Feeds, speeds, and other topics for the home machinst

[David]

While still avoiding the firebox cladding I decided now was the time to do the safety valves, and to get the NSWGR profile correct it seemed like a good idea to bring a works drawing into F360 and use RapidTurn to do it. So I set up the RapidTurn this morning.

You have to take the motor off to install it so you can reach one of the holding down clamp nuts.

I plunked it on the table and aligned it per instructions. It took a while but I can see once you've done it a few times it should be easy. Then reinstall the motor.

Switch everything on, reference the axes, and break the rotational sensor cable! I had slung it over the mill head to keep it out of the way and the cable isn't long enough for that. As the head goes up to it's limit switch the cable goes taught and pulls out of the 5-pin DIN socket in the control cabinet, bending the pins and leaving the round outer shield behind. I bent the pins back as best I could. They look like they'll break at any time, hopefully not in the DIN socket. I'll have to replace the plug at some point obviously.



Better keep going and see if it still works, turn the spindle on and it goes backwards. Bob Rogerson hit this and the manual has it in the troubleshooting section as "very rare". I'm not sure I believe that. Swap the U & V wires on the motor and keep going.

A test indicator shows about 0.0001 run-out on the 5C taper of the spindle. I don't know if this is good or bad but given the generally dodgy nature of the whole tool setup it's probably not significant.

Now mount the criminally rough casting that holds the dinky toolpost. The M10 bolt won't even start from the front so I clean out the thread and read about aligning it. You're meant to run a dial indicator back and forth over the "front" face to get it aligned with the table X axis. This face is so rough it looks likes it has wave crests, as is the one at 90deg to it. Not much to be done about it just now so indicate it as best as I can and clamp it up. Put the toolpost stud in and it's drunk to the tune of 1mm over it's about 50mm length. Clamp on the dinky toolpost and really crank on the nut to try and bring it flush with the wave crests on the casting. I can't see much rigidity here. If Tormach don't agree to send me a better one I might have to attempt to mill these surfaces flat myself.





At this point it's time to go to the running day where I spent most of the afternoon in loco trying to figure out why my injectors still don't work. One of them was blocked but would still pick up above 80PSI which was confusing. Then after that was fixed the plastic tube from the tender to the loco is just too short so kept getting pulled off the spigot at the tender end, where it cannot be seen. So that was fun.

Come home and try to find a bit of imperial stock to fit in the 5C collets that come with the RapidTurn. Tormach users are obsessed with imperial measurements and aluminium. Can't find any so turn down some 24.xx mm steel to 15/16" - it wouldn't fit in the collet to begin with, even though it should have. The lathe chuck comes off because the camlocks came loose. Knowing today was not the day for parting steel I use the bandsaw and the blade binds and comes off.

Put the steel in the 5C collet, stick it in the RapidTurn and follow instructions. I have to guess at the SFM for steel because I assume aluminium figures from the manual would not go well but the depths of cut look ok - less than I'd use on the manual lathe.

Initial confusion when the program moves the tool right rather than left but that turns out to be just it going to M30.

The program starts and things look a bit hairy but I let it go. Then I see steps forming at the ends of the cut, to the tune of 2-3mm wide. I finally figure out this is because the steel is undersized for the collet and is getting pushed back into it.

Break out the 3 jaw chuck on a 5C taper. Start the lathe and it looks like it's about to fall off! Bob Rogerson also hit this - the chuck is not screwed to the backplate properly. The screws in mine were quite loose. I took them out with the intention of cleaning between the surfaces etc but it doesn't want to come off without some force and I don't want to break something (else). So I tighten up the screws and hope for the best.

Put the steel in and have another go. Still hairy and the mill head (now the X axis with the lathe tool) drops a few mm at times when it should not. No idea why but if it does that in the middle of a cut it's going to be exciting and break a tool. That's really scary and I'm pretty miffed about it.

I forgot to reset my X value so the diameters are wrong but it gets through the tutorial program, with a couple of those alarming drops and what looked like a dig in at one point too. The smaller diameter is 0.003 oversize comparing it to the larger turned diameter to take out the mistake of not setting X. That's pretty ordinary. But it basically did what it was meant to.

As a final exercise I reset my X value and do another OD turn program with a higher SFM but 0.015 doc for roughing and a little curve or chamfer to blend it into the larger diameter. That goes ok - no idea if the radius is correct but the finish is better and the cut looked and sounded better. This time the cut hits the number to the nearest 0.001. Then I round over the end with a 2mm radius and this gets a bit heavy but seems to work.

