Member postings for Gwil

"How It's Made" is often interesting to watch, but the narration is often rubbish. The script sounds like it was originally written in French (the show is made in Canada) and then clumsily translated into English by someone who hasn't watched the film and doesn't understand engineering anyway.

There is an interesting description of the accuracy (or not) of getting holes in the right places using marking out and co-ordinate methods, and also jigs, starting on page 20 of Tubal Cain's book 'Simple Workshop Devices'. (The message I took from that is that drilling is a distressingly unreliable business.)

This may be useful:

**LINK**

There is also a revised edition but that may only be available secondhand.

Edited By Gwil on 14/10/2016 12:36:51

Hi Muzzer

What I meant was that constant current drive is the most efficient approach if you have a number of LEDs to drive and you have a high enough supply voltage (e.g. 12V) that you can connect a number of them in series, leaving just a few volts across the regulator. Under these circumstances I'm not sure that a boost converter with its own losses followed by a constant current regulator would end up being any more efficient, and it would be more complicated. However I haven't done any real world calculations on this arrangement.

I should probably have mentioned constant current drive (like Michael G I was starting from basics) but now others have dealt with it. It is certainly the most efficient approach if you have a high supply voltage to take advantage of.

LEDs behave differently from incandescent bulbs - a small change in the voltage across them means a large change in the current going through them. It is always necessary to control the current somehow, usually by making the supply voltage a good deal higher than the voltage across the LED(s) and adding a resistor in series. For example to run a single white LED (3.5V forward voltage) at 20mA from a 12V supply one needs a resistor to lose 8.5V. Ideally that would be 425 ohms, which unhelpfully isn't a standard value but 470 ohms is standard and would only reduce the current slightly. Two LEDs in series would have 7V across them leaving 5V for the resistor (250 ohms ideally this time), but you have to be aware that the smaller the voltage across the resistor the more sensitive the arrangement is to variations in supply voltage and LED voltage, which can change with temperature. Series and parallel wiring works but you always have to consider how the current is being controlled.

Rummaging about in my computer I found another shot I'd taken (not that it adds a great deal, but here it is anyway) -

Edited By Gwil on 03/08/2016 17:09:17

As requested, some pics of the SMEE model wall engine:

Very interesting. Thanks for the link.

A question for Jason - you say "I just use mild steel to pin mine..." but the photos seem to show something filling the joints, although I can't tell what it is. Could you clarify please? Thanks.

Roll of (soft) toilet paper kept next to the lathe or other machine. Use like mini kitchen roll to wipe off oil, swarf,, blood etc.

I've just realized that this thread should probably have been started in the 'Discussion about articles in Model Engineer and Workshop section'.

A couple of thoughts which may or may not be worthwhile -

(1) I don't think the test described on page 38 actually checks that the spindle is perpendicular to the table, only that the table is flat - the spindle could be cocked over at an angle and the clock reading would still not waver as the table was run to and fro. To check for perpendicularity you need to mount the clock on an arm at a fixed radius, pointing downwards, and swing it round to different positions.

(2) Obviously the LH x-axis handwheel shouldn't be rubbing against the endplate. However, "...the right hand side of the x-axis leadscrew had the same dual thrust bearing arrangement as the AMA25LV..." (page 39), apparently to cope with thrust in both directions, so it isn't clear to me why the LH end would need a thrust bearing as well. After all, two thrust bearings at one end work for the y-axis. Maybe this is why the makers omitted the extra bearing on the AMAT25LV.

Hi Daniel

Don't know how relevant this is, but anyway -

OK, the pivot screws into the cylinder. Has the pivot got a shoulder, larger in diameter than the threaded part, turned on it? If it has, and the shoulder doesn't seat squarely against the cylinder, then there is something wrong with the hole. If there is no shoulder and the design relies on the thread coming to an end to locate the pivot, there is a good chance that it may end up cocked over anyway (unless the thread is a very close fit) even if the hole is perfect. Another thought - the hole may be perpendicular but was the axis of the tap when it went in?

Just thinking aloud ...