Member postings for SillyOldDuffer

Partly other things to worry about, partly measuring Q isn't vital to my project, partly I'm not quite fully convinced I'm doing it wrong yet, partly because I don't want to stop the current test run, plus a technical issue to solve.

My electronics produce a relative measure of amplitude, so in theory I could simply change the governor values so the pendulum is never impulsed, which can be done by remote control.

Unfortunately, SK questioning IR sensor performance led me to check how well my cheap Arduino comparator performs. It's good on beam-break detect, but the beam-open detect timing is delayed. Thus my relative amplitude value is wrong, possibly by a constant, not confirmed. Doesn't affect clock operation, but it means the decay curve can't be trusted. Unhelpfully, my amplitude measure is low resolution because the code clips the logged amplitude value to 4 places of decimals to save space. Needs a recompile and install to fix.

Often wish I'd built two clocks, one for long test runs and the second for testing new ideas on. As it is, I don't want to touch the clock until the present test comes to an end. Been running for 43days, 22 hours. During that time I've not been able to apply any hardware or software upgrades. At the moment the clock is 6.297933s slow.

Looking at the data I'm 99% certain there's an error in the way I calculate temperature/pressure compensation, cause unknown. And a suspicion the Arduino applying compensation with single precision floating point numbers is losing accuracy.

Just for laughs here's the Allan Deviation.

If I drew the graph correctly and understand Rawlings/Bateman it means 'Progressive change ins system parameters' with 'Normal type of rate fluctuation. Could be horribly wrong - I don't have the right sort of brain for maths and statistics!

Dave

I'm waiting to see what happens when BT convert my phone. Shouldn't be difficult. There's an extension, wireless base, ADSL router and handy powerpoint. As my landline and broadband are both BT provided, I'm expecting BT to replace the copper line (from pole to house, easy access) with fibre and give me a new router and handset. Or maybe just a router and I have to buy an IP handset. The switch-over should be seamless except for me having to connect all my computers to the new router.

My mum is a bit more complicated because she has a big button phone that she can see and understand, an ADSL router set up by me, and a red-button alarm box. There's a power socket, and I think the red-box has an ethernet socket. With luck it's ready to go, otherwise the red-button device will have to be replaced, hopefully on the same day. I expect I'll have to attend to make sure it all works and buy a new big-button uncomplicated IP phone that mum can cope with. I have an idea grannie-friendly phones are pricey.

Worst case I suspect will be anyone who finds change difficult, has a complicated internal copper set-up, no power-point where the cable comes in, and lives in a hard to wire home.

I've long passed the age where I resent change being forced on me or my technically hopeless family members. They all look to me to make thing work! Should have kept my hobbies secret!!!

Dave

A gripping result!

I'm having trouble conceiving how it works. Presumably low Q is caused by friction between coins and top-plate resulting from the magnetic force. I wonder if Q could be improved by increasing bob weight until gravitational force almost balances the magnetic force.

I'm guessing friction follows the same rules whatever type of force brings two moving objects into contact. Does anyone know?

Dave

You're working much too hard on a time-waster; Mother Nature always wins in the end. Throw the fork away and move permanently into your workshop. It's a safe space where only good things happen.

Dave

Fascinating. I love this stuff.

Graphing, zooming in and doing the statistics revealed many anomalies in the data collected by my Mark 1 pendulum, which had many mechanical and electronic shortcomings compared with yours. Though the latest build is better made, I'm still getting odd patterns and events, especially over long runs. For example, seems the Q-factor of my pendulum varies significantly during runs. May be my methodology is faulty - forum peers haven't approved my way of measuring Q.

I sample every beat, about 0.933s and after 46 days the log file is up to 370Mb. Loaded for statistical analysis into my computer, a 370Mb input file eats nearly 4Gb of RAM. A 20 millisecond sample rate is scary.

Dave

If only!

Seriously though, I'd expect your pendulum to do better than 12000. The build quality is high, the measurement method seems sound (best amplitude measurement yet), and the graph looks right.

Other possibilities: swinging in a draught; frame flexing; floor not solid; suspension not level; elliptical path; external vibration; rod bending...

I've found pendulums to be very sensitive to tiny faults.

Dave

Getting it working is the main thing, installing and learning Python for a one-off requirement is daft if you already know another language.

If 'C' were a road vehicle, I'd say it was the most powerful motorbike on the planet. Wonderful, but has to be ridden with great care and skill or the rider crashes. 'Segmentation fault (core dumped)' is only one of many exciting possibilities the programmer has to manage. Python is more like a posh pick-up truck - much easier to drive with reasonable performance, comfortable, safe and very capable.

