NSW 'B' class mogul

[David]

The cutout in the cylinder was done on the manual mill, so I was able to take it easy. I just wish I'd thought to put it further down in the vise jaws and not at the V!

A good cut you're getting. I only have cheap 4 flute, non centre cutting HSS endmills, aside from a few 6mm 4 flute non centre cutting no-name carbide ones. My endmill situation is a bit poor ATM for CNC.

[jasonb]

The lack of ctr cutting may be why they complain on the plunge down after a tab though you can set it to ramp although that means more to clean up later.

I don't have many top brand cutters just a few Dormer and YG, On the manual mill most cutters below 6mm and 1/4" are the HSS "throwaway" FC-3 type often from Hertel which cost about £4GBP which are good general purpose cutters for both slotting (full width) and edge cutting. I also tend to use 3-flute ones for initial clearing on the CNC unless it's aluminium then I tend to use the two flute ali specific ones.

On the CNC it's mostly carbide, a few New Centuary which is YG's chinese factory but more from ARC's premium range which Aussie do down your way or similar quality cutters from APT as they have a bigger range so you can get things like corner rounding or extended shank lengths, cost a bit more than the direct "take a chance" from China but still a lot less than the high performance ranges with typical cost being £5-6.GBP for 2-6mm diameters.

One reason you may not be having much luck with adhesive holding is that a blunt cutter will put a lot more side force into the work rather than cut easily through it, this also generates more heat in the part which will soften the adhesive.

[David]

I think this will do it for silver solder on the air cylinder. The remaining parts will be threaded, soft soldered, or glued on.

After watching Joe Pi's series about making a stationary engine I took note of how he held the crosshead casting and made a little jig that allowed me to hold the cylinder with the flanges upright and at a known distance from the rear vise jaw. It worked well enough, although not spot on - I moved the hole about 0.2mm from where the DRO and my calculations said it should be when it looked off at the start of drilling. A 6mm endmill opened the hole out and nothing tore or otherwise went south.




It's all fluxed up and a rusty steel ring is holding the cylinder top so it doesn't move when the solder melts. I added a little very thin silver solder but I think most of the work is the current stuff reflowing.




Pretty pleased with how it is going so far.




I put some small head 10BA screws into the top flange then cut and sanded them level with the bottom of the part. This will either be glued on, or two little cylinders that go through the holes in the wings will thread into the cylinder. Those holes unfortunately coincide with where the ribs come through the top of the cylinder so drilling and threading may not be an option due to the edges of those holes.



I think that's the most difficult bits of the Westinghouse pump done. The top steam cylinder is simpler than this with just an offset through hole to confuse matters, but that should be easier than these ribs and bosses. The shuttle valve has some ribbed end covers but again, it shouldn't be too bad. I might try doing them on the CNC for laughs, and fall back to the traditional way if necessary. They'll need a mighty small endmill on the CNC.

[Jim]

You're doing a cracking job there David creating what will be a gem of pump that you can proudly say "I made that."

Jim

[Roger]

Nov 9, 2020 9:33:30 GMT David said:

I think this will do it for silver solder on the air cylinder. The remaining parts will be threaded, soft soldered, or glued on.

After watching Joe Pi's series about making a stationary engine I took note of how he held the crosshead casting and made a little jig that allowed me to hold the cylinder with the flanges upright and at a known distance from the rear vise jaw. It worked well enough, although not spot on - I moved the hole about 0.2mm from where the DRO and my calculations said it should be when it looked off at the start of drilling. A 6mm endmill opened the hole out and nothing tore or otherwise went south.




It's all fluxed up and a rusty steel ring is holding the cylinder top so it doesn't move when the solder melts. I added a little very thin silver solder but I think most of the work is the current stuff reflowing.




Pretty pleased with how it is going so far.




I put some small head 10BA screws into the top flange then cut and sanded them level with the bottom of the part. This will either be glued on, or two little cylinders that go through the holes in the wings will thread into the cylinder. Those holes unfortunately coincide with where the ribs come through the top of the cylinder so drilling and threading may not be an option due to the edges of those holes.



I think that's the most difficult bits of the Westinghouse pump done. The top steam cylinder is simpler than this with just an offset through hole to confuse matters, but that should be easier than these ribs and bosses. The shuttle valve has some ribbed end covers but again, it shouldn't be too bad. I might try doing them on the CNC for laughs, and fall back to the traditional way if necessary. They'll need a mighty small endmill on the CNC.

That's a great result, you're right to be pleased with that.

