7mm Mickoo's Commercial Workbench

mickoo

Western Thunderer
Cheers Mick, I looked at the prints I did last night (and washed very hurriedly this morning before leaving for work) and despite the IPA being very grey, they seem to be pretty clean. Have a look at the solebar number plate on my thread, though to be fair it has had a thin wash of grey.

I don’t have a suitable pot to put in the curing station. Might have to go to Wilco or somewhere similar, but probably not before the weekend. Definitely safer than raiding ‘er ladyship’s cupboards…

I guess if I were printing 23 hour specials, I might be a bit more cautious.

Ozzy, I haven’t yet, but I’ll use some cheap coffee filters, or, failing that, I’ll fold some kitchen roll into quarters, and then use that.
No old jam jars then?

Maybe it's the type of resin, my experiences was from a few years back and once bitten I've not tried again. IPA is an expendable product, much like nitrile gloves and rolls of kitchen paper.

Frankly it takes me longer (£/hr) to filter and reuse IPA than it cost to buy and fill fresh, my life is too short to be faffing around with such economies.

Fair do's to all that have and do, but not for me, I'd rather be down the pub having a beer :cool:
 

mickoo

Western Thunderer
Apologies for the photo bombing.

A quick blow with filler primer and then a physical test fit to see if it clicks into place on the etch core, it did.

There are a couple of niggles, one easily fixed, one which might kill the whole part dead in the water.

The RH side is sat a little too high above the gutter, a quick measure shows that this side of the gutter is lower than the LH side. I could skim a little off the roof print on the RH side but then the crown might not sit snug up under the overhanging plate. The other option is to secure the roof and then gently heat the gutter and push it up snug under the roof, it's only 0.2-0.3 mm gap so not really a big issue all told.

The big new elephant is the facets or lands that stl files exhibit, they show up on large gentle curved surfaces, the larger the surface the more obvious they become. They'd been present on the test shots but a blow over with filler primer and smooth down made them almost invisible. However once the whole roof printed the export function has made them really obvious and with all the rivets, not easy to sand down.

However, there is a small glimmer of hope, one of the front test shots showed no facets at all, a little root around in the back up files found that one had been merged and compiled differently. Looking at the part in detail the lands are still there but nearly 60% smaller.

I've got another couple of test shots going through now, one with the revised compiling and one with some post export edits (which take an inordinate amount of time to apply), I'm hoping one or both will open some new doors in getting rid of the dreaded facets.

Either way they need to be resolved or the printed roof is dead in the water as are other large curved surfaces like the VL80 cab.

Not obvious here is the revised smoke box module, the floor has now been raised in correlation to the smoke box door opening, that also required the tube plate to be adjusted and all new pipework to suit.

IMG_11703.jpg

IMG_11704.jpg

IMG_11705.jpg

IMG_11706.jpg

IMG_11707.jpg
 

Dave Holt

Western Thunderer
Mick.
The raised smokebox floor shouldn't impact on the tube-plate as it slopes down behind the centre cylinder before it meets the (full circular) tube-plate. Well, it does on a rebuilt MN.
Dave.
 

mickoo

Western Thunderer
Mick.
The raised smokebox floor shouldn't impact on the tube-plate as it slopes down behind the centre cylinder before it meets the (full circular) tube-plate. Well, it does on a rebuilt MN.
Dave.
It will if the tube plate was already too low ;)
 

adrian

Flying Squad
Frankly it takes me longer (£/hr) to filter and reuse IPA than it cost to buy and fill fresh, my life is too short to be faffing around with such economies.
Fair do's to all that have and do, but not for me, I'd rather be down the pub having a beer :cool:
Quite true - sometimes life is too short. The price of IPA has dropped significantly from pandemic silly prices so I see little reason to expend effort in reusing IPA at all.

My better half doesn't see a problem as we have a few garden light/heater units that use IPA fuel with a ceramic woven wick so any IPA I buy is for the garden lights - honest guv! Now it might have a cycle through the 3D washer unit first but it gets recycled in the lights rather than tipped down the drain so I'm not going to lose any sleep over buying another 5L when I need it. :thumbs:
 

mickoo

Western Thunderer
Recycle or leave to evaporate is the best way.

Tipping it down the drain is not be the best way to dispose of it, especially if it's been used for 3D resin cleaning, some of those resins are quite toxic to the eco system. I tip mine into an old 5 litre IPA bottle and just and leave the top off. The IPA evaporates and leaves the resin sludge which over time turns solid due to exposure with UV. Once solid it can then be disposed of in normal household waste.

Mind there are washable resins and I'm sure less toxic ones, I wouldn't personally wash washable ones and dispose of the water in any drain myself but other people seem to have no qualms.

