New Forum for CAD?

[Midland]

I was given a copy of Turbo CAD deluxe 20 some years ago. Just installed it and it looks frightening. Any views on whether it is any good??
Cheers David

[silverfox]

AS the OP on this you might see that i am still stumbling a bit with AutoCad (2007). I have done some 'slot together' drawings and have sent them off. However they need to be polylines and as such the CAD submitted isnt suitable. I am in corres with Malcolm High and he has given me some instructions but feel a bit relunctant to keep asking him as he has his own work rather than holding my hand

So can anyone give me some assistance on getting the drawing converted . in simple terms that is... meanwhile i am still on my teach yourself learning curve plus whilst i want to get some metal in my hand, the fact i have been on this for the last 6 months another week or so isnt going to affect the whole scheme of things

Thanks

Ron

[92220]

Hi Ron.

If you want to convert a series of lines into a single polyline, type pedit on the Command line, then select the first of the lines you want to join up. The command line will then ask if you want to turn it into one, meaning a polyline and there will be a (Y). Hit Return, then when the next line pops up click on 'Join' or type J. It will then ask you to select object. All the individual lines that are connected to each other will then be joined up and converted to a polyline.

Hope that helps.

Bob.

[92220]

Dec 3, 2017 12:56:41 GMT Midland said:

I was given a copy of Turbo CAD deluxe 20 some years ago. Just installed it and it looks frightening. Any views on whether it is any good??
Cheers David

Hi David.

I've not used Turbocad, but when AutocadLT was introduced, Turbocad was in it's 'infancy'. It had been out a couple, or so, years before. It was a toss-up between the 2 systems for a lot of users. I have heard good things about it, so good luck with it. Persevere and you should have a decent system to use.

Bob. 

[silverfox]

Bob

Is it really that simple??

I thought brain surgery was going to be a doddle!!

will let you know how i get on

Thanks

Ron

[joanlluch]

I believe that many of us who have gained some experience with 3D CAD will agree that it’s a mistake to begin with a 2D only software. 2D CAD was invented/developed at a time nothing else existed and it was inevitably based on hand based technical drawing concepts. I still have to see any great advantage with AutoCAD and similar products over drawing by hand, except for documentation and sharing purposes. The actual drawing experience and benefits to the designer are not that great. In fact these early 2D drawing packages have been always quite frustrating to use. And they continue to be.

3D CAD on the contrary is completely another subject. You actually have fun and are productive with it instead of hating the software. For starters, the main difference is that you are actually drawing parts, not just a pile of disengaged lines. Unlike 2D drawings, 3D CAD drawings have actual notion of part and assembly geometry in the space. This means that 2D views, sections, dimensions, lists of parts, are already implicit from the drawings and can be automatically created at any time with not a single missing line or feature.

So even if your geometry is flat, such as a locomotive frame, draw it in 3D. Export the main view to AutoCAD and send it to your laser cut provider. Much easier, faster, and effective than fighting and trying to draw the same view directly in AutoCAD LT (or turboCAD 2D btw).

My advice is, just use 3D and forget about 2D.

[Oily Rag]

Dec 4, 2017 13:40:05 GMT joanlluch said:

I believe that many of us who have gained some experience with 3D CAD will agree that it’s a mistake to begin with a 2D only software. 2D CAD was invented/developed at a time nothing else existed and it was inevitably based on hand based technical drawing concepts. I still have to see any great advantage with AutoCAD and similar products over drawing by hand, except for documentation and sharing purposes. The actual drawing experience and benefits to the designer are not that great. In fact these early 2D drawing packages have been always quite frustrating to use. And they continue to be.

3D CAD on the contrary is completely another subject. You actually have fun and are productive with it instead of hating the software. For starters, the main difference is that you are actually drawing parts, not just a pile of disengaged lines. Unlike 2D drawings, 3D CAD drawings have actual notion of part and assembly geometry in the space. This means that 2D views, sections, dimensions, lists of parts, are already implicit from the drawings and can be automatically created at any time with not a single missing line or feature.

So even if your geometry is flat, such as a locomotive frame, draw it in 3D. Export the main view to AutoCAD and send it to your laser cut provider. Much easier, faster, and effective than fighting and trying to draw the same view directly in AutoCAD LT (or turboCAD 2D btw).

My advice is, just use 3D and forget about 2D.