The finishing cutter size on a CNC machined version is determined by the corner radii, and I can't see why that couldn't be at least 0.5mm if not more without it looking wrong. I suspect the casting has generous corner radii.
The shape of the cylinder means the cutter doesn't need a very long flute length, so even a 1mm ball nosed cutter would be pretty robust. I'd rough it out with a 6mm, leaving say 0.6mm all over, then do a horizontal finishing pass with a 2mm slot drill, leaving say 0.2mm to go. Then just finish off with a 1mm ball nose and Parallel Finishing with about 0.1mm overlap and 150mm/min. Just start it off very slowly on the first X/Y pass and then leave it to get on with it. It could take several hours to trace over that with a 1mm ball nose, but who cares!

On the business of cutters that don't cut to the middle, they're of little use on CNC machines unless you're doing a simple side entry. For a top entry, they're just not suitable.

[David]

4 cylindrical bits have been made and soft soldered on. The bottom part, sitting beside the cylinder has 6 10BA small headed hex headed screws in it. Before soldering the bits on I deliberated on where the 10BA threaded holes were relieved to allow the screws to seat properly... and mis-thought about it. After screwing them all in, trimming them, and filing them flat I realised I'd got it wrong so had to take them out, relieve the other side of the threaded hole, and try and get tiny screws with 1.5mm of thread back into their holes. I blame it on the fact I was doing this about about 7:30am and was tired.



The slots on the gland were cut with a 1mm endmill in a hex collet block. The manual mill can't go anywhere near fast enough for these so it was 0.25mm deep cuts hand fed very slowly. I have to make another one for the top cylinder. I'm not convinced these will be noticeable after painting but I changed the model to have 2mm wide slots before removing it from the collet block and they looked too big.

[David]

I've made the top cylinder and it's gland, but not the shuttle valve or steam inlet/outlet.

Mistakes were made turning the cylinder, but recovered from ok.

I don't have a 3/4" hole saw and didn't feel like trying trepanning on the lathe, esp since I can't hold the part securely in the jaws without marring it. So I used the CNC mill as a very expensive rotary table. I spotted and drilled the dummy bolt holes and then drilled over 30 1.4mm holes around where the slot for the spacer goes.

The drill is in the hole for the steam inlet/outlet and is used as a visual guide to try and orient the cylinder how I want it. It isn't critical so by eye is good enough.



I used a 1.5mm endmill, 5000rpm, 60mm/min, 0.25mm depth, 10 passes. I stopped between each pass to blow out the chips. I had a program that did a single pass and kept adjusting the Z reference by 0.25mm so the program ran 0.25mm lower each time.



You can see some runout in the slot. It was all done on one setting with the same program so I don't know where that came from.

But the spacer fits fine.



That all went well, so it was time for a stuff up! I turned the part over so I lost the X reference at least. I didn't trust the Y reference any more either so I tried to redo that but couldn't get the edge finder to the front of the part because it was below the vise jaw top and I only had 3mm copper as a soft-spacer. I took the rear at X=0, then put what should be halfway into the DRO. I checked it by sending a drill down at X=0 Y=0 and for some reason didn't think it was centered and doubted myself, moving the Y axis by 2.5mm, the radius of the edge finder. Spotted the dummy bolt holes on the *top* of the cylinder and they were off, and visible! Luckily they're not very deep and I must be able to fill them with epoxy or something.

I've tapped all the holes and put the bolts in for looks but no photo of that because my son has borrowed my phone for 3 days.

[Roger]

Hi David,
That looks like it came out plenty good enough. I'm curious to know why you drilled all of those holes? A 1.5mm cutter is reasonably robust, and it's only Brass, so I'd have roughed and finished it with the 1.5mm cutter. Ok, it takes a bit of time, but you don't have to stand and watch it.

Personally, I wouldn't set up anything round with an edge finder. If you haven't got something aleady, I'd take time out to make something to allow you to attach the clock to the spindle so you can spin that round whatever you're trying to clock up. It's a much more reliable and accurate way in my opinion and well worth the time doing. It means you can also clock up on any hole if you lose position and then tell the DRO you're at whatever that hole location is. If you don't know where it is, create a CAM output using the same zero with just that one hole in it so it will tell you the centre position. Then you can do a quick sanity check in the air to make sure it looks plausible and carry on as before.

[David]

I *do* have to stand there and watch it because every time I don't I seem to come back to a problem! I drilled all the holes because thought it would make life easier for me and the cutter. It didn't take long. I did that slot in separate passes so I was standing there, blowing out the chips between each pass etc.

I would like one of those spindle indicator fixtures, plus a few other ones. But an edge finder is plenty accurate enough for most things I do so long as I use it properly. I could have sat the cylinder on a parallel to get the top lip of it above the vise jaw but didn't think of that in time.