I don't think I'd trust this thing to get within 0.005 of any given dimension. For stuff like my safety valve columns or a smokebox door I'd say it'll be fine. I doubt it will do wheel tyres like I was hoping it would though.

I have to consider my wrenching of the sensor cord and stuffing the pins could be having some detrimental effect, but I'm not sure how it could cause the alarming head drops.

The casting for the toolholder is a disgrace.

It's enough of a faff to install that I'll try and leave it on until I've produced something usable with it. I'll try a few simple models and toolpaths from F360.

[David]

I did some googling and found other people had problems with erratic table movement too. The RapidTurns ship with one of two sensors and one of them seems to give trouble. I took mine apart and of course I have the dubious sensor. I'm guessing there is no guarantee of which one I'll get if I order a replacement and I'm thinking after exchange rate and postage I'll be looking at $200 AUD. I've written to the importer and Tormach with the details and a pointer to the forum thread where the problem is discussed.

However! When starting to put the mill back to milling mode the first thing I did was open the control cabinet to replace the RapidTurn programming key in the VFD with the milling one. And the one I pull out of the door pocket (ie not installed in the VFD) says RapidTurn!

So that was a stupid mistake... the keys are clearly labelled, I knew what I had to do, and still got it wrong. Some people are just incompetent, and I seem to be one of them.

Now I was honour-bound to put the RapidTurn back together and do some more testing to see what difference the programming of the VFD makes. Other people with the iffy sensor say their machines are okay up to a certain rev limit.

I used conversational outer-dia cuts, bringing the spindle RPM limit up through 300, 600, 1200, 1300 RPM at which point things got erratic again. The table was starting to stutter and the RPM display for the 1200 limit test was showing 1300, and for the 1300 limit, 1400 was displayed while running. In last night's test where I was trying to limit the spindle speed to 1500 - 1600 RPM I was clearly going past what the sensor could handle.

So it looks like my sensor is only good to about 1000 - 1100 RPM, which is better than what some others found. The spindle is meant to be usable up to 3500 RPM.

The finishes I'm getting on mild steel (not free cutting) are nothing you'd put on YouTube to brag about but they're probably better than I'd get manually. And I've never had any idea what feed and speed to use on a manual lathe so that won't be helping matters.

I got some shiny mirror inserts that must be for non-ferrous for the turning tool so I'm going to try them on brass - my safety valve bodies. Sadly only the end that screws into the bush will be done with it, but I hope to see a good finish. Not sure how to turn this yet other than using a grooving tool for most of the body.

[uuu]

I have some inserts for Aluminium, which are very shiny. They do cut brass well, too - but come with a health warning. Yours may be different, but mine have a very aggressive top rake, so they are prone to digging in. They work best with very light cuts.

Wilf

[David]

Thanks Wilf, I'll keep that in mind! Some of my insert packets - not those from eBay - have a material code on the back. The shiny ones show a green N which google tells me aluminium and other non-ferrous and non-metals materials. Given brass and bronze are well known to require no top rake so they don't dig in I wonder at this. Industry makes a lot of turned brass and bronze parts but those metals have no group or geometry of their own so I assume the aluminium geometry must be good enough.

[David]

A lot of time spent getting the safety valve pillar program going this week. Overall I'm pleased I can get two of them that looks similar on a casual inspection from a distance. I'd hoped for more but I think it's going to take a great deal of effort and learning to get it.

The repeatability is still hopeless, 0.005 at best, and often something terrible happens between runs of a program and it's a lot worse. I have to make a skim cut, measure, and reset the diameter before every run of the program or I could end up anywhere within 1mm dia of where I want to be, in either direction. That's disgusting.

The mill head seems to be pretty sticky once it's stopped. I put a 0.0005 DTI under it and it takes a few clicks of the finest jog key to get it to move appreciably, and it loses something like 0.0005 - 0.001 on a change of direction from down to up. But it did seem to return to the indicated zero if you rapid moved a few mm up and back - not sure how. I don't think the indicator is sticky, it seems to work pretty well in all other situations.

I know others have managed good results so it's either me or the machine. I really hope it's me as I don't want to be fiddling with bearing preloads and who knows what. I spent a stupid amount of money so I didn't have to do that but I think I was mistaken in my expectations of the quality of the machine.

