Member postings for Andrew Johnston

Good grief, SoD must be really bored!

Andrew

I wondered about that too, or if the work was being curved upwards due to pressure from the vice. But the OP states he had the same effect on a 40x6 part, so presumably not?

Andrew

Well, that's an interesting conundrum. Never come across it on any of my milling machines. Tool deflection does happen, but it's small, especially with the light cuts being taken. And, as the OP says, tool deflection would cause the bottom of the work to stick out, not vice versa.

Feeds and speeds don't seem unreasonable, although we don't know if the cutter is HSS or carbide. If the cut is unstable then it should generate noise/vibration - did it?

Where's SoD with his vibration monitor when you need him?

Andrew

Not necessarily. Wire drawing occurs when steam is throttled and the input and output velocities are similar. This means that the expansion is isenthalpic, ie, enathalpy stays constant although the pressure drops. Since the pressure has dropped but the enthalpy stays the same the output steam is drier, or even becomes superheated. That may be an advantage as condensation in the cylinder will be reduced and hence less energy lost due to condensation and re-evaporation.

With Stephenson's valve gear, at short cut offs, the input ports are only partially open so wire drawing is almost certain to take place, which may, or may not, be an advantage.

Losses due to poor passage design and sharp changes of direction are different.

Andrew

I use option 2.

Andrew

That is interesting, you learn something every day; thanks very much.

It works fine in my old version of VisualMill:

I had in mind that using a facing operation with bosses would overcut the boss; but obviously that's not the case. I think it arose from my experiments with pockets that are open on one side. Pocketing doesn't properly finish the corners that meet the open edge. But facing overcuts the three sides that are not open. I'd better go back and do some tests.

Andrew

Definitely selecting a surface and the selection is called a FlatArea. But the system still highlights boundaries, so may not be really selecting a surface. Same result on a much newer demo version. Looks like it's a limitation of my CAM.

Andrew

Pocketing doesn't work in my CAM software, which is why I didn't suggest it:

It sort of does what is sensible, ie, keeps within the outer boundary and outside of the inner boundary, even if it isn't what is wanted. Presumably Fusion360 has extra knowledge of the part geometry and knows it can go outside the outer boundary?

Andrew

It's one of those seemingly simple jobs which are a PITA. If you trying facing the system will machine the outside perimeter with no problem but will overcut the boss. Conversely if you profile between the boss and outline it won't completely machine the corners of the outline.

A simple but very inefficient solution is to profile the boss but have a number of 8mm stepovers such that the whole of the blue area is covered.

Alternatively one could create a rectangular outline more than 10mm from the outside of the blue area and profile between the rectangular outline and the boss.

Andrew

There seem to be three fundamental issues. Roughly in order of importance:

The new cutter shown seems to be poorly ground. The used cutter shown is badly worn with the cutting edges rounded. It would seem that these cutters are simply not up to machining steel.

Runout on the collets - it's way bigger than the chip load. So one tooth may not be cutting while the next will be taking double or more the expected cut. So the loads on the cutter are much higher than expected from simply looking at the rpm and feedrate. Varying chip load is also a recipe for vibration and chatter.

There is still quite a lot of cutter stick out. even with the shank cut down.

One and three could be ameliorated by buying quality cutters intended for cutting steel. Two is more difficult, short of acquiring a different mill. A good clean and check for any small burrs may help. Reducing the stick out of the cutter would also help.

Andrew

It's single enveloping, ie, the worm wheel wraps partly around the worm giving line contact. Double enveloping is where the worm also wraps around the worm wheel. Which means that the thread on the worm is cut on a circular path on a plane which also contains the axis of the worm.

Andrew

Man, that's hot. Higher than Heathrow reported today I think. One normally expects airports to be slightly hotter than the surrounding areas due to all the engines pumping out hot gases. But not at the moment. I went down to Stansted yesterday to renew my Civil Aviation Authority medical. Never seen it so quiet. Very few cars and no buses around the terminal building and the long term car park was empty.

