EEVblog Electronics Community Forum
Products => Crowd Funded Projects => Topic started by: MacAttak on May 07, 2013, 01:28:58 am
-
http://www.kickstarter.com/projects/otherfab/the-othermill-custom-circuits-at-your-fingertips (http://www.kickstarter.com/projects/otherfab/the-othermill-custom-circuits-at-your-fingertips)
It looks much like many 3D printers, but is a CNC mill. Designed to drill and route PCB boards, but can work in other "soft" mediums too like wood. They claim that it can be used for metals too - but I am highly skeptical of that claim. Metal milling typically requires water cooling and some beefier hardware due to the higher torque needed and lateral stability.
It clearly is targeted square at the DIY electronic hobbyist as it appears to natively support Eagle files for input.
Generally speaking, I don't see major problems with the project itself. But not everything seems right.
The things that seem questionable to me are:
- This thing is based on snap-together HDPE plastic panels. I find it very hard to believe that it will be rigid enough for precision work.
- It probably should not be used for fiberglass boards, just paper ones due to the dust it will generate.
- It's not very big - you won't be working with large boards.
- In the pictures provided I only see three stepper motors (and they aren't that big). Not sure if that is enough to push a routing bit especially in X and Y axis?
- I find it odd that they are giving so much attention to OS X support for their front-end tool. They still support Windows, but it seems like OS X is where their heart is. Seems odd for a tool geared towards OS X over Windows in this Windows-dominated space.
- I can't imagine using this thing for metal milling. Wood, plastic, PCB would be fine... but metal?
It does look intriguing though. However the price is a bit too steep unless you are spinning boards very frequently.
(https://s3.amazonaws.com/ksr/assets/000/552/496/b0234ac36b5ecd467088e4c08cd11899_large.png?1367273662)
(https://s3.amazonaws.com/ksr/assets/000/568/342/52bba9e93e939767c8e15705ddddb69c_large.jpg?1367798900)
-
Also, an image of the resulting PCB quality. With a very dubious soldering job. Surely they would have done a better job with something that exemplifies their product results.
(https://s3.amazonaws.com/ksr/assets/000/564/026/a663d28510fef569d3c397bc9909f8dc_large.jpg?1367606106)
-
The traversal speed is pretty slow and given the size of the usual cutter, there shouldn't be any problem cutting copper. You probably wouldn't want a bit larger than 1/8 inch in it anyway.
A pretty clever design and it looks rigid enough for PCB fab, although the envelop is pretty small.
I'd wonder about long term accuracy since it's using what looks like threaded rod instead of ball screws.
-
Given the relatively small size of the mill, I doubt that being made of HDPE(?) will have a big effect on accuracy, at least when milling PCBs.
I wouldn't expect it to be able to handle materials other than plastic and balsa that well.
At my local hackerspace (http://hobarthackerspace.com/), we've been doing some PCB stuff on a 6040 CNC router (http://www.carving-cnc.com/) with acceptable results.
I'll find out the specs of the PCBs (track size etc.) at tomorrows get together.
-
I wonder how do they get this thing flat enough for pcb milling. This job requires <1mil flatness and squareness. Especially when going with wide angle engraving bits (which last longer).
-
It doesn't sound unachievable, I mean, at the price level (not exactly cheap, but not absurdly expensive) I'm sure that a CNC mill that looks like a 3d Printer can be made for that.
I'd question the precision though, they make a bit of a deal about it being light and portable, "You can grab it and take it on the bus, toss it in your trunk, or put it in your bike basket."... yeah, none of those things sound like something that would be good to subject an expensive machine which needs to operate in a precise and *repeatable* manner. If the structure is really that interlocking plastic (rather than just being a shell), nup, I just don't see that being resilient enough to work "fully assembled out of the box" like they want.