I noticed I was putting the tool holders on backwards so cut pressure could force them off their depth stop rather than into it and really hoped that was the problem. But I changed all my tools and reset all the offsets, and the results were no better.

I'm still feeling my way around depths of cut, perhaps I'm being too heavy handed.

I tried to set the drill chucks up but I think it's too hard to be worth it.

I can't get a decent thread out of F360, but the PathPilot conversational threading tab produced a usable one first time so I use a program I saved from that. I'm asking F360 to use the canned threading cycle but the parameters it's giving the cycle are no good. But it is magic doing threading on a CNC lathe. I'd love to try thread milling too, but cutters are crazy expensive down here.

I'm not greatly enamoured of the Tormach machine but PathPilot is great. For usability it looks light years ahead of anything else I've seen on YouTube. Obv I have no direct experience of anything else. The HAAS control panel looks like a nightmare in comparison.

I wrote to Tormach and the importer last weekend about the dodgy sensor, but no word back from either yet. Not impressed.

I thought I'd be able to turn the safeties with a RH and a LH tool, but it seems the easiest way is with a 2mm grooving tool. I think if I used to RH/LH pair I'd end up with a ridge where they meet or overlap because I can't measure tool offsets accurately enough to avoid it.

[Roger]

Hi David,
Are you homing the machine between each run? I presume not since there's no reason to do it. I doubt if homing repeats very well, just like on the mill.

It's probably worth running a few experiments to try to track down where these repeatability problems are coming from. They could be mechanical ones with the tool post not seating accurately, or you could be losing steps through overloading or over accelerating the stepper motors. Maybe you're running out of travel and the stepper motors are still driving. It's an open loop system, so anything can happen and the control won't know.

I'd find a place on the machine where I could put a clock and return to a known position so you can see if it comes back to the same clock reading. It ought to be easy to see whether the basic machine is repeating by that method. Does it repeat in the length direction or is the issue only on the diameter? Again, you can check either of those.

Another issue is backlash which we know isn't controlled in PathPilot. You're probably going to have at least 10 microns lost motion in the leadscrew, and if you're not approaching a diameter in the positive direction, it may or may not take up the backlash depending on the depth of cut. You might want to arrange for all of the finishing cuts to be approaching inwards so the backlash is always taken up. In other words, if you're adding a chamfer and diameter while working towards the chuck, you'll be approaching that diameter with an outward movement, and that's not good. This is why backlash compensation is so important and why I can't for the life of me understand why Tormach can't appreciate that.

Grooving tools are great because you can avoid the issue of trying to get two tools to precisely machine the same diameter. This is really hard to get spot on.

[jasonb]

Also check that your 2mm grooving tool is not flexing sideways under a heavy cut and giving odd sizes, mine will certainly flex if I push it hard on the lathe. You can get aluminium specific inserts for them which would also cut well on your nonferrous parts.

[Roger]

Apr 22, 2019 6:22:08 GMT jasonb said:

Also check that your 2mm grooving tool is not flexing sideways under a heavy cut and giving odd sizes, mine will certainly flex if I push it hard on the lathe. You can get aluminium specific inserts for them which would also cut well on your nonferrous parts.

This is a good point. I always plunge with it like a parting tool, using it in the strongest direction and then cleaning up with it moving sideways. Mind you, mine is 3mm wide so it's probably a lot stronger than the one David is using.

[David]

Thanks for the info guys, lots to think about. I don't home the machine between runs. If the homing switches are as ordinary as you think they are it probably wouldn't help. I'm guessing they're really there just to avoid over travel, not set up a consistent set of coordinates between reboots etc. I do usually take a skim cut on the stock between runs to ensure my X is still ok but with my crappy measuring tools it's hard to know if the variation is in them or the machine! Hopefully my shiny new measuring tools will help determine that.

My grooving tool toolpath does a series of plunges to close to diameter then the finishing passes drag it along the profile. I'm sure it's suffering a bit and I did dial the load back and got better results on the latest two bodies I made.

I've stopped working on it until I can test the two I've bored out, and I'm waiting on springs and balls. The springs are $6 each! That's a whack to swallow in terms of selling these. It'll be the most expensive part I think.

Tormach finally got back to me and after I reset the sensor location again as they requested (I'd already done it) and the problem still showed up they agreed to send me a new sensor. They also said their QA procedure changed 2nd half of last year to try and catch the dud sensors before shipping. As part of the test I did an online upgrade of PathPilot - was quick and easy with no problems. Nothing I could see in the release notes was useful to me though.