The 'under the hood' comment is correct. Apart from protecting the programmer from hard crashes, Python claims to be 'batteries included'. Comes close - it's achieved by having a huge library of modules. Almost guaranteed that someone has written a Python module that supports what you need, which avoids much grunt work and debugging. If necessary you can write your own module. Many are written in C for high performance and efficiency.

Dave

For info, Torx screws for holding inserts come in different sizes and the keys are colour coded:

Dave

Machinery's Handbook covers more ground than anything else but it's thousands of pages of small print. One for a clean Design Office rather than a grubby workshop.

In order of preference:

• For speed, the internet. But beware, not everything on the internet is correct!
• For common workshop needs like screw pitch sizes etc, - laminated A4 pages of info printed off the internet. Oil proof after being laminated.
• my favourite handy reference that meets most home workshop requirements - Model Engineers Handbook (Tubal Cain)
• Machinery's for when in-depth or unusual info is needed. How valuable it is depends on what you do: Tubal Cain covers most common questions.
• For older British, I have a collection of references of which the best is Newnes Engineers Reference Book, 1948. Mostly 1948 modern, including metric, but it covers quite a few legacy items. Books covering British engineering from, say 1840 to 1940, are fascinating because they cover dozens of long forgotten obsolete proprietary standards. British Standards didn't really get going until after WW1 after industry had caused huge problems due to everybody doing their own thing. Most standardisation was done between 1918 and 1945, and reference books from that period are full of changes - date sensitive, approach with caution. They're occasionally useful to me and I expect a restorer would use them much more.

Machinery's hales from the USA, and I doubt even an early edition would cover early British engineering. I have the 20th Edition (1976), and it does US, British and Metric well. Not felt the need to buy a more up-to-date edition yet.

Dave

Excellent!

Though I too would have expected a higher number, the Q-factor is 'about right'. That said, it suggests something can be improved, as does the way Q increases as the pendulum decays.

The Q-factor calculation assumes a particular exponential decay, which may not be achieved in practice. Check the graph's curve to see how close the curve is to theoretical. Could be the Q is higher. (I can't look up the maths now - have to go out.)

I'd start by suspecting the suspension. What could possibly go wrong with a knife pivot? Could be the edge is rounded - not a point or a circle, or it slips sideways at high amplitudes, or is cutting a groove. Might be a 'running in' fault that diminishes as the edge wears into a comfy shape.

Next suspect is the bob shape because air resistance increases with speed. Although the rugby football shape should perform well, maybe it's construction edges or curves cause excessive turbulence. A smoking joss-stick might provide a clue.

Dave

Edited By SillyOldDuffer on 30/08/2023 09:39:46

Well this is what ScamAdviser says:

I wouldn't risk it!

Dave

Edited By SillyOldDuffer on 29/08/2023 17:48:26

Drool! Pure porn. I can't wait to see how it works in practice.

Dave

That's not how this forum operates Tony. We're engineering forum, not a mutual admiration society. Everyone can all be challenged - politely!

Posts often contain technical mistakes, misunderstandings, and methods that can be improved. Members are free to comment on these, and what they say is usually valuable. Model Engineers are not easily offended unless the comment is abusive. We understand having a positive attitude to criticism and other ideas allows us to improve.

With respect, you can't be offended after posting things like:

Now if you mean tricks of the trade from the experts on this site and other sites like this - no thanks, I prefer to teach myself by actually doing, rather than listen to someone who thinks his way is the only way and the correct way. The correct way is what works for the job, not how you were taught by someone who was taught by someone else. That's the problem with experts: No imagination, no visualisation, no thinking outside the box, just do it the way you were taught to do it and don't confuse things by actually thinking for yourself.

The statement is controversial and based on ill-founded assumptions. For example how do you know you're dealing with 'someone who thinks his way is the only way and the correct way'? And everybody on the forum thinks for themselves, not only Tony. Sounds like a personal belief system to me; prejudice not engineering.

Easiest way to avoid negative feedback is not to provoke it. What response was expected to 'I'm sure you'll all hate this, sorry no apologies!? What's sauce for the goose is sauce for the gander: expect to be treated in kind. Please think about what you type and don't accuse others of being 'plain ignorant and self serving' just because they suggest alternatives.

Nothing wrong with 'Tony Method' in a singleton home-workshop, but it doesn't scale up or support team work. That's all we're saying. There's no need to take the hump.

Dave

Not necessarily a second. I recall surface plate standards allow a margin. Te flat surface doesn't have to be correct all the way to the edges, perhaps because margins are harder to grind perfectly and are prone to damage.

Dave

The usual Arduino CNC program is GRBL, which is a free download. Various ways of installing it, you might try searching 'Arduino GRBL' , which will find various sites and tutorials like All3DP.