I'll look into the IPA light/heater units, might have a need for some of them in the summer eve's :thumbs:
 

adrian

Flying Squad
I'll look into the IPA light/heater units, might have a need for some of them in the summer eve's :thumbs:
My better half bought a couple of the "bio-ethanol" garden lights. I can't find the exact units but the "bio-ethanol" fuel seems very similar to IPA and uses a ceramic woven wick. The IPA works just as well in the units as the official ( and expensive ) bio-ethanol fuel so it's a win win for me!

Something like these - but ours don't have the stones!
 
Last edited:

mickoo

Western Thunderer
Finally!!

It's a long, painful, but worthwhile route to increase the fidelity of Autocad stl files and the dreaded facets. Made all the more complex by not having the latest OS on the big tower PC but on the laptop, itself being no slouch.

It takes a while and there are I'm sure, easier ways but this works, it's free and that'll do me.

Image.jpg

The lower casing is a raw stl file that pretty much all CAD packages export, that's about the limit of fidelity a stl file can handle. You can see the facets in the whistle depression, the upper one is the revised smoother one, a roughly 4:1 improvement. I suspect a pukka 3MF from a handsomely priced payware CAD package could do better, but how much I don't know.

Much smoother than this and I think your file sizes are getting to the unwieldy size and the slider may baulk, it's sweating a bit now with this one.

One trick I did apply was to carve the model up into two parts, one with all the little details where the facets cannot be seen, aka rivet heads etc; the other being the main casing skin; I suspect if you smoothed all the details as well in none file then everything would slide to a stop. The trick then is to merge the two files in the slicer; each file has an exact copy of an object in the same 3D space, when the objects are perfectly merged then both files are in alignment.

All of which means that all previous versions feed the bin monster, re slice and start again :D
 
Last edited:

SimonT

Western Thunderer
Mick,
I wonder if the stl triangle/faceting issue is a product of Autcads export routine. I haven't seen such problems with Rhino. I can control/tune output attributes such as maximum distanc, edge to suface (The only one I have changed from default, I use 0.01mm) and six other attributes. I can choose which stl file format to use, either Binary or ASCII. For the Hall firebox the number of triangles is 132938 giving file sizes of 6.6Mb or 31.1MB respectively.
We can try running the model through Rhino if you like but that, of course brings in the variables of different dwg/dxf export and import routines!
Simon
 

Big Train James

Western Thunderer
Simon, Autocad doesn't give you that sort of control over the creation of the stl file. There is a function to increase the "facet resolution" as they call it. But it is limited, and gives no control over things like max distance from theoretical surface to the triangular face that represents it.

I have a licensed copy of Rhino, v5.something. I can theoretically open Autocad dwg files directly in Rhino, but when I do that nothing ever shows up. It's a version thing, I think the Autocad files are too new for the version of Rhino. I do need to try saving my dwg files to a much earlier version of Autocad, just haven't gotten around to it yet.

The workaround is to export the file in an iges format, then import that into Rhino. Then I can get much higher resolution stl files. The same sort of exercise could be performed with access to Solidworks or Fusion, but at the moment Rhino is the only non-Autocad software I have access to.
 

SimonT

Western Thunderer
James,
I started on Autocad, the last version I used was v.14, bought for $10 in the souk in Kuwait as part of Enduring Freedom! I'm now on Rhino 6. As you know, Rhino has a very impressive list of file formats that can be imported and exported. However we tried swopping file formats within F7 and found it to be somewhat iffy. (The three of us were using one of Autocad, Turbocad or Rhino!) The exchange of igs files between Turbocad Hand Rhino worked reasonably well but I don't remember us trying that kind of swop with Autocad. However, igs is a good alternative to the vast array of different dwg/dxf standards. We'll see if Mick wants to try processing through my machine.
Simon
 

mickoo

Western Thunderer
Thanks for the offers of alternative process but the answer is no for many reasons.

First is time, mine and your's I run several test iterations a day, I can't sit waiting for files to come back from others days later, big files get worked on during the day and then would need a conversion elsewhere, often around 01:00-02:00 before slicing and running over night.

The BLP roof alone has had 17 test shots in 10 ten days, I doubt anyone wants to convert that sort of file processing in my deadlines.

Second, I will not release my files elsewhere, there are build attributes and tricks to get good prints, that took time and research, I'm not giving that away to anyone.
 

mickoo

Western Thunderer
Mick,

Soooooooo, does this mean water tank sides are printable ???? ;)

Ian
No for several reasons.

1st: Warp free requires some internal tricks and specific supports, your tank walls are too thin to do that easily if at all, I'd need to add the supports with breakaway points to minimize damage, I won't know where they are until I test print some panels. Then I have to get hold of you, explain them and get you to draw them and then run a new test print; we may have to do that six or seven times, possible for each type of panel. It's time consuming for you and me, time I don't have right now, plus you'll see how I support stuff internally ;)

2nd: Cost, with respect you probably can't afford it and I certainly cannot afford to tie the machine and my hours up to do it. The BLP roof alone is near two weeks work and a couple of bottles of resin, add in the hourly rate and you're looking at a price tab of £4-500+ for the first roof. The signal heads were easy, small stuff done many times before, but even then we had to go through several iterations to get the final result that was workable. I was able to offset the 'true' cost as I used spare space on the build plate for other stuff whilst printing yours.