I sent them photos of the casting the toolpost bolts to and they agreed it was too rough and hopefully they'll send me another one of them too. If they do I'll mill the original one flat myself given I'll have a spare if something goes wrong.

I'm getting the impression this machine is only good for 4-6 thou (inch) tolerances. That's pretty dire but seems to be widely acknowledged. Some people have no trouble at all and others have nothing but. The safety valves don't have a single critical dimension on them which is ideal, but I struggle to see how you'd make batches of parts with 1 or 2 thou tolerances for customers. Maybe if you're good at programming, measuring, compensating, etc it's possible.

Anyway, it'll probably make anything I want for myself. I'm just wondering how I'll make it pay for itself at this point. There are many, many things I could buy or make to make it easier to use but that pushes the sunk costs up ever higher. A fixture plate (a big expensive piece of steel!) an electronic touch probe (I'd love one of these), like all our machines the list is endless!

I got a cheap electronic tool length probe in the post today, about $100 AUD. I splurged some of the money I got for the tender sides. If it can do better than 0.1mm it'll be better than the height gauge I got from Tormach. Which did come in handy for the firebox cladding - first time I used it for anything other than measuring tools!

[Roger]

Can you post some pictures of your setup, I'm finding it hard to appreciate what the machine looks like? I can't see why it won't repeat better than a thou, that seems to be the minimum tolerance for it to be remotely useful. 4-6 thou is frankly a joke in my opinion.

[David]

I mean the whole machine, not just the lathe part. I don't think even Tormach expects it to be able to hold tolerances of better than 0.004. I don't know how much of that is the machine design itself, the factory and assembly quality (these are built down to a price, not up to a standard) and how much is just the expectation that the users won't/can't maintain and tune it any better.

Remember it's basically an RF-45 clone with stepper motors on it. It's not actually built on the RF-45 style castings but design-wise that's what it is.

I'm sure you (as in Roger) could tweak it and dial it in and get better results. Until I have to I don't want to spend the time and risk making it worse!

To see the lathe part, look up RapidTurn videos on YouTube. In particular I think you'd like the ThreadExpress channel, a clever bloke in NZ.

[jasonb]

I'm sure you should get it down to a thou, there was a video I saw where someone ended up sending theirs back as he had been told it would work to 10ths but he could only get it to a thou. Will see if I can find it.

I've recently got a Sieg SX-3 and straight out the crate without any backlash compensation set or adjusting the gibs I did an internal square 20mm x 20mm and that came out at 19.99 x 19.98 which is 4 tenths. Z has a bit more backlash in it than X&Y. The few parts I have made with it so far all fitted together which is what counts

Mine was a customer return that had known Z axis issues which I have just about sorted, while doing this I ran the same pprogram multiple times with a DRI stuck to the column and measuring where the head stopped at at the end of each run, got within 0.01mm repeatability over about 10 runs.

[Roger]

Apr 26, 2019 23:38:52 GMT David said:

I mean the whole machine, not just the lathe part. I don't think even Tormach expects it to be able to hold tolerances of better than 0.004. I don't know how much of that is the machine design itself, the factory and assembly quality (these are built down to a price, not up to a standard) and how much is just the expectation that the users won't/can't maintain and tune it any better.

Remember it's basically an RF-45 clone with stepper motors on it. It's not actually built on the RF-45 style castings but design-wise that's what it is.

I'm sure you (as in Roger) could tweak it and dial it in and get better results. Until I have to I don't want to spend the time and risk making it worse!

To see the lathe part, look up RapidTurn videos on YouTube. In particular I think you'd like the ThreadExpress channel, a clever bloke in NZ.

I thought Tormach claimed 1 thou accuracy, although without backlash compensation it may not be easy to achieve that unless you have good quality ball leadscrews. It seems to me that there's something very wrong with either the leadscrew, the thrust bearing arrangement or the gibs. You have to remember that the control only sends steps to the motors. How that translates to the physical motion of the table is entirely determined by how well the machine is made and set up.
Checking lost motion is simple enough, just put a clock on the direction you're interested in and tell it to move back and forth onto and off the clock. On a dial clock that might only be 0.1mm but it's enough to get a feel for what's going on. I think it would be well worth the time to put locomotive projects to one side for a while and really get to grips with what's wrong with the machine.

[David]

Jason, what material were you cutting and what was your feed and speed?