I second what Joan has said. I come from the paper/drafting board generation, then I got into 2D CAD via the Text books. The snap accuracy is super cool. But one is still in the draft board and paper mode and it is slow.
I got into Solidworks 3D. It took me some time to get my head away from the paper pencil frame of mind and working. However the "Inside Solid Works" is one of the best Technical Text Book I have ever come across. I worked from page one on-wards, yes it takes time. Many evenings working through the chapters but it is written very well and my dim mind followed it with perfect clarity.
Not sure of the other breeds, the writing of manuals is something that most companies spend as little $ as possible. (my wife is a technical writer)
So once one has their head around the simplicity and speed of creating a 3D part you will never look back.
Then there is the magic of dragging and dropping the 3D part into a page and it converts to 2D and with all the view options. Dimensioning is easy and fast and then print off for the workshop floor. Missing dimensions, easy to check and add. Then the assembly of parts which sorts out a huge volume of potential problems. The assembly of the parts is not infallible but heck, if Don Young, Martin Evans, Keith Wilson, LBSC etc had 3D available to them, a huge amount of errors, waste and anguish would be eliminated. They would have embraced it. Keith was using 2D CAD later before he was gone. Not all, but a lot of errors are removed and sorted and the updating the master set of drawings is simple.
I am currently working off 3D to 2D drawings for my simple garden steamer that seem to be just plain lazy creations and it appears there is not a lot regards editing or updates as these are not the first editions. 2.5mm shaft into a 3/32" hole sort of stuff. So every part I make I redraw into my own 3D-2D drawings and I am very pleased that I do! I appreciate that when one is of the age that learning this stuff is too much bother then OK, but if you are a new or perhaps a young person, where this is the industry mainstream it will become the norm in the ME hobby in the decades to come. I cannot recall any one in industry creating paper and pencil drawings today.

[joanlluch]

Dec 6, 2017 3:10:46 GMT Oily Rag said:

if Don Young, Martin Evans, Keith Wilson, LBSC etc had 3D available to them, a huge amount of errors, waste and anguish would be eliminated. They would have embraced it.

Actually, some of these old model engineers are often regarded as geniuses for the wrong reasons. Their greatest achievement was creating significantly good 2D plans out of paper and pencil alone. Also I bring a lot of merit to the old teams of mechanical engineers who were able to design and draft accurate plans of every single part of the latest full size steam locomotives without relying on any sophisticated graphical or visual tool.

Among the great model engineers still contemporaneous to us, I like the work of Kozo Hiraoka. He has published several books, and at his 80's he still runs a build series on the "Live steam & outdoor railroading" magazine. He supplants his possibly limited English language skills by the provision of highly detailed, amazingly clear, 3D drawn sketches of every single step of the building procedures. Definitely, not someone's work to miss in my opinion. 

[oldnorton]

Like several others I guess, I am toying with the move to a CNC mill. I have got a SX3 sitting there and the hardware conversion seems straightforward and DIY achievable for under £1,000 with reasonable parts bought locally. That sort of machine conversion is a project that I will enjoy.

It is the software steps that seem harder to fathom. I have used Draftsight in the past and now TurboCAD 3D, but I always seem to slip back to 2D. I have a MAC indoors, which is why I like to have MAC CAD software, but this is a rarer product I admit.

Then one has to output those drawing files to a CAM cutter program, and vCarve seems a good 2.5D middle ground, then output the G-code through something like Mach3 that actually drives the mill motors. This all costs a few hundred more. This all runs strictly on a PC only and Windows 8 seems a good way if you use the newer USB drivers. I have looked at doing this all in Linux, at very low cost, but the free software seems a bit more basic. If you want Mach3 for machine control, and a nice hand-held manual controller at £100, then it has to be Windows and Mach3.

I have read of people singing the praises of Fusion 360 so I have had a look. First big plus is that it is a 3D CAD system and sitting there on the same screen are all the CAM routines so that you can run tool paths, and modify the drawing as you test those. Secondly, it seems to be completely free for all Hobbyists (and any commercial business under a certain size). Thirdly, it runs under Windows and on a Mac, saving all its work to the 'cloud', so I can do a drawing in the evening on the Mac, go out to the workshop in the morning and files are already copied to the machine PC. It is obviously a big winner for me. (edit: I ought to add for beginners like me, the Fusion 360 output is in G-code that still has to be read by Mach3, and Fusion 360 seems to have that specific machine output, as well as dozens of others).