One thing that puts me off making the indicator holders is the little dovetail. How do you make that - just file it? Or do you grip the indicator some other way?

[Roger]

Nov 18, 2020 22:34:59 GMT David said:

I *do* have to stand there and watch it because every time I don't I seem to come back to a problem! I drilled all the holes because thought it would make life easier for me and the cutter. It didn't take long. I did that slot in separate passes so I was standing there, blowing out the chips between each pass etc.

I would like one of those spindle indicator fixtures, plus a few other ones. But an edge finder is plenty accurate enough for most things I do so long as I use it properly. I could have sat the cylinder on a parallel to get the top lip of it above the vise jaw but didn't think of that in time.

One thing that puts me off making the indicator holders is the little dovetail. How do you make that - just file it? Or do you grip the indicator some other way?

Hi David,
A dovetail cutter can be made by turning the required shape on the end of a piece of Silver Steel, machining it in half and hardenind and tempering. It doesn't take much effort to make. I machined a slot first then finished it off with the cutter.
I've also made a fine adjustment arrangement which makes it much easier to use. By all means PM me with your email address and I'll send you a link to the design. I can output the parts as a .stp file so you can load them into Fusion360. It's a CNC friendly design which can mostly be machined from one side.

[jasonb]

Most lever indicators come with a couple of fittings to go onto the body's dovetails and then the holder just clamps onto that

I don't think I would have predrilled it, just done it with a constant ramp down blowing out as it went round and round

[David]

Here is the bracket that holds the lubricator on. Nothing to report with this, it went as planned.



The other bits you can see are part of my continuing efforts to make the pump out of as many bits as possible. Here they are on the machine and you can see how close I skirted to one of the hold down screws - there was far less than 1mm in it. I also noticed the collet nut on the toolholder barely missed the vise jaw so the second part was machined closer to the center of the screws and a parallel went under the piece of wood to keep the toolholder away from the vise jaw.

These were cut with the same 1.5mm cutter from yesterday, same speeds and feeds and depths. They both came out with virtually no finishing to the edges required.



They are part of the air filter for the pump. One of them was tapped and soft soldered to the cylinder. The other had the holes drilled 10BA clearance. The screws hold them together. I did it like this because in the photos it is clear this is two bits bolted together and I wanted to try and replicate that look. I've chamfered the edge of the one soldered on to make it a bit more notiecable. After I soldered it on, I drilled through it in a couple of places and tapped 10 BA into the cylinder and put some brass studs in to hold it together when the feet get soldered on later. The studs are filed flush and hidden by the top part that is screwed on. The top part still needs a hex soldered to it.





Almost all the bits made so far.

[coniston]

It's all looking very pretty, glad your 'stuff up' wasn't too much of a disaster. I bought one of these, have used it a couple of times and seem sok, probably not as good as a more expensive one but my Mercer DTI fits perfectly. www.ebay.co.uk/itm/Universal-Adjustable-Indicator-Quill-Clamp-Plunger-Dial-Gauge-test-DTI-Holder/253802548142?epid=1407795436&hash=item3b17cf8bae:g:4OMAAOSwB8xfnB0M

Chris D

[David]

Thanks Chris, that's exactly what I'm looking for. I'll see if AU ebay lists them.

Made a few more bits.

* The 'piston rod' goes into a tapped blind hole in the top cylinder and then all the way through to the bottom where a not squeezes everything into place.

* A hex had to go onto the top of the air cleaner/box (not sure what it's name is). I found some brass hex of the correct size and parted a bit off with a little stub on it to fit through the existing hole and soldered it on, then drilled the stub out.

* I remade the very bottom plate (up on the post from 14 Nov) because when I looked at the assembled parts from the side I noticed I made it with the two cylinders that drop down from it are on the centreline of the cylinder, not in line with the ribs. I checked my photos and it seems these cylinders should be in line with the ribs and the plate should have the same asymetric bolt/stud holes as the top plate.

* The top and bottom plates of the bottom cylinder are now held in alignment by one of the rear bolts whose hole has been extended though the cylinder flange. These holes should all go through the flange but the real pump has some odd shaped bolts with semi-circular heads that are too difficult to model so we have nothing there at all.

So the lubricator, air box feet, shuttle valve assembly, and all the pipes and mounting bracket to go now I think.

I'm pretty sure I've made a months work out what other people would spend 3 days on but it looked like something that would reward the extra attention, and it's the only thing I have good close up photos of.

No photos, oldest son lost my phone on the first day of his excursion :( I probably shouldn't have told him not to lose it!

