Member postings for Andy_G

The fundamental problem they face is that the LASER beam is affected by thermal gradients, turbulence and convection currents in the air. (Even having somebody stainding near it is enough to cause a beam deviation of several hundred microns over a few metres).

I designed & built a LASER guided cutting tool in the late 1980s (using a quadrant photodiode as the target). It was easily capable of micron level resolution, however the beam deviation due to air currents within the machine enclosure rendered it useless.

Hi Russ,

If you search on the likes of Aliexpress for "3D touch probe" there should be quite a few options - probably VAT to be added to the headline ~£40 figure.

I will add a link to an example, but suspect it might not be acceptable to the Mods (mods: in which case, please delete it)

https://www.aliexpress.com/item/1005004670392373.html

Too late to edit the above, but you may also like to take a look at bNC which (I think) will run on the Rasberry Pi to use in conjunction with your GRBL controller.

I haven't used it, as it requires python to be installed on a Windows machine.

https://pypi.org/project/bCNC/

Edited By Andy_G on 28/03/2023 15:49:03

Interesting - I was idly wondering about 3D digitising and had had a scout around for suitable software without really finding much. I hadn't come across GrblGru though, so thanks for that!

(Decent looking 3D probes are available from the Chinese direct sellers for about £40)

I also looked at the '3018' type machines, but wanted some (modest) metal cutting ability. I eventually came to the conclusion that I'd need to upgrade / replace every component before I'd be happy with it, so I set off making my own, thinking it wouldn't cost much more - be careful of getting drawn in!

I (foolishly) added up what I'd spent on it - best part of £1000

Edited By Andy_G on 28/03/2023 15:01:39

SHX files are Autodesk font files.

This suggests that the drawings include text or symbols from a font other than one of the mainstream Autodesk offerings. (Maybe inadvertently!). The missing fonts *should* get replaced with one of the defaults after the warning.

I haven't used Autocad for more than a decade, but I have vague memories that there used to be an option to embed the fonts with the .dwg file (tick box or similar) - Do NanoCAD / SE have a similar option?

If you take note of the .SHX file name it should be possible to copy the file from the source computer to the destination computer.

Some more information here

Edited By Andy_G on 26/03/2023 21:11:48

I'd be interested in this, too - happy to PM, but it might be of general interest?

Truly outstanding!

Very well done.

#1 Sort out the gib strip in the compound slide

#2 Bolt it to something solid

#3 Indexing holes in the chuck backplate & a toolpost spindle.

I watched this youtube video last night and thought that it contained a lot of good advice and common sense that doesn't normally get an airing:

No, the lack of rigidity was in the spindle itself - the frame of the machine is ~6kN/mm measured near the spindle, but the original spindle itself  was only ~0.8kN/mm, and the re-worked spindle was ~2.5kN/mm. I'm in the process of building a new spindle which I hope will be stiffer. It's a "rabbit hole"

Edited By Andy_G on 13/02/2023 01:20:40

Before you go too much further, work out what RPM will be required for the materials / tools you want to use (from published cutting speed information) - High RPM may be more of a curse than a blessing.

For example, if you want to cut steel with a 6mm diameter tool, the maximum spindle speed is ~3000 RPM for HSS and ~8000RPM for carbide. I think that both of these are below the minimum RPM for the first type of spindle.

First off, I would say "just buy a mill". If you ignore that, then I would steer you towards the second type of spindle for a mill.

I built a router type machine (CNC) with a cheap, air-cooled spindle motor, and it was nowhere near rigid enough.  After some re-engineering, it will cut aluminium (fairly happily) and steel (not so happy). It is still not very rigid and I'm in the process of making a new spindle along the lines of your second option.

Cutting test for material removal (it runs a bit happier with lighter cuts):

Good luck

Edited By Andy_G on 11/02/2023 10:01:23

Given that you're machining 304 stainless, I would start slow (with plenty of the best cutting oil that you own). With such a fine tool, the natural tendency is to be tentative with the feed and, on stainless, the result is likely to be that the tool rubs, hardens the surface and then, when you apply a little bit more pressure, digs in and breaks.

