Crossing out with CNC

Not really, I was aiming for a decent fit in the slots in the front wheel hubs. A stack of slips gauges measuring 0.750" fits, whereas a stack measuring 0.751" will only just fit, with a bit of a shove. So most of the spokes should be a nice fit, with just a lick of the file needed on occasion.

Ian: It sounds like you're well on the way to designing to make best use of CNC. Had I used a separate cutter for roughing and finishing, the cutters would have been identical, but the one used for finishing wouldn't have been worn away by the roughing. Using separate cutters is fairly common if one is chasing tenths. Having to mount a part in more than one orientation has several cost implications. One, as you say, is the cost of manually moving it. Two, the cost of a second fixture and third, problems with ensuring that the part is accurately located, especially if critical features cannot all be machined in one orientation.

In principle it should be cheaper to have lots of parts machined from one bar for separation as a last operation. However, I'd talk to the CNC house and see what they think. Don't assume they know what they're doing! The company that had a go at machining our heatsink gave us a lot of waffle about their old Ferranti CMM measuring to microns. Given that the room it was in was essentially a corridor between the workshop and offices proper, with people coming and going all the time, I just asked if the room was temperature controlled.

Regards,

Andrew

.

Ian,

If you do feel the need to read ... try this.

MichaelG.

Edit: sorry, can't get the hyperlink to insert properly, will try again later.

Edited By Michael Gilligan on 19/11/2013 22:18:46

No ... Oxford is not playing ball.

Just search Google for:   instability of invar

The first hit will probably be for a book by Matthys ... Chapter 23

Edited By Michael Gilligan on 19/11/2013 22:28:56

Interesting book Michael however the abstract of chapter 23 is a bit confusing. At first it states that Invar was found to grow by 50 ppm in 27 years. Later it states that modern Invar is more stable but goes on to state that it has a dimensional stability of 2-27 ppm/year. Does this mean that it grows by 27 ppm in the first year reducing to 2 ppm after some (undefined) time, is it a variation between batches, or is it a random variation?

What about "Super Invar"?

Russell.

Russell,

My reading of Matthys, and other sources, indicates that the problem was discovered [but not quantified] soon after its invention. Then, in 1927, published data revealed cumulative growth of 50ppm over 27 years; but did not accurately describe the curve, although it was seen to be "... exponential, gradually slowing down with time ...".

Subsequent, very detailed, measurements have shown that the curve is not truly exponential but [one might say] a "digitised" version with random variation ... The overall trend is to grow; with the rate of that growth slowing over the years; but the detail view shows the behaviour to be complex and unpredictable in the short-term.

My assumption is that we are seeing chaotic behaviour as the crystal structure of the alloy "settles" ... rather like the creaking and groaning of a ship's timbers!

Super Invar's Temperature Coefficient is certainly better, but I believe that [to some extent] it still suffers from these instabilities.

.

Current thinking is that Fused Quartz is the best material for a basic pendulum rod ... but then there's the small matter of contriving suitable end-fittings.

.

There can never be a perfect timekeeper based on a pendulum [or any mechanical oscillator], because of the Earth's own instability ... but the best makers have got very close.

{Note: Your interpretation of "very close" may vary}

MichaelG.

.

Ian,

Have a look at this ... then buy Lotto tickets !!

MichaelG.

.

Edit: Cybaman UK homepage.

Edited By Michael Gilligan on 20/11/2013 09:22:15

Accuracy is based on cost - pure and simple.

At the home shop level which is lets face it, the lowest level on the food chain we have all the disadvantages and only a few of the advantages

Each extra advantage is a plug and play module until by the time you have got where you 'imagine' you want to be you are many tens of thousands into a big hole and no longer at the home shop level.

Servos get mentioned over steppers as being more accurate, they are but again not at home shop level.

A stepper system if well designed and run within it's safe limits is very accurate, most of the early Apollo and Shuttle hardware was machines on stepper driven BOSS machines.

Servos we are told are more accurate because they have feed back that tell the motor where is is and where it should be which is very true but how this is implemented is never mentioned.

There are two systems, Best and proper one is that the feed back from the encoder goes back to the controller and it forces the motor into position and checks all the while.

The other way is the encoder feeds back to the driver and tells it where it is and should be, unfortunately without the controller connection it has to assume, just like a stepper it has gone there. If you are running a fast moving program it never has time to catch up before moving onto the next move.

High end controllers [ read £££££ ] use the first system, low end controllers like Mach use the second so that although you have moved up market onto servo's they are only of the same use as a well designed stepper in that you hope they go where you say.

Next hurdle is stickshion if that is a word. Basically it's it the force needed to overcome a static load on a sticky surface.

All machine tools suffer from this whether you realise it or no. Weight of the machine and lubrication play a big part but it is there at all times.

What happens on the home shop level is you command the machine to move 0.01mm and nothing happens because the slide is glued to the base, you command another 0.01mm move and again, same thing happens.

Third time it jumps probably 0.04mm as it overcomes stickshion and catches up. Same applies when you reverse the direction of travel. This is why Andrew probably got the varying results on his spokes.

High end get away from this by better design, again read ££££ . Linear bearings, turcited slides, drives well overspec'd for what they have to do. I have seen servo drives rated at 350v and 60 amps on laser cutters just for positioning as laser cutting is a non contact operation. That's serious power that could kill you in a heartbeat or rather lack of a heartbeat.

Next is heat, we will forget the temperature controlled rooms but they do exist but concentrate on the heat generated by machining which is often overlooked.

A lump of aluminium machined aggressively continually at high feed rates which is what CNC is all about grows as it expands. Not so much a problem in industry as they flood cool everything with shed loads of coolant.

In the home shop we are not geared up to it, don't want the mess and her indoors certainly doesn't want the smell. Result is who rough machines out then allows to cool, re measures and runs a finishing path ?

Don't bother answering that question because it's no one.

So this basically in a nut shell is why we in a home shop environment are limited. If we accept that CNC can work for us using reasonable tolerances then it's fine. If we set our sights to high then we either have to shell out enough money to take you from a home shop to small industry or work manually and slowly to achieve the results.

A decent home CNC will work to a thou most of the time, 1/2 a thou with care.

.

Ian,

Just coming back to this ...

If you fancy a challenge, there is a very substantial-looking "base" in the For Sale list ... it's the Main Casting of an Astra Slotting Machine.

Alternatively, there is a rather nice looking BCA currently listed on ebay.

MichaelG.

Edited By Michael Gilligan on 01/12/2013 09:08:36