[David]

I've spent the entire day making three small parts - that's why it's taking me 10 years to make this thing!

An absurd amount of time went into making and fitting the little bar on top of the cylinder. Hours and hours. It's been flooded with solder to fill the gaps and try to give it a bit of a casting look.

One of the end-caps of the shuttle valve housing has been done on the CNC. The other side is still going, over 2 hours so far. F360 disabling rapids is not helping and I think was a dog move. Disabling more than one tool should have been enough! The spindle is running a lot longer than it needs to. This side was really quick, the other is taking so much longer because it is 3 times higher and has some curved ribs on it.

The finish in the photo is worse than what it looks like in real life.



Everything so far:

[Roger]

That's a great result David, it's definitely worth the effort when you get a reward like that.
On the CNC front, you sound surprised that a run takes 2 hours. Some jobs run all day, and the longest ever single pass ran into a second day with a pause overnight. Run time is only important if you're up against a deadline or you have to stand over it. Figuring out how to machine unattended would be my priority. It would be much less boring and free up time to do something more interesting and productive.

[jasonb]

This is where running the spindle as fast as possible will help as you can get a faster feed rate for the same given chipload which means the "new rapids" will be as fast as your feed. If the tool does not have to cover such long distances in rapid then that will also help so play about with things like clearance heights and staydown percentages and lengths. If using adaptive clearing then think about setting it to cut both ways then at least you are removing metal on the way back to the usually heavier climb cut. .

I had a look at a part I had cut previously that had about a 1hr run time per side for the adaptive and finish passes and altered the rapid rate down to the feed rate I was using and it only added just over 4 mins to the time so a lot will depend on the actual part but less that 10% increase I can live with and expect if I played around with things could get that time down as I only altered the rapid rate

[David]

I've taken to using 1mm for clearance heights etc to try and avoid too much Z height movement, although the Z moves are also the easiest to find and change back to G0s. Most G1 commands to a positive Z should be G0s. I've also tried using a 'marker' value like 1.2 so when I can do a find/replace on G1 Z1.2 but that doesn't work very well for a few reasons. Plus there are many X/Y rapids this doesn't catch.

I left the 2nd end cap running overnight, but when I checked it this morning it only went for about 5 minutes after I left! The spindle was really hot when I left it last night even though it was mostly cutting air. I'm wondering if I stuffed the bearings when I did the big plow though the table.

The 2nd run is so long because I used 2 3D 'parallel' toolpaths at 90deg to do the curved ribs with a 0.25mm step-over. I've done the whole 2nd cap with the 1.5mm endmill at 5000rpm, 30mm/min, up to 1mm depth of cut on the pocketing operation. It seems happy in brass with those parameters. I thought I'd stay down at 30mm/min due to the depth of cut. If only everything could be made of brass!





My 4mm 3 flute cutters arrived, but they have 4mm shanks and I don't have 4mm collets! I assumed they'd have 6mm shanks like all my other small endmills do. So I've ordered one of them now but it's on the slow eBay boat from China. Banggood charges postage but things seem to arrive a few weeks quicker which is why the endmills got here relatively quickly.

[Roger]

Personally, I don't use any rapids because they don't make up enough of the tool path to make much difference the way I program the cuts. Maybe if you're using a full 3D type of path then it may make more difference.
I use F1000 and a linear move which gives a greater ability to override it manually when setting up. It also means I don't get caught out when typing X10 for example from the command line. G0X10 is flat out speed, whereas G1X10 is at whatever the feedrate is. I just think it's less dangerous. On a light machine it's not such a problem, but mine will happily shear off a 16mm cutter like a carrot which really wakes you up.
I use 1mm or 0.5mm clearance depending on the situation.

My advice is to be ultra conservative until you're really confident you can walk away, it's not a race.

[jasonb]

One of the perils of buying from sources that don't give much info on what you are buying. The 4mm shanks are not a bad thing as they do allow you to reach further down when cutting upto an edge, a standard length cutter on a 6mm shank will be more limiting, I've also go some on extra length shanks so you still have plenty to hold even with quite a bit sticking out of the collet.

If you look at the manifold and engine end plates half way down this page you can see that I needed to reach 16mm deep against the manifold flanges and into the endplates that would not have been possible with a 6mm shank cutter.
www.model-engineer.co.uk/forums/postings.asp?th=141724&p=13

As Roger says and my example in the previous post depending on what you are actually doing the loss of Rapids may not make a lot of difference, loss of Steep and Shallow is what I will miss most as that gives as good if not better finish that two parallels in less than half the time, no wonder that was removed to stop commercial use.