That said, never really can see the attraction of milling PCBs myself, slow, and I don't see how you can get anything like the tolerance you can with even toner transfer, let alone photolithography. The only thing it has going for it is no chemicals and you can make the mill drill through holes for you with a bit change (absurdly, none of their samples are through hole drilled that I saw!).
Now, ignoring the precision issues in the design for the moment - if they made it so that you could swap out a milling head, for a 3d printer hot end, easily... that would be a winner right there I reckon. It's got to be do-able, a single machine that can both mill and extrude. The mechanics are essentially the same except for the tool head. Add a laser cutting tool head as well for a real factory-in-a-box.
-
Nice toy.
Specially when you keep in mind, that a CNC self-assembly kit made of steel costs about three times more.
But this mills made of steel are able to cut non-ferrous metals without problems.
This CNC is nice when the goal is to engrave.
But when the goal is to mill more then balsa, this CNC will be overloaded.
A hotend in this plastic-toy would not end good.
The heat could soften the CNC.
It would be better to put a milling spindle in a small makerbot.
I knew, I am a little bit crazy.
But like Dave would not use el cheapo multimeter, I will not work with this type of CNC.
A CNC milling machine should be at least as heavy that to move a pallet jack is required.
-
- It probably should not be used for fiberglass boards, just paper ones due to the dust it will generate.
From their FAQ:
Some of it is. That’s why we include non-toxic PCB FR-1 boards with every mill. NEVER EVER mill anything with fiberglass in it, including PCB FR-4 stock. Fiberglass dust is poison. Seriously, it killed my grandpa (this is Martine talking), don’t do it.
-
Looks nice, but at the price of 2 layer boards from Seeed/iTead etc. it just doesn't look worth it for $1,000. I can see it being useful if you make hundreds of prototypes over it's lifetime.
-
This thing looks neat, but I can't see it working well as a PCB router - and certainly not as a metal cutter.
The dimensional stability of plastic with temperature is terrible, as is the rigidity. Looks like ACME screws instead of ballscrews, so they will likely be using software backlash compensation, which means positional error. Add in the flexibility of the structure and that means more positional error... add in the rigidity of the spindle and that's more error. When milling PCB's, the loads are not very high, but you must be very precise - which is what I think this machine won't be. As someone else said, you need those PCB's ultra flat - sticking it down isn't really adequate, so I am not sure what method they use to adjust for flatness or even to measure it.
I don't understand at ALL why they are using Mac control only? The vast majority (almost all of it, actually) of stuff done with CNC is on Windows or sometimes Linux. I can't think of any reason anyone would choose to do this on Mac unless that's all they knew, but it seems really strange to me. MANY more features could be done using something like Mach3 or EMC. I wouldn't waste time making it read Eagle files natively - use Gerber, way more useful and universal.
I wish these folks luck, but I foresee problems with their approach and their hardware.
-
Looks nice, but at the price of 2 layer boards from Seeed/iTead etc. it just doesn't look worth it for $1,000. I can see it being useful if you make hundreds of prototypes over it's lifetime.
I agree but I think one unspoken benefit of milling boards yourself is the turnaround time. There is a certain non-zero value in that.
-
Regarding flatness, how critical would that actually be for a PCB mill? It seems that it wouldn't matter much if it milled a little too deeply into the board in a few paces... since you would only be milling 1-layer boards (or maybe 2-layer if you can get really good horizontal alignment).
-
Most cutters are conical cut with a flat tip therefore the width of cut is directly proportional to the depth of cut so it is a big deal if you want to do relatively tightly spaced traces with reasonably controlled widths.
-
Uh.. As far as I can see, this one ain't worth the price. It's just small, flimsy and I seriously doubt it'll be good enough even for PCB making. :--
-
Regarding flatness, how critical would that actually be for a PCB mill?
Very. Bits are usually conical so depth controls width. Additionally, the angle of the tip tends to be large (60-90 degrees) so they don't break off too easily. The tip will probably have a width of 0.2mm (they are never a proper point), and simple maths tells you that a slight change in depth can make a thin track disappear. A board I did last week of mostly DIL but with a SO16 package had a 15mil cut width from a cut depth of 0.09mm, as an example.
