What did you do today?

[wierdscience]

Did up a new driveshaft for the hydraulic pump on a friends crane truck. Normally there is something left of the old shaft that can be reused. Not this time,both companion flanges were shot, the slip yoke was shot, the tube yoke was shot and the splined end was missing it's poly. So $265 worth of new Spicer 3150 driveline parts and an hour later.....

[danlb]

I spent the afternoon chasing a wild goose. It started with a cable modem that was randomly disconnecting from the internet. The cable company tech asked me to remove and attach the co-ax at each junction. Nothing he suggested did anything to improve anything, so I let him go. I hooked up the main house router to the DSL feed since I'm going to be out of town for a bit. Two hours later the wife tells me that both Tivos have a banner saying that they can't reach the internet. Then they say that they can't reach the DHCP server.

4 hours later I finally realize that the Tivos are connected to the main router via a nicely hidden switch that I can't reach. I eventually trace the problem back to the ethernet cable from the main router that was plugged into a patch panel just below the co-ax. It apparently made a bad connection when I jostled it for the first time in more than 15 years.

Sigh. All is well again, other than the cable modem that I will have to attack next week.

Dan

[Bented]

Old school parts for a 100 year old paper manufacturing plant. Someone produced this drawing from a known good sample, as is often the case with old parts the "as measured" dimensions are close to 64th" fractions as was common practice in the past.

The 2.0472-2.0468" diameters are 52 MM press fit bearing bores, this is a known dimension, the intended ideal of the other dimensions are unknown and include any acceptable errors.

Much to my delight I dug around in the ring gage drawers and lo and behold there is a 2 5/8-12 gage. Was not looking forward to measuring 20 threads over wires.

[Bented]

I finished the first side of 11 parts yesterday morning, at around noon the tool turret on the lathe began to misbehave.
After much fussing with the turret without resolution of the problem I wisely walked away from the job until Monday.

They had a decent surface finish for HRS stock.

Only a few more to go.

[Cenedd]

Did something that'll probably set a lot of your teeth on edge. I'm working with some pre-bent borosilicate glass tubing at the moment to watercool a PC. Necessary? Well, no but that really isn't the point. Most of the cutting (there's only 5 cuts so no point over-tooling) I've been doing with a Dremel and a diamond wheel as the cut is usually too close to the end of the tube to score and snap. This one, once I'd cut the first side it was short enough to chuck and spin so I did that to score and eventually got it deep enough to snap - this is 16mm OD tube with 2.5mm walls so it's quite tough! VERY unforgiving of dimensional mistakes as I'm finding out - not much flex in a solid glass tube!

Caveats before I get flamed: full face mask with P3 filters was worn while doing anything that may produce glass dust. Lathe was not run under power, just turned by hand and the tool (60° carbide threading tool) advanced very slowly - in the order of 0.01 to 0.02mm DoC per rev.

More pictures of it (partially) assembled here if anyone is interested.

[Bented]

Excellent work, how many failed before the process was successful?

[Cenedd]

Excellent work, how many failed before the process was successful?

None First score I did have to re-chuck it a few times and score deeper before it'd snap by hand. Ended up feeding in about 0.3mm in the end. The Dremel was a little less accurate (user, I'm sure) with multiple shallow plunges eventually meeting up in a circle. It eventually fractures by itself well before completely cut through all round. Despite being more ragged, I could clean the edge up with the side of the wheel and then roughly chamfer it by spinning it with the wheel at an angle. Finally some 120 grit wet and dry and then a quick finish with 400.
I think actually that the leg I thought was a mm too short might actually be ok.

Loved parting and chamfering acrylic tube - which is also used but I was just mocking up before using the one-shot at glass I had. Much nicer and neater than hacksaw and plumber's chamfer tool. Might have to use that for my son's when we get round to that.

[Bruno Mueller]

Yesterday I have made a little thing. Firing pin for a Sig Sauer 225. Cold work steel, turned, hardened and tempered.

[Bented]

None First score I did have to re-chuck it a few times and score deeper before it'd snap by hand. Ended up feeding in about 0.3mm in the end. The Dremel was a little less accurate (user, I'm sure) with multiple shallow plunges eventually meeting up in a circle. It eventually fractures by itself well before completely cut through all round. Despite being more ragged, I could clean the edge up with the side of the wheel and then roughly chamfer it by spinning it with the wheel at an angle. Finally some 120 grit wet and dry and then a quick finish with 400.
I think actually that the leg I thought was a mm too short might actually be ok.