FWIW, my advice would be:

- Run slowly (a few hundred RPM, or less);

- Plenty of oil;

- Feed steadily 'through' the surface of the part (I.e. don't wind the tool until it touches the surface and then start feeding);

- When you see the curl of swarf appear keep feeding steadily until you're at your target depth. DON'T STOP.

On the tool: Make sure that the tip is the widest part - it looks from that photo that the tool may have tapered towards the tip (although the actual tip has gone) - you don't want the tool to wedge itself into the slot it has cut, or it will jam and break.

Difficult to tell from the photos, but it looks like there is quite a large relief angle on the front of the tool - if so, reduce this to the minimum possible (5° maybe?) to better support the cutting edge.

I am a heathen, and rough out any HSS tools I need with an angle grinder - much quicker than a wheel - and sharpen them on a belt sander. That said, 3/8" is a big chunk of HSS to start with if you want a 0.3mm wide blade. Broken centre drills, etc. will need less grinding, but you will need to make holder for them.

Exhibit A (set up for internal grooving):

(3mm grooving tool facing the camera; 1mm from a previous job on the other end.)

My 2p, anyway.

SOD's maths are out by a factor of 60

(200kHz at 25k ppr is 480 RPM)

Higher microstepping divisors can help with smoothness and noise reduction, and might have some use in getting smooth rotation at very slow speeds, but I also think that the highest numbers are overkill.

If the DS18B20 reading is initiated by the software in the Arduino, then the problem surely lies in the software?

SoD gives a reasonable list of possibilities, but specific to the Arduino - if relying on the 'millis()' function to determine delays, the counter overflows every ~49 days (for a 16MHz device) which can result in unintended consequences. Your problems occur more often than that, but something similar may be happening - do all your timing related variables have sufficient precision? For example: storing time related values in a LONG type variable could result in overflow in ~24.5 days (should be UNSIGNED LONG).

Once the software is off it's intended course, the results could be anything - e.g. failing to zero a variable resulting in values accumulating and giving a false reading.

Sould the numbers I have put in bold above increment with each reading?

If so, the problem would appear to be in the software, rather than the sensor.

I agree - I'm pretty sure it's to mimic the old 'half stepping' drives that were in use before PWM became common.

In any case, a genuine DM556 would not have this setting.

As an aside, I'm pretty sure that isn't a genuine DM556 driver - it looks just like all the ones badged as TB6600, etc.

The real DM 556 doesn't have the 2A/2B modes and goes up to 128 microsteps -

DM556 driver manual

'Generic' TB6600 etc example

I believe that the *real* TB6600 chip (42V / 4.5A) only goes to 16 microsteps - these all go to 32, and are therefore "fakes" based on some other chip. Like John, I have heard of problems with these, where the rating of the actual chip used is less than the sticker on the case.

Interesting article here:

I only found out about this *after* I bought some (very cheaply!). Touch wood, they have been performing OK on my little CNC at 36V and ~3A.

(A happy New Year to all, btw! )

The difference is in the waveforms used - have a look a pages 18 and 19 of the TBA 6600 chip datasheet:

**LINK**

I can’t see that there would be much practical difference, but one mode may reduce motor noise or be less vulnerable to resonance problems than the other in a specific application,

I suspect that the issue might be that that the '3 O'Clock' (looking at the lever) hole has wallowed out a bit, and shouldn't be able to align with the detent ball in this position - there is visible wear in this hole in the original photos (the hole is at 9 O'Clock when looking at the lathe). If the detent is '1/2 a hole' out, it shouldn't engage.

How well do the balls fit in the holes in the handle? If they have room to wander, they will be able to engage in the detent holes sooner than they should. (If they are loose, the holes in the handle may be able to be reamed out to take a slightly larger ball, or a small, round ended dowel - made to suit.)

My 2p, anyway.

Edited By Andy_G on 30/12/2022 10:28:56

The bushes will be oil impregnated, so I doubt that glue would stick to them.

The bushes in that picture look like sintered ones to me.