I haven't seen them claim 0.001. I have had some good results - the tender axleboxes I made came out well and possibly within a thou of each other until I got impatient doing the slots for the hornguides where it all went to hell. But that was me making bigger cuts than perhaps the machine was happy with. The guides the boxes were going into were not machined so had no great accuracy so I don't think this mattered.

I have seen the video by the guy who returned it, he was trying to interpolate a large outside diameter. I wouldn't expect a Tormach to do very well on that job and I'm surprised any machine can make a good circle using just X/Y moves. Having said that the 13mm bores in the above axleboxes were milled and the finish was excellent and the fit of the bearings in there was better than I could have done manually. Lucky I thought to take another spring cut before tweaking the program.

The safety valve fixture turned out really well too, no wonder all the YouTube heroes work with aluminium. It took no skill to get a great finish and fit. ThreadExpress does work with steel and I imagine his machines are dialled in pretty well.

It could be I'm still just asking too much of it or doing it wrong. But if I went much slower the thing would be stopped. I'm sure the machine would benefit from the bearing preload and jibs being checked and adjusted if necessary - I'm not too good in that department. There are tons of videos about it though and the manual probably has the procedure in it - the documentation is pretty good.

I think I wrote above the head seemed like it would come back to a zero on a DTI with a rapid move away and back despite showing much stickiness when jogging. A mystery.

[jasonb]

Typical, that was the one set of code I have not kept but going by what else I was trying around the same time it would be: 6mm thicK 6082 T6 aluminium, 6mm dia three flute HSS cutter, 3000rpm cutting at the full material depth, 150mm/min feed 1.0mm DOC on finish pass.



I also did another part in steel that had four raised "D" shaped leugs that were drawn with 20mm between them, this is the fit of a 10-20-40 block between them



I have posted some of the other things machined in this thread on ME forum, if you click the video's and watch in Youtube the feeds and speeds are listed there if I have not put them in the posts. Various materials, Mild steel, Corian, Bronze, EN8. Andrew who posts several times in that thread has had a Tormach for several years and he is a bit a**l when it comes to accuracy.
www.model-engineer.co.uk/forums/postings.asp?th=141724&p=1

J

[David]

Thanks Jason. It seems your steel speeds are 3000-3600rpm, 150mm/min, with 0.25mm width-of-cut. That's a bit more than twice the rpm and 5 times the feed I would have used! How is your cutter life? I didn't even cut brass at 150mm/min. 120mm/min was as far as I got I think.

I have some simple parts to do for a friend in steel so when my new 6mm cutters arrive I'll give your settings a shot and see what happens. My cutter will be 4 flute but that should mean less load on each tooth.

That will also be an interesting test for the effect on accuracy. I have a couple of the same parts on my bench cut at my more conservative rates that I haven't sent off yet so can compare them.

I took my 1400 rpm for a 6mm cutter from a video by Stefan Gotteswinter. I think he computed that speed for HSS but he did tests with a carbide cutter too at the same speed so I assumed it would be ok. It did give good results for a while but the cutter didn't last all that long until the results started getting worse. Not bad enough to ruin the part, they will clean up fine with sandpaper.

[jasonb]

That was quite a shallow DOC as I was cutting the full 8mm thickness and it was also EN8. I did a second one of those star wheels for someone else in EN1A and fed at 200mm. Both still a bit tame as I'm getting used to the machine and can react quicker if something is not right at those sort of feeds. Plenty of life still in the cutter it hardly looks used.

The small bronze gland that I showed was at 5000rpm ( my maximum) and 300mm/min and could happily have been fed faster.

Just because you have more flutes should not mean you reduce the chip load. You should keep the chip load the same which requires the feed to be increased by approx 33% when going from a 3 flute to a 4 flute provided you can clear the swarf.

What make are your cutters? most half decent makers publish feeds and speeds for various materials and situations such as slotting or profiling. You can base your settings on those but reduce a bit for the lighter machine. For example the 6mm dia 3-flute carbide cutter I used on the star wheel gives a speed of 4100rpm and feed rate of 590mm/min when profiling at 1xD height and 0.05D DOC ( 6 x 0.3 ) in steel less than 30 rockwell. As I was cutting 8mm I reduced the DOC to 0.25mm and speed was less as I don't have cooling or air to clear the work set up yet and still find those feeds scary fast at the moment. For the 4-flute version they give 5100rpm and 605mm/min with a greater DOC of 0.1 x D

Talking with Andrew when I see him the feeling is that these lighter machines work better with a fast running and quickly fed small diameter cutter rather than slowly hogging it out with a big cutter. Combine that with adaptive tool paths and they can shift metal at a good rate if time equals money.