As you might imagine, there tends to be a lot of focus on sticking the PCB down flat on the mill bed (and assuming that the mill bed is itself flat and perfectly perpendicular to the Z axis). Indeed, I have tried most things and settled on carpet tape, which works brilliantly for me but may not for anyone else. However, an alternative way of attacking the problem is to compensate in software for the non-flatness of the PCB.
The method I am aware of, which I know is used by someone making boards for commercial products, is to use the milling bit as a touch probe, and then use that to map the surface of the PCB. The result of that is then used to modify the PCB CNC depth data, and the milling bit then tracks the surface of the PCB perfectly. The process of mapping the PCB surface is quite quick, and the grid used doesn't have to be very fine - 10mm x 10mm is apparently easily good enough for a PCB warped by not storing it flat.
So, whilst on the face of it the precision required is quite high and this plastic mill may seem not to meet that, it is a resolvable problem so long as the imprecision is repeatable (that is, so long as it doesn't change as things move around). A non-flat but reasonably rigid bed doesn't have to be a problem.
-
If you mill a flat surface to put your pcb on, it will be completely flat. I'm typically using just a MDF board, mill it flat, use a 40 degree V-bit, 0.2 mm tip width, and that gives very good results, looks like factory produced. Then remove excess copper with a 1.5 mm end mill, drill with 0.8mm drill bit, and cut the board to size with a 1/8" end mill.
You can use a tool like http://rapid-pcb.com (http://rapid-pcb.com) to generate everything for you. The flat surface, the signal contours, drills and cuts, and it also does the depth * angle = surface diameter for you and shows that in the preview.
(http://rapid-pcb.com/screenshot.jpg)
Best regards,
Paul.
-
If you mill a flat surface to put your pcb on, it will be completely flat
Indeed, your flat surface will be very flat. The trouble is that PCB's can be not-flat, having warped a little during storage, so they don't sit flat on the nice surface. Just a mil or two lift can make or break a PCB with 10mil track/gap.
There are solutions: stick the PCB down with tape or superglue, suck it down with a vacuum, etc. None of them seem to beat the surface scan and gcode adjustment, though :)
-
Ah, sure, I forgot to add that I use double-sided tape. It seems I need to try your method too, with the surface scan. :-)
I've attached some images of a board I just made.
Best regards,
Paul
-
They look very nice :)
Er... doesn't it cost you a fortune in tooling to rub out all that copper, though?
-
Don't know yet. But I've milled ten of those boards above with the same tool, and I can't see a wear or difference in the result. The rubbing takes only 3 minutes, so it's pretty quick, much quicker than the signal contours.
If I can do 10 pcbs with one 1.5 end mill, and I can buy a 10-pack of those at Amazon for $16 (e.g. http://www.amazon.com/dp/B00E5X7844?_encoding=UTF8&camp=15041&creative=373501&linkCode=as3&tag=tube00-20) (http://www.amazon.com/dp/B00E5X7844?_encoding=UTF8&camp=15041&creative=373501&linkCode=as3&tag=tube00-20))
, then the cost per PCB is not very high (IMHO). I'm just a hobbyist, though, so I don't create hundreds of cards :-)
Best regards,
Paul.
-
Paul, did you use the Otherfab to mill those boards, or something else?
-
For those I used a K2 CNC KT-2514S, a Bosch 1617EVSPK router, high-precision collet, Eagle for the PCB design, rapid-pcb.com for the GCode and Mach3 for the actual running of the machine. Tools: Flat milled with 1/2" end mill, contours with 45degree 0.2mm carbide V-bit, rubbing with 1.5mm carbide end mill, drilling with a 0.8mm drill bit and finally cutting the board with a 1/8" end mill. Double-sided tape for holding the PCB down.