In outline, a controller board powers the mill's motors, and the controller is told what to do by an Arduino running GRBL. GRBL is fed a list of G-Gode commands that move the axes as necessary to make the cuts needed to produce an object. G-code can be handwritten, but it's more usual to model objects in 3D-CAD and then use CAD to generate the G-CODE.

The process is similar for 3D-printing, except the commands makes a 3D object by adding layers of plastic, concrete or metal, whilst a mill or lathe subtracts metal to create 3D objects. CAD software usually supports both methods. Not tried CNC, but basic 3D-printing is rather easy, almost plug and play.

I watched a Youtube video a few months ago where a chap who knows his stuff enthused about the Proxxon Mill. I suspect they have an undeservedly poor reputation because people expect too much of them.

Dave

Sutter Creek, now that's an interesting historic location!

Is there still Gold in them thar hills? Hope so, I've swapped the wife for a shovel and am on my way!

Dave

I like it very much, but what's with all those holes in the brass plate?

Very pretty, but having plenty of mass at the bottom is a good thing. On second thoughts, it makes sense if the column is bolted to the wall, with the base plate hanging off it to support the optics.

An observation rather than criticism - fascinating stuff.

By the by, it's worth chamfering the bob top and bottom to reduce turbulence. I guess a large round fillet would be best, but a small flat chamfer would do good.

Dave

Don't overcrowd the earth terminal : one big wire with an eyelet is best. The problem is dealt with by an earth block like this example from Screwfix:

Ideally the block is bolted to the metalwork near the VFD, but it can be free-standing. The main requirement of earthing is that all connections be secure - you don't want earth wires vibrating off an overcrowded terminal screw. Insulated terminal blocks are available too:

ABS if you must, but metal cases are stronger, fireproof, and shield Radio Frequency emissions. An unfiltered, unshielded VFD is a potent source of electrical filth. Putting the VFD inside an earthed metal box helps contain its anti-social output.

Dave

Yes, but it depends on what sort of wood was used to make the charcoal. How much ash, water, air, and volatile chemicals are in the wood. Not much heat in Balsa charcoal because Balsa is so light. Other woods have a high volatile and ash content, so a lot of heat is lost when the wood is cooked evaporating oils and acetone etc, leaving a Carbon diluted by ash.

The wood used to make Charcoal for Gunpowder comes from a few high carbon low ash species. The wood used to make BBQ charcoal is highly variable, but as a rule of thumb hardwood charcoal contains more energy per lump than softwood charcoal.

You will need to sample several brands to find experimentally which works best in your loco. And don't assume that a particular brand will always be consistent - there's no guarantee the maker will carbonise exactly the same species of wood every time: more likely he carbonises a mix decided by the availability and cost of wood. Choosing woods that flavour food are more important to him than heat value: the needs of locomotive drivers aren't considered at all. But the right sort of charcoal competes well with good coal.

Most coals are unsuitable for model locos too. Coal is a natural product created over millions of years. It's qualities depend on the type of vegetation, high or low ash content etc, that grew in a primeval forest then on the circumstances in which it was submerged, how much pressure was applied to it by succeeding geological layers, and for how long. Brown coal is very poor, whilst hard coals have a high heat value. Welsh Steam Coal was the best in the world for locomotives. It was a hard coal, created from a low ash vegetation, highly compressed, for a very long time - as good as it gets. Other coals, notably from the Eastern US, are nearly as good, but most are more contaminated. or have lower heat value than the very best. Coal suitable for making coke is a good bet if you can find any, but most of it goes for steel-making. Power stations burn inferior coal, where high ash content and impurities like Sulphur and Phosphorous can be managed. Older Brits will remember 'Nutty Slack', a dusty highly contaminated coal product left after removing the best lumps. It burnt reluctantly in domestic grates and the pollution caused smog - anyone with a lung or heart problem died young! Now we have Smokeless Fuel, coal treated to remove dirt, not quite coke. I don't know if it's any good for model locos.

Not sure what the problem obtaining coal in Australia is : Australians consume more coal per head than any other country in the world! Most likely reason is it's all exported, mostly to China, and there isn't a domestic market.

Dave

Up to a point. They're more rigid, which is good, but can get in the way, which is bad. I have a selection of tools between 12mm and 6mm for a WM280. Most used 10mm with 12mm a close second.

From memory my 280 has about the same motor power as a 290, and it can take deeper cuts with 12mm tools than I'm comfortable with. I like to run it so it sounds as if the motor is working hardish, not being flogged. Although the motor and VFD are both fan-cooled, it's still a hobby lathe.

Dave