[David]

Last night, spurred on by Jason's success in such a short time I decided to have a go at the G1 wagon wheel I modeled a few weeks ago, as the mogul is still in the doghouse. I hoped the material was cast iron but it was free machining steel which was fine.

The 6mm cutter just cut down from the top of the stock to the top of the spokes and I used 3000rpm - 150mm/min feed. No trouble there as the cut is so shallow- 0.6mm. It plunged into a pre-drilled hole that was used to clear most of the material between the spokes. The hole was 5mm dia so only 0.5mm around the outside of the EM plunged into steel. The plunge rate was 60mm/min. I had the lead in/out speeds at 60mm/min too, they should have been the same as the cutting feedrate.

For milling out the spokes I drilled with a 5mm drill, followed up with a 4mm EM where it seems I forgot to change the settings and F360 is telling me I used 1400rpm - 60mm/min. Oops. It was going down 1mm at a time but I'd say could easily handle 2mm or 3mm so will increase that too. Maybe full depth with a spring cut would work fine.

I finished them with a 2mm EM at 2700rpm - 90mm/min. That makes no sense either, the rpm should have been higher than the 6mm cutter I assume. This cutter was going down 1mm at a time, perhaps it could have handled 2mm.

EDIT: I remember why those speeds went down on the 4 and 2mm cutters. They're HSS not carbide so I thought maybe I should slow them down.

What's missing is a fine line to represent where the tyre joins the wheel. Probably won't miss it in G1.

The only mishap was that I forgot to tighten the collet nut on the 2mm EM which I will assume is the problem the caused one of the spokes to be .1mm too thin.

As I've already turned 50mm of the steel down to size (33mm) I'll do a second wheel tonight with no loose tools and the updated toolpaths.

[David]

I did some more wheels, changing the feeds and speeds. Then I removed an adaptive clearing op which took 5 mins and made a lot of table moves and replaced it with a simple profile that took somewhere less than 2 mins I think.

I tried the 2mm cutter at full depth because it didn't have much to remove but it broke on the 4th pocket. I had another so have changed that toolpath back to start at about 2mm deep and spiral down the other 2.1mm. Perhaps I should take it back to the original 1mm. The finish from the full depth toolpath was clearly not as good anyway despite a spring cut. A 2mm HSS EM might not be up to much...

All the endmill cuts have already had most of the material removed by drills - none of these are working hard IMO. I think the 6mm cutter finish was getting worse over the course of the 4 wheels even though it's doing almost nothing, but it was pretty rough even when I started this test. I'd have to make at least another 4 to see if the 2mm cutter lasts longer without the full depth toolpath.

6mm EM, carbide, pretty rough: 3000rpm, 150mm/min
4mm EM, HSS, was new: 3500rpm, 150mm/min
2mm EM, HSS, Dormer, new: 4000rpm, 90mm/min.

From a runtime of 20mins, 30sec to 18 mins. I was expecting about 16 or 17 mins after removing the adaptive path.

I still need to make a spider so I can work on pre-cut blanks, and do some ops on the rear of the wheel, add in a step for reaming the axle bore, and then use the RapidTurn to cut the tread and flange so another fixture for that - some sort of arbor that will positively locate and drive the wheel to ensure it doesn't slip or wobble.

Writing off the cost of the arbour and spider, I'm guessing each wheel will need a few mins parting off the blank (wouldn't trust RapidTurn for this and I'd have to stand there feeding it stock anyway) , ~20 mins for the front, say 10 - 15 mins for the back, and maybe 5 mins for the turning. So 30 - 45 mins each.

An equivalent set of Slaters injection molded wheels is about $19 AUD and they almost certainly look better than these.

However, it is a good first step to doing loco wheels!

[David]

The new sensor arrived today so that's very good service. It was quicker for the sensor to arrive than for them to answer the e-mails.

I also tried to figure out the wiring on the cheapo tool length sensor last night. I'd thrown out all my test leads and lost the alligator clips for my meter so that made things a bit challenging. The instructions are in Chinese and the wire colours are different but I think I've figured it out.

Didn't have time to solder a plug on today, maybe tomorrow.