I also adjusted the accellation in Mach3 down to the minimum (3). For that particular board, it took 40 minutes from start to finish, including the adjustments, milling the flat and all tool changes. No processing was done on the board afterwards, no sanding or anything like that.
Best regards,
Paul :-)
-
it took 40 minutes from start to finish
Ah! I was thinking that 3 mins was very impressive indeed, but now see you did say that was for the rubbing out. My mill is retrieved from the bin :)
-
Looks nice, but at the price of 2 layer boards from Seeed/iTead etc. it just doesn't look worth it for $1,000. I can see it being useful if you make hundreds of prototypes over it's lifetime.
I agree but I think one unspoken benefit of milling boards yourself is the turnaround time. There is a certain non-zero value in that.
I can photoetch a board in a fraction of the time/cost of a miller and can do boards up to A3 size (although have never gone that big, nearly A4 size is the biggest I've done). Can also do double sided (although not thru plated as yet) it doesn't even have to be a messy process (milling isn't exactly clean given the dust produced) and I could make a HELL of a lot of boards with the $1000 for a tiny PCB mill.
http://www.flickr.com/photos/23129101@N00/sets/72157623875853916/ (http://www.flickr.com/photos/23129101@N00/sets/72157623875853916/)
http://www.flickr.com/photos/23129101@N00/sets/72157623874786792/ (http://www.flickr.com/photos/23129101@N00/sets/72157623874786792/)
http://www.flickr.com/photos/23129101@N00/sets/72157625931640050/ (http://www.flickr.com/photos/23129101@N00/sets/72157625931640050/)
Note I have an updated etch tank (glass tank from IKEA since these photos were taken), The vacuum pulldown UV box took a day to make and cost about £100 in parts.
I really don't understand the obsession on the web with PCB milling when there are far better, quicker, cheaper methods.
Martin.
-
there are far better, quicker, cheaper methods
Have you milled any boards?
I have the bubble tank and everything out in the garage, which is what I used to use before I tried milling. Haven't touched it since because it's such a drag hauling it out, plumbing it in, disposing of the waste, carefully not splashing any etchant around the kitchen, etc. And at the end I still had to drill the off-centre holes. Plus generate a workable printout, expose the boards, etc. There are more steps in etching, and each step leaves some physical thing to dump when you spot the mistake just too late. With milling there is nothing physical until you start the drilling (although running an end mill through some clamp can be a bit more expensive then throwing away a printout, admittedly).
Sure, if I wanted to do A3 boards I probably wouldn't mill them (perhaps - ask me again if I ever need to do some), but the fact is that most of my boards are smaller than 100mm x 100mm, so that kind of potential isn't going to feature on my radar. YMMV, of course.
OTOH, I'd like to see you etch a board with an irregular outline and a big hole in the middle.
doesn't even have to be a messy process
It is a wet process, which is enough to be a major drag just from the disaster potential, never mind if you actually do spill something.
milling isn't exactly clean given the dust produced
A car vacuum cleaner, switched with the spindle motor, takes care of that here. Indeed, my PCB mill sits on the same desk as my programming PC, and leaving it milling as I program doesn't require me to don a mask of any sort. I don't even hoover especially afterwards.
I could make a HELL of a lot of boards with the $1000
itead could run off a HELL of a lot of baords with the $1000, and they would be double-sided with solder mask and silkscreen. We must be mad to make our own (etch or mill), right? No, there are many considerations, of which up-front tool cost is but one.
I suspect that the best method is whatever you use, because you are already au fait with the process. If your method works well for you, it is hard to see the benefit of a completely different method, but that doesn't mean that it doesn't work well for someone else. And just because you might do this board with that process doesn't mean you will forever do every board the same way. Horses for courses.