Loved parting and chamfering acrylic tube - which is also used but I was just mocking up before using the one-shot at glass I had. Much nicer and neater than hacksaw and plumber's chamfer tool. Might have to use that for my son's when we get round to that.

Enlighten me, what material attributes make glass coolant lines preferable to say acrylic or polycarbonate tubing?
Are the cooling fluids highly corrosive, must they be sterilized between uses for instance?

Do you use ammonia based compounds as coolant?

[Cenedd]

Well, if you're asking me to justify the effort and expense, I can only do so as well as I could for my lathe, mill...and half the tools I own
Coolant can be just distilled water but you can get algae growing unless you add biocide or a coil of silver. You can also get corrosion - galvanic corrosion if you have nicely dissimilar metals in the loop - so an anti-corrosion agent is wise. It ends up that most people buy a pre-mixed coolant off the shelf. You want both the above plus conductivity as low as you can get....just in case People also use coloured coolants but apart from the fact that some can settle out and clog the tiny fins in the water blocks (equivilant of heatsinks) a friend of mine once wisely said "Which colour will look best when it leaks everywhere and stains your carpet?!".....so I've been using clear and colourless ever since.....and proving him right too!

Mainstream options are

• soft hose - easy to work with but can have incompatibilites with some coolants. Also can discolour over time.
• acrylic tube - looks nice (if done well), doesn't discolour, doesn't really have incompatibilites that I'm aware of but it can fracture especially if you have the thin walls that make it easier to bend.
• PETG tube - easier to bend, not as brittle but also slightly absorbant so can discolour as the coolant soaks into it. Not compatible with coolants with glycols in. Starts to go soft at about 40°C so not a huge amount of safety margin.
• Glass - fairly new, not affected by any coolants, doesn't discolour, the clarity is much better than any of the other options.....but not the easiest to work with. Mostly for straight lines with fittings for bends but you can get the pre-bent sections I tried a couple of. It seems to be possible (according to YouTube at least) to bend it with a blowtorch with some practice but I haven't got one and the accuracy required would have made that difficult too.

The choice is largely down to aesthetics to be honest. Watercooling is much more efficient and means you can run hotter things (CPU and graphics card mainly) without quite as much noise as air cooling would make. It also keeps them cooler so they tend to live longer. The flip-side is more cost (although you can re-use, of course) and always the risk of a leak - although there is a new product that's interesting that lowers the pressure of the whole system below ambient pressure so you can actually drill holes in things (as a sales pitch) and it still won't leak.

I mainly went watercooled for the quiet for the quiet - my rig is almost totally silent. It's far more over-the-top than it needs to be partly for the hobby side of it but also I decided I wanted to build something a bit more unusual that I was proud of. It's also what led me into machining originally so it's nice to revisit the project and put some of the tools that I now have back to the use that caused their existence.

[Bob Engelhardt]

... dimensions are close to 64th" fractions as was common practice in the past.
[snip]

A 1/64 is 0.016... (+- 0.008). Yet it's dimensioned to 0.001 - dimensioned to 0.01 seems more reasonable. Or is the tolerance +- 0.008?

[Bented]

Old part VS new drawing.

The tolerance on the new drawing is +- .005" unless otherwise specified.

The original may easily have been +-1/64" in the pre-cad era.

It is very easy to draw a part in cad leaving the dimensional settings at .000, a feature on a part that will not effect its performance in use if it is +- 1/8" is still dimensioned .000.

If one only has a drawing to go by then every single dimension is important.

Individual dimensions often have to be manually changed.

I can draw this way all day long, set the linear dimensions to 8 decimal places.

This benefits no one.

[Bob Engelhardt]

Old part VS new drawing.

The tolerance on the new drawing is +- .005" unless otherwise specified.

[snip]

OK - so the tolerance isn't implied by the number of significant digits of the dimension.

Another question - how do you locate the 0.56 radius without knowing one of these points? Do you know the overall diameter from a Note? That would give you the lower point.

[challenger]

Painted our great room.
Yeah, that sucked.

[Bented]

OK - so the tolerance isn't implied by the number of significant digits of the dimension.

Another question - how do you locate the 0.56 radius without knowing one of these points? Do you know the overall diameter from a Note? That would give you the lower point.

The OD is 4.304", using the start point in Z and the finish point in Z that maintains a .560 radius and the diameter is more then good enough.
As an added bonus the customer has no method of measuring such a dimension short of contracting a 3rd party vendor to do so.

Cardboard factories rarely have an in house metrology lab.