help buying a good PCB mill

[mrpackethead]

This may be me whinging a little bit, but the most amazing thing in the world for me would being able to get prototype boards at a reasonable price at about the same speed Digikey can get parts to my house.

I mean, having to work out a schedule for various projects (schematic for project A, layout for project A, order boards for A, schematic B, layout B, order boards B, receive boards A, test boards A, schematic C...) isn't the worst thing in the world, the context switch is a bit painful.

But... cost-wise, I don't yet have enough projects going through to justify the cost (both training and equipment-wise). The boards from seeed come fast enough and are cheap enough.

As a curiosity, has anyone tried the Othermill for doing prototype boards?

Just stop. milling pcbs will just create you time delays and fustration.

[FabLabMaster]

  My own pursuit of a good PCB router lead me to this post. I've had an LPKF S63 for about 5 years now. The joke here at work is "it works great... when it's working". During the 5 years, I've had to send the milling head back 4 times. Each of those times was due to "manufacturing and/or design flaws" and each time, the machine was down for several weeks including the week of troubleshooting and dealing with their tech support. This does not go over well with my superiors. Also, the machine is "dumb" and can damage itself if an error occurs and you're not right there to babysit and, trust me, there are plenty of errors to go around and plenty of expensive damage that can occur (I got this machine to free up time, not devote even more time to PCB production and machine troubleshooting and repair). The software (CircuitPro) is not very user friendly and it's always being updated for bugs, etc. The software throws errors almost every usage and nowhere can you find info about the errors. I've stumped their tech support many times.
  I've learned many of the machine's quirks but, so many of these errors are unexplainable by me or LPKF tech support. I've been looking into Accurate CNC for a replacement as the ROI on the S63 has been satisfied. I want a machine that is smarter with less downtime and software that makes more sense and is less quirky. LPKF sold me on their machines ease of use telling me that high school kids were setting these machines up and cutting PCBs in a day. That didn't happen for me as they shipped me a machine that had not been calibrated and was missing the vacuum table I bought. Took me over a week (with tech support) to figure out why the machine was not working correctly. Tech support had to guide me by phone through the entire factory calibration process to get it working (they are in OR and I am in NY which leaves a 5 hour window, when they finally get back to you). Like I said earlier, the machine works great... when it's working. But, you just never know when the next error or tool change collision will happen or why. I'm looking at AccurateCNC now.
Best of luck to you whichever way you decide to go.

[BMF]

I've used LPKF mills (c30 -s100) for about 15 years. I've had very good luck with them and I respect the company. I would caution that the economics of making boards in house is not what it was 15 years ago. They are very good for low volume simple boards like adapters and antennas but complex boards can be very costly. If have made single boards that cost $150 in tooling. Tooling costs, design limitations, operator skill requirements, and unique design for manufacture issues are serious limitations. I now send out most of my boards for fab. I'm probably not going to replace the machines if they ever wear out. If you choose to buy a machine be aware that the software is very important. LPKF and MITS are good machines from what I know. LPKF software is very good. I have no knowledge about MITS software. You'll want a dedicated PC, sound proof enclosure, quality substrates and tooling, a rivet system like EasyContac, a tool to set ring depth, inspection microscope with reticle scale, and an extra hose for the hepa vac to do cleanup.   If you want to bump things up a level you can do hole plating and masks but be prepared for additional labor. The economics don't justify those steps for me. A second issue is design flow. If you are designing your board based on prototyping equipment and not production equipment you are not in a direct path. ANY edit to a pcb design has risk. I like to avoid that design shift risk whenever possible. On the other hand the machines are really nice for making PCB mockups to mount connectors and controls to verify the locations for mechanical fit tests. There are also handy for building test jigs.   

[mrpackethead]

I've used LPKF mills (c30 -s100) for about 15 years. I've had very good luck with them and I respect the company. I would caution that the economics of making boards in house is not what it was 15 years ago. They are very good for low volume simple boards like adapters and antennas but complex boards can be very costly. If have made single boards that cost $150 in tooling. Tooling costs, design limitations, operator skill requirements, and unique design for manufacture issues are serious limitations. I now send out most of my boards for fab. I'm probably not going to replace the machines if they ever wear out. If you choose to buy a machine be aware that the software is very important. LPKF and MITS are good machines from what I know. LPKF software is very good.

The software that LPKF supplied me was so good that after spending 3 hours milling it went and milled a 45 degree line right across my board.  Wonderful. rendering it useless.  It was so good that peridically it would drive tools way too deep.

My experience of LPKF was absolutely hiddious. 

[mikeselectricstuff]

On the other hand the machines are really nice for making PCB mockups to mount connectors and controls to verify the locations for mechanical fit tests. There are also handy for building test jigs.