-
I can photoetch a board in a fraction of the time/cost of a miller and can do boards up to A3 size (although have never gone that big, nearly A4 size is the biggest I've done). Can also do double sided (although not thru plated as yet) it doesn't even have to be a messy process (milling isn't exactly clean given the dust produced) and I could make a HELL of a lot of boards with the $1000 for a tiny PCB mill.
http://www.flickr.com/photos/23129101@N00/sets/72157623875853916/ (http://www.flickr.com/photos/23129101@N00/sets/72157623875853916/)
http://www.flickr.com/photos/23129101@N00/sets/72157623874786792/ (http://www.flickr.com/photos/23129101@N00/sets/72157623874786792/)
http://www.flickr.com/photos/23129101@N00/sets/72157625931640050/ (http://www.flickr.com/photos/23129101@N00/sets/72157625931640050/)
Note I have an updated etch tank (glass tank from IKEA since these photos were taken), The vacuum pulldown UV box took a day to make and cost about £100 in parts.
I really don't understand the obsession on the web with PCB milling when there are far better, quicker, cheaper methods.
Martin.
Wow, that toaster oven has seen a few too many Hot Pockets :)
-
I used to etch my DIY boards, but after building a decent CNC a few years ago, I will never go back to etching.
Sure it's not all clean and the dust is not healthy at all, but I got mine enclosed in a box with an acrylic window.
For mounting, I use a scrap piece of wood on the bed that I plane once a month or so, then simply double sided tape to hold the board in place during milling and drilling.
Only takes like 20-30 minutes for a medium sized board from start until finish with holes and everything.
I don't remove all the copper, just enough to make sure there are no shorts and it's easy to solder.
Like this one:
(http://imageshack.us/a/img707/73/k6a2.jpg)
Sure the cost of a decent setup would pay for a lot of manufactured boards, but the convenience of just doing a quick layout and having something you can start soldering within an hour is very nice.
I am however not convinced the advertised "Othermill" machine will work very well, also it seems very expensive for that size and materials.
-
Have you tried those tiny packages like 0.4mm ptich TQFPs?
-
This was a successful prototype for a LGA-14 (0.8mm pitch) package. It would be tricky to go smaller, I reckon.
-
You can now pre-order these outside of kickstarter. I had considered one for quick small jobs, however, looks like they've raised the price, currently $2200.
http://store.othermachine.co/products/othermill (http://store.othermachine.co/products/othermill)
Oh well, guess the search is still on.
-
You can now pre-order these outside of kickstarter. I had considered one for quick small jobs, however, looks like they've raised the price, currently $2200.
http://store.othermachine.co/products/othermill (http://store.othermachine.co/products/othermill)
Oh well, guess the search is still on.
The 6040 mill that GeoffS linked to above is $1.5k and looks a lot more sturdy and useful to me.
-
Just thought I'd add some information on the OtherMill. [Disclosure - I bought one through Kickstarter, and liked it so much that I do some consulting for them now.]
I wrote up a post that shows how I made one of my projects with the machine I bought. Here is the link:
https://medium.com/@tomsimon/lipo-charger-board-for-the-arduino-nano-39e709f459bc
I have done larger projects including things with cut-outs for 10W LEDs, or weird shapes and various connectors, such as micro USB.
Thanks
-Tom
-
I wrote up a post that shows how I made one of my projects with the machine I bought. Here is the link:
https://medium.com/@tomsimon/lipo-charger-board-for-the-arduino-nano-39e709f459bc
Stencil for this board ? srlsy ? Or is it just to show capabilities of CNC machine ? :)
-
I know it seems a bit much, but I show it in my write up mostly to demonstrate the technique. I found it much easier than using a toothpick to apply solder paste. Once I made my first stencil there was no going back for me. Now I just do it as a matter of course. It does not take much to make the stencil - maybe 15 minutes total time spent. For my projects, mostly LED things, I make about 4 to 10 of them, so it is well worth the time to make a stencil.
-
I wrote up a post ...
If I didn't already do this I'd want to do it after reading that :)
As for the stenciil, if it's no trouble then why the hell not. Nice job :-+