I would be interested in any comment from our more CAD/CAM experienced friends. Meanwhile, for those of you wondering how it all works, I found this particular Fusion 360 tutorial to be very informative. www.youtube.com/watch?v=_nb1VNnz4Uc

Norm

[Roger]

Dec 6, 2017 10:40:32 GMT oldnorton said:

Like several others I guess, I am toying with the move to a CNC mill. I have got a SX3 sitting there and the hardware conversion seems straightforward and DIY achievable for under £1,000 with reasonable parts bought locally. That sort of machine conversion is a project that I will enjoy.

It is the software steps that seem harder to fathom. I have used Draftsight in the past and now TurboCAD 3D, but I always seem to slip back to 2D. I have a MAC indoors, which is why I like to have MAC CAD software, but this is a rarer product I admit.

Then one has to output those drawing files to a CAM cutter program, and vCarve seems a good 2.5D middle ground, then output the G-code through something like Mach3 that actually drives the mill motors. This all costs a few hundred more. This all runs strictly on a PC only and Windows 8 seems a good way if you use the newer USB drivers. I have looked at doing this all in Linux, at very low cost, but the free software seems a bit more basic. If you want Mach3 for machine control, and a nice hand-held manual controller at £100, then it has to be Windows and Mach3.

I have read of people singing the praises of Fusion 360 so I have had a look. First big plus is that it is a 3D CAD system and sitting there on the same screen are all the CAM routines so that you can run tool paths, and modify the drawing as you test those. Secondly, it seems to be completely free for all Hobbyists (and any commercial business under a certain size). Thirdly, it runs under Windows and on a Mac, saving all its work to the 'cloud', so I can do a drawing in the evening on the Mac, go out to the workshop in the morning and files are already copied to the machine PC. It is obviously a big winner for me. (edit: I ought to add for beginners like me, the Fusion 360 output is in G-code that still has to be read by Mach3, and Fusion 360 seems to have that specific machine output, as well as dozens of others).

I would be interested in any comment from our more CAD/CAM experienced friends. Meanwhile, for those of you wondering how it all works, I found this particular Fusion 360 tutorial to be very informative. www.youtube.com/watch?v=_nb1VNnz4Uc

Norm

Hi Norm,
In my opinion, Fusion360 is the way to go. There's little point in going down the 2D route, that limits the machine to 2-1/2D work unnecessarily. I'd certainly use it if I didn't already have another system.

I don't understand the reluctance towards using Windows, specifically Windows 10 which is rock solid and the best version yet in my opinion. I run not only the CAD/CAM but also Mach3 and stream Music and surf the internet, all on one Windows 10 machine. What's not to like?

Personally, I'd go for Mach4, it's been around long enough now to be stable. I'll be switching to that in due course. I would use an ESS (Ethernet Smooth Stepper) interface embedded on a breakout board. Mine cost around £400 and that's all you need to produce the step and direction pulses for the three axes. I wouldn't recommend using a Printer port arrangement with Mach3, that puts a lot of computing burdon on the PC which it wasn't designed for. An ESS requires a dedicated Ethernet port, so you will need to add a second one if you want to connect that PC to a network or the internet as I've done.

Mach3 is a bit amateurish and buggy, but it does work well if you don't try to do things like starting part way through a job. I hope Mach4 is an improvement in this regard, it's been re-written from the bottom up, so it certainly ought to be. The biggest complaint about Mach3 other than that, is the inability to customise it. It's really an old DOS application, so although you can customise some functions, if you move buttons around, you have to create a brand new mask for your new layout. There are no tools for this that are of any value. I believe Mach 4 is fully customisable which is something I really want to do. The default layout is dire, with so much unnecessary clutter and confusing buttons that are of no value.

Again, I don't really understand the motivation for using Linux over a PC for the machine control. In the days when PCs and Windows were a bit flaky then that might have been a good reason. That's simply not the case today. Integrating the CAD/CAM and control on the one PC makes life easy when you want to modify tool paths while running the machine, something I do all the time. Making the control separate just adds another step in that process.

A friend is using Fusion360 for all his CNC work now, and although it took a bit of getting used to, he's now confidently machining all manner of complex 3D parts.

[steamer5]

Hi Guys & girls,
If you want to see what’s possible with Fusion 360.....check out this link.

www.modelenginemaker.com/index.php/topic,7530.0.html

Chris is doing a great job...enjoy!