3rd: The tank parts are large, even if you resolved the warping and internal supports or even if just flat panels with bespoke supports, you're looking at 5-6 hours to print each panel, just no build plate time spare at the moment.

4th: Fit, getting 3D parts to fit together with tight joints is hard work, you can butt joint but the adhesive will fail over time, the resin is not a plastic that bonds well with most known adhesives. You'd need to design some sort of locking assembly or joint alignment aids, that takes a lot of testing to get it right so that the gaps marry up on each piece. Then you have to test print and make sure those areas remain stable and do not warp or deform.....on top of the main panel not warping or deforming.

Here's the EMD spartan cab, note beveled corners, lobes and sockets to aid alignment, even then this joint is a very weak spot and not something I'm happy with at all. The two highlighted parts are bespoke sacrificial support beams to aid printing, they're cut off the final product but are needed to ensure those edges print straight and warp free.

Image.jpg

The top one is an inverse warp item, basically that top edge, no matter the orientation or settings would always bow upward, therefore the sacrificial support has an inverse bow that when printed becomes flat so that when you get to print the actual part the stresses are already neutralized. That simple inverse part took near on a dozen test prints to fine tune, alter the orientation in any axis and you're back to square one with a new inverse part.

There's over 50 iterations on the cab front to get it warp free and print all the detail, cab sides are over 20 and the rear wall is much easier and follows the cab front so only 5 maybe six iterations.

This cab alone is near three years 3D dev time and materials, time is never free so you're looking at something like £2k costs and now I have to redo it all with the new machine and settings I have found on the BLP roof.

IMG_1420.jpg

IMG_1424.jpg

Note slot in the rear of the nose, this interlocks with the cab front and pulls that face flat and square, without that slot the cab front bows inward, the cab sides (despite all the bespoke sacrificial blocks) still have a small bow inward, that'll get taken out with a thick brass floor to hold it all square.

The best way to print the cab and tank is one whole piece, takes a long time but you have no joints to worry about, the BLP roof could be broken down into three sections, it'd make dev time much easier and the final print much quicker; you can get all three sections on the plate in place of the one big one. I would have preferred that but getting a near invisible joint on the three sections would have been near impossible with all that rivet detail.

I still maintain that 3D is not the best medium for the water tank, brass sheet is, given it's size and form, there are some objects where 3D is not the right application.
 

Lancastrian

Western Thunderer
Mick,

Wow, that is a significant cost At least it's something I'm not having to build, lol. The EMD cab is looking good though. Your reply does highlight, not only to myself, but others, the headaches that can arise with 3D printing.

Still need to sort out the other files I sent you too.

Ian
 

mickoo

Western Thunderer
Mick,

Wow, that is a significant cost At least it's something I'm not having to build, lol. The EMD cab is looking good though. Your reply does highlight, not only to myself, but others, the headaches that can arise with 3D printing.

Still need to sort out the other files I sent you too.

Ian
Dev costs are never really added up, if you do you'd scare yourself silly, I do to a certain extent but it's not that accurate and always on the light side.

If you're self employed then any dev work you do means you're not working for a paying customer, therefore that time is not free, you have to account for it.

If you think 3D is pricey then kit dev is worse, probably close to £20k (test builds, etch revisions, drawing 3D for casting) in my time alone, let alone materials and other costs. It's not a problem when you have a full time job to offset it, but now it's a cost I cannot sustain/hide.

Afraid to say the returns do not cover the costs anymore for group builds; I've done two, there will be no more. Independently it has a slim future as you potentially recoup 100% of the sales.

Your parts are sat on a build plate with a load of other pieces for other customers, it's just finding the plate time to get them in. I am acutely aware that a 2nd machine (possibly a third) is probably the answer, but I've just not quite reached that work/cost threshold.
 
Last edited:

mickoo

Western Thunderer
Mick,
I wonder if the stl triangle/faceting issue is a product of Autcads export routine. I haven't seen such problems with Rhino. I can control/tune output attributes such as maximum distanc, edge to suface (The only one I have changed from default, I use 0.01mm) and six other attributes. I can choose which stl file format to use, either Binary or ASCII. For the Hall firebox the number of triangles is 132938 giving file sizes of 6.6Mb or 31.1MB respectively.
We can try running the model through Rhino if you like but that, of course brings in the variables of different dwg/dxf export and import routines!
Simon
Certainly is, so it's time to explore alternatives and move forward, the workaround I have is long and convoluted.

Current roof is 1.639,517 triangles at 89Mb I'd say that should be reasonably smooth, won't know until the print pops out tomorrow though.
 
Top