Andrew

Assuming free hobbing (hob and blank are not geared together externally) then it depends upon if you need a set number of teeth. One can free hob using a tap and a blank with no gashing:

But it's a crap shoot as to how many teeth you end up with; in the case above it didn't matter. If a set number of teeth is important then you need to gash the blank at the helix angle of the worm -

- to give the hob something to engage with and drive the embryo worm wheel:

The DP of the worm wheel shown was approximately 6.281, although I used a commercial 6DP involute cutter for the gashing. The hob is home made.

Andrew

A quick visual comparison with the rule in the picture will show that the shank is 3mm.

My apologies for not reading the thread sufficiently well to realise that an expert had already expounded. At least crawling back into my cave will be some relief from the heat.

Andrew

The basic rule for worms and worm wheels is that anything goes. But it is common to specify the worm wheel in terms of a standard DP or Mod number. The downside is that the tpi/pitch of the worm is then an irrational number. On my lathe one can modify the change gears so that the gearbox will screwcut worms to approximately the correct tpi or pitch for standard DP and Mod values.

Andrew

Here's a picture of a 2mm diameter 2-flute uncoated slotdrill:

Note that the working length of the cutter is quite short. I'd be running this with the collet closing at the top of the flutes, so less than 10mm stick out. I used this style of cutter to form a series of fins on a liquid cooled heatsink for an experimental high power inverter. The gap between each fin was 2mm and they were 4mm deep. I ran at 4000rpm, full width, 1mm stepdown and 150mm/min feed in 6082 aluminium. Each heatsink had 18 gaps and I machined 8 heatsinks, all with the same cutter.

Andrew

Andrew

There's no polite way of saying this, but the finish shown is awful.

The pocketing on both sides of this steel part was done with a 3mm cutter - no deburring needed:

However, the reason that the cutters are snapping is waaaaaaaaay too much stick out of the cutter. Looking at the finish on the flat parts there appears to be chatter, caused by the excessive stick out. In due course I'll take a picture of one of my 2mm cutters for comparison.

Andrew

I looked up a previous CAM file using a 2-flute carbide endmill cutting 316 stainless steel on my CNC mill. I was running at 3500rpm, 70mm/min feed and DOC of 0.425mm. Nothing special about the DOC, it just gave an integer number of steps down for the thickness of material. This was done 10 years ago, now I'd be at 5000+rpm and a commensurate increase in feed. With a 3-flute cutter, full width and DOC of 0.4mm in low carbon steel I'd be running at 7800rpm and 100mm/min.

All the above is with quality cutters, Guhring and YG.

It's a funny thing, but I find that industrial feeds and speeds (from YG) work very well in my home workshop on my hobby CNC mill. Although of course the cutters don't know what machine or environment they're running in. But with the caveat that it only applies to small cutters. I don't have rigidity or horsepower to drive large (>20mm) cutters at their recommended feeds and speeds.

Andrew

Finished the valve chest cavities:

The LP cavity is nearly 2" deep so a fair stick out on the 10mm endmill! All machining was done by numbers, X and Y from the DRO and Z using the handwheel dial. It's pleasing that finished dimensions are within a couple of thou. The holes for the valve chest cover studs were drilled and tapped using the DRO on the vertical mill while the holes in the cover were done on the CNC mill. Fortunately the cover fits, studs are 1/4" BSF, clearance holes were drilled 6.4mm.

Next task is to finish bore the cylinders on the horizontal mill.

Andrew

Boxford didn't make a 280 lathe, but Myford did.

The fitting is Camlock, so the spindle doesn't have a radial register or threads. The Camlock system primarily mates on the outer taper. Measurements on my lathe spindle Camlock give a runout on the back face of 0.01mm and a runout on the outside taper estimated to be 0.002mm. The inside of the taper plays no part in mounting the chuck so it's runout, or otherwise, doesn't tell one anything.

The runout on the back face seems a lot to me, but it would be a bad idea to try and machine it. I think Camlock parts are hardened which makes life more difficult. I'd agree wth JasonB, the best bet is to machine the chuck backplate in situ. But make sure you mark it so that it can be replaced in the same orientation each time, otherwise it won't run true.

Andrew