Cheers Kerrin

[oldnorton]

Hi Roger

Thank you for the thoughts on Mach4, I will look into that and see if I should go straight there.

I like my Mac CAD option only because I have a Big Mac indoors (non-edible) and really don't want to run Parallels software (windows software on an Apple Mac) or put a new PC next to it. The Win8 or 10 PC in the workshop however is a given as is the use of an Ethernet Smooth Stepper, or a USB version which seems to do the the same processing job at half the price. I also agree that the LinuxCNC route is for older PCs, parallel drivers and a much more hands-on G-code approach at virtually no cost. A lot of people still seem to like the LinuxCNC route if they come from a computing background.

I am looking at motors, screws and bearings from a small company close to me in Milton Keynes. What I had not appreciated was the quality of screw bearings that you ought to fit if you hope to get close to 5um performance. For example, for each axis I am looking at ball screw assemblies at £60, then a floating bearing assembly at one end for £25, and the twin adjustable thrust bearing at £48. In all the amateur SX3 conversions I have seen on the web they seem to use basic home-made bearing assemblies which I guess 'easily spoils the job'.

Norm

[Roger]

Dec 8, 2017 10:35:24 GMT oldnorton said:

Hi Roger

Thank you for the thoughts on Mach4, I will look into that and see if I should go straight there.

I like my Mac CAD option only because I have a Big Mac indoors (non-edible) and really don't want to run Parallels software (windows software on an Apple Mac) or put a new PC next to it. The Win8 or 10 PC in the workshop however is a given as is the use of an Ethernet Smooth Stepper, or a USB version which seems to do the the same processing job at half the price. I also agree that the LinuxCNC route is for older PCs, parallel drivers and a much more hands-on G-code approach at virtually no cost. A lot of people still seem to like the LinuxCNC route if they come from a computing background.

I am looking at motors, screws and bearings from a small company close to me in Milton Keynes. What I had not appreciated was the quality of screw bearings that you ought to fit if you hope to get close to 5um performance. For example, for each axis I am looking at ball screw assemblies at £60, then a floating bearing assembly at one end for £25, and the twin adjustable thrust bearing at £48. In all the amateur SX3 conversions I have seen on the web they seem to use basic home-made bearing assemblies which I guess 'easily spoils the job'.

Norm

Hi Norm,

I can't comment on a USB interface version of a SmoothStepper, but I can't see any reason why this would be inferior to the ESS. The bandwidth required is pretty small.

The biggest thing that would improve the accuracy would be to use closed loop with linear scales, but sadly Mach3/4 is open loop. I'm sure there are systems out there that will do that, and Linux may support that too. In reality, top quality Ground Ball Leadscrews can give good accuracy. If you don't have recirculating ball leadscrews I doubt very much if the accuracy would be acceptable over the long term. It's not just a matter of backlash, but of wearing out the part where it spends most of its time, creating non-linearities and making backlash compensation impossible.

I blew the budget on precision ground ball leadscrews, the largest I could fit in for maximum stiffness. That still gives about 6-8microns of lost motion when you change direction. Fortunately, you can dial this out with a bit of experiment, putting figures into Mach3/4 to apply when you change direction. It seems to work during automatic control, but I'm not convinced that it works when you're jogging. I found that the figures you put in as compensation values don't really make sense, but that doesn't stop you finding values that give a good result.
I don't think many people use this feature. I reported back my thoughts and they didn't really want to engage with me about it.

My advice would be to go for 1 micron steps for the control system. This is going to help improve the repeatability. I can put a 2micron clock on mine and see it jog one micron at a time. You're only rarely going to use this when setting up, but there's no penalty in doing it, so long as your Stepper  drive system will give an adequate rapid traverse speed.

[atgordon]

Like many posters, I would welcome a CAD/CAM category (hopefully with support for 3D printing as well!).

I was traditionally trained on a drawing board (A level TD), which I also taught to A level and C&G level. I started electronic CAD with 2D ACAD in 1985. I now use Fusion 360 for all my design and drawing work. I sometimes have to dip into Inventor for some more complex stuff.

The CAM side of F360 is still in it's infancy compared to MasterCam. So I create parts in F360, and then export them as STEP files to M/Cam for post processing (an extra step, but it works much better than the current generation of CAM ops in F360). I should add that M/Cam is the worlds worst 3D drawing package, although it does have the ability to create 3D solids: it is much better to import from your preferred flavor CAD package and edit when in M/Cam (which is pretty easy to do so in spite of its somewhat anachronistic interface). M/Cam post files work seamlessly on my mill, and are much shorter than most comparable F360 posts for anything more complex than facing ops. Also, F360 posts often generate errors (even after a simulated run shows none in F360).

