutes, only to find on my return that the clock had stopped, because the mercury switch got stuck inverted. This has left the coil of the solonoid energised and upon investigation, it was found that the coil had heated up to 67 degrees celcuis! Actually not at all dangerous at that temp, but it can burn skin! The wire will melt at over 300 degrees C and the enamel will burn somewhere between 130 and 300... The coil will most likely reach equilibrium way below that temp (that means energy coming in equals energy going out, or heat from the power to the coil is equal to the heat radiating to the air). BUT I am not keen to try it.

I have tried many types of fuse values and makes, but have not found a solution. Upon talking with a family friend (Tony Roe), who IS an electrical engineer, I found 2 other types of components that I am sure will have it sorted out. Only time will tell, as they are now added into the circuit. The first is a resettable temperature switch. This is mounted on the back of the coil itself, and cuts the circuit if the coil measures a temp in excess of 60 degrees and reconnects the circuit at 45 degrees. The other component I have put into the circuit is a PTC or a "polymeric positive temperature coefficient ". Put simply, its a fuse that when current is drawn through it above its rating, the polymer inside it melts and stops conducting electricity. The current of the coil draws 250 mA for periods of about .7 of a second. The PTC internals melt when a current of over 100mA flows for around 7 seconds...and stay melted untill the current is turned off. Which is actually better than the temp sensor, because the temp sensor will oscillate between on and off as the circuit turns on and off!

So its been agreat week of breakthroughs.
See ya round like a rissole!

Monday, November 28, 2011

Beta of the Electron eata goes like a bought one!

Here she is!  And she's running like a beaut.
Here's a pic of the mercury switch...It now only switches a transistor on and off instead of switching the actual current load that energises the coil. But there is 2 advantages to that. 1 is that I can alter the amount of current that energises the coil with the use of a variable resistor...instead of having to add or subtract windings on the coil, and 2, when the mercury switch was carying the full current load, it was sparking inside and carbon from the 2 electrodes was coming off and blackening the glass. I put a tissue behind it so you can see the switch better...
Here's the transistor with standard resistors...Yeh I know it looks bodgee!
Here's how the mercury switch works...below...
There are 2 systems, the drive system and the time system. What the mercury switch is used for is turning on and off the coil(red thing at the bottom right), which impulses the pendulum and keeps the clock running.

In the first picture below, the pendulum (vertical bit on the right) is swinging from right to left, and the small horozontal arm has caught on of the teeth and is turning it anti clockwise. This action will eventually bring one of the cam lobes around on the black wheel, which will tip the backwards shaped tick arm, which then tips the arm that the mercury switch sits on. It works just like the points in an old cars distributer, because it only impulses the pendulum on and off at a time and duration determined by the shape of the lobe.

Here's how the time keeping system works. Each swing of the pendulum makes the horozontal arc shaped arm (below) catch a tooth on the small count wheel, which slowly tuns the black snail shaped wheel (it takes 1 minute to do one revolution) While the snail shaped wheel is turning it is lifting the black horozontal arm which is resting on it. There is a clear arm hanging from the black horozontal arm and it is lifting a tooth on the large count wheel.

When the snail wheel completes a full revolution, it drops the black arm and the arm clear arm hanging from the black arm drops to pick up the next tooth on the large count wheel.
Below you can see the vertical clear arm enguaged with the large count wheel. The large count wheel has 60 teeth for each minute of the hour. Sorry for the ugly bits of bluetack, but their needed for adding weight. It is a prototype afterall...
Hope you like the pics, I'll put a video up of it soon. It looks great when it's in motion!

Monday, November 7, 2011

Beta of the electron eater!

This is (I think) the final design to be made into the production model. Everything in flouro green (including the red and purple count wheels) has been altered because of design shortcomings discovered with the 1st prototype!
The red and purple count wheels went to the dentist to have their teeth reshaped so that the the backstop and pulling arm (respectively) fell into the wheels, not leaving a recess for backlash to occur. (see top right and bottom left yellow bits)

The face no longer supports the centre shafts, as it created too much friction of there was a small amount of twist in the clocks chassis. The new face design looks better too!

There have been 2 lobes added to the cam wheel (above), to make a total of 3 lobes. This will make the impulsing circuit energise the coil every 10 seconds, instead of every 30 seconds. I think that should be enough to keep it running. Only a long run like for an hour or 2 will tell, and I am yet to install the 3 lobe cam into the 1st prototype to see how it goes. Here is what the 1st proto type currently looks like...
I don't like the all black, it was just easier to cut all of the pieces in one cut. When this was first assembled, I wasn't feeling that is was going to all that successful. With the new mods made to the design, I am really confident that its going to run well! Fingers crossed.

I think I am going to try a Mercury Switch!
Here is a design that could work. The mercury switch is the spaceship looking thing mid left. The long arm on the lever is to amplify the small lobe on the on the cam. We'll see