I am fortunate in having a MasterCam "seat" that I can use for free ...

[92220]

Dec 3, 2017 17:51:03 GMT silverfox said:

Bob

Is it really that simple??

I thought brain surgery was going to be a doddle!!

will let you know how i get on

Thanks

Ron

Hi Ron.

Sorry. I missed your post. Yes it is that simple. just draw lines that connect to the previous one and then highlight all the lines that are connected end to end (it won't join lines that meet, say, at the mid point of another line...only the ends of lines) then choose or type pedit (polyline edit). When the list of choices comes up, pick Join, or type J. Job done. It will then tell you how many lines have been joined. If one hasn't been joined in, just redraw it making sure the start is exactly at the end of the previous line. If you want to make the collection of straight lines into a smooth wavy line then after it becomes a polyline, highlight it again and then choose Spline or type S. This will then convert the jaggedy looking straight lines into one smoothly flowing wavy line.

Bob

[silverfox]

Bob

The info both you and Malcolm High supplied, was pearls beyond price ( or something like that) and like riding a bike, it is easy when you know how lol
What i did find is how sensitive the joining of lines are. I had to enlarge to the Nth degree to find out that there was a gap that just showed up, but the programme wouldn't jump it. by deleting the offending bit jutting out and redrawing the line. bingo!
Nearly finished all the plans, and your suggestion of converting them to Sline is going to be taken on board

The only drawback is after drawing the part and doing the polylines and the J command, when one of these miniscule gaps is there, i cannot delete either of the offending lines and redraw, as when i deleted it, the whole of the polyine vanished!!along with most of the drawing. So i did a copy file of the drawing and when i found one of these little blighters i amended the original, made another file to pedit and so nr. I think the highest redraw i got to was copy 15!!
r
I think i will have a go at the £D ones now, are there any freeware/shaeware ones about i could play with later on? You never know i may then convert my Warco WM16 to do it all for me, especially as i know 3/5 of sod all about electrics and electronics

Thanks again for the holding of the saddle and letting go when i wasnt looking

[Roger]

Dec 11, 2017 16:20:28 GMT silverfox said:

What i did find is how sensitive the joining of lines are. 

Usually there are symbols that pop up near the cursor to show whether you're over a particular feature, say the end of a line, or just on the line somewhere etc. You need to make sure that the symbol is showing before you click to start or finish a line for example.

You'll also notice that while you're sketching, most packages flash up guide lines as to what it might snap to. It's sometimes better to draw a feature, say a line, away from anything, being careful to make sure it doesn't snap to an existing feature. Then you can use a constraint to, say, make that a tangent to a circle. This is the way I construct things like the classic belt guard shape, ie two circles joined with lines at a tangent and the inner parts of the circle deleted.

[silverfox]

Roger
you would have had a fit watching me draw a tangent between two circles of different diameters and the centres were 2in apart on the X axis and 3in apart on the Y

Could find anything in the dummies book, so in the end i got a 6in ruler and put it to the screen to get an idea where the radius would intersect the circumference. One that was done it was 'easy peasy' not!
I was a few degrees out but not enough to lose any sleep. in fact what i did was to then trim the circle of the 'inner' part and then draw a line between the two remaining segments!!

Bit luddite but it worked

[Roger]

Dec 11, 2017 17:59:58 GMT silverfox said:

Roger
you would have had a fit watching me draw a tangent between two circles of different diameters and the centres were 2in apart on the X axis and 3in apart on the Y

Could find anything in the dummies book, so in the end i got a 6in ruler and put it to the screen to get an idea where the radius would intersect the circumference. One that was done it was 'easy peasy' not!
I was a few degrees out but not enough to lose any sleep. in fact what i did was to then trim the circle of the 'inner' part and then draw a line between the two remaining segments!!

Bit luddite but it worked

That made me smile. You'll gradually find the tools, I'm sure they're all there and it's worth taking the time to explore them.

[RGR 60130]

To join up two straight lines with a slight gap between them you can usually use the Chamfer command. You set each leg of the chamfer to 0 (zero) which actually produces a corner rather than a chamfer. Clicking on the Chamfer command then each of the two lines you want to join will join them.

Reg