I have a cheap Chinese 3040 router which is also pretty good for drilling jigs and doing outlines. In principle it may be feasible to rout PCBs on it but I don't have the patience to tweak it.
That being said the PCB results I've seen Micah Scott get on an Othermill have been surprisingly impressive.
 

[BMF]

[/quote]

The software that LPKF supplied me was so good that after spending 3 hours milling it went and milled a 45 degree line right across my board.  Wonderful. rendering it useless.  It was so good that peridically it would drive tools way too deep.

My experience of LPKF was absolutely hiddious.
[/quote]

I had a similar crash once on the c30. The second time I ran to board I had no problem. I used the same file so I assume the data was good but somehow the PC to machine communication corrupted the data. It could have be a weird corruption in the machine. I haven't had that issue on the s100. My biggest complaint is in managing tool wear. I find it best to start complex boards with new cutters and endmils. I don't like second passes at a phase because the machine head does not track as well on the air cushion. I save the used tools for small boards. It is difficult keeping track on tool life. I usually just use a microscope to read the tool life.   

[technotronix]

What about Prometheus Desktop PCB Mill? Check more from here. http://makezine.com/2015/11/16/first-look-prometheus-desktop-pcb-mill/

[G0HZU]

I've used LPKF mills (c30 -s100) for about 15 years. I've had very good luck with them and I respect the company. I would caution that the economics of making boards in house is not what it was 15 years ago. They are very good for low volume simple boards like adapters and antennas but complex boards can be very costly.

Yes, that's my view too. At work the PCB mill has always been a highly valued piece of equipment by our RF team because you can try out various RF filter/oscillator/switch concepts and do several iterations in one day on various exotic RF PCB laminates.

I suspect that the naysayers on here who think these machines are a waste of time will still not understand what I'm talking about. If I was in the business of making two layer digital boards on FR4, eg microcontrollers, CPLDs etc then I would use the mill a lot less and I'd just order a PCB from a PCB company. But to approach a PCB company to do several iterations of a critical RF filter design on an expensive PCB material would be crazy in terms of cost and timescale (and frustration at all the waiting around for the next version of the dev board to arrive).

Also, the mill can be a great tool to have if you want to make a little piggy/bandaid PCB to diagnose/fix an issue on an existing PCB. This doesn't have to be an RF PCB, it could be a mini piggy PCB on a digital board. The time taken from spotting a problem to producing a PCB design and mill it and build it and test it and prove it can be less than a day.

[nctnico]

For small non-critical boards you need quick it is probably easier & cheaper to just etch the PCB and I don't mean messing around with household chemical in an old jar but a decent professional setup. For RF however milling is the only option I know to get a good enough accuracy.

[G0HZU]

Agreed but the mill can also make the little piggy PCBs with custom PCB outlines. Eg L shaped or with round or square inner access holes. It will rout any PCB outline you want within reason to fit inside a tight space. If you went down the etching route you could cut the shape by hand with a saw but it would be very fiddly to do. They are great for making quick and dirty FR4 test jig PCBs or laminated tools as well but only if you are in a hurry. Usually it's much better to get a real test jig PCB made with FR4 with silk screen etc.

[edavid]

What about Prometheus Desktop PCB Mill? Check more from here. http://makezine.com/2015/11/16/first-look-prometheus-desktop-pcb-mill/

What about it?  It's been over a year since that article, and it still hasn't shipped 

[mrpackethead]

Also, the mill can be a great tool to have if you want to make a little piggy/bandaid PCB to diagnose/fix an issue on an existing PCB. This doesn't have to be an RF PCB, it could be a mini piggy PCB on a digital board. The time taken from spotting a problem to producing a PCB design and mill it and build it and test it and prove it can be less than a day.

Agreed. If it worked.  Sadly the LPKF stuff never worked. The Support was atrorious.   Left a very bad taste in my mouth.

[G0HZU]

Also, the mill can be a great tool to have if you want to make a little piggy/bandaid PCB to diagnose/fix an issue on an existing PCB. This doesn't have to be an RF PCB, it could be a mini piggy PCB on a digital board. The time taken from spotting a problem to producing a PCB design and mill it and build it and test it and prove it can be less than a day.

Agreed. If it worked.  Sadly the LPKF stuff never worked. The Support was atrorious.   Left a very bad taste in my mouth.

I think some of the issues with the newer/better? machines may be that LPKF are trying to make the machine into a PCB vending machine rather than something that relies on operator skill. So instead of having an experienced operator setting up the equipment for registration or tool depth using manual (slow) methods they have tried to automate it all on the newer machines. Add to this the buggy nature of modern PCs and operating systems and it must be harder to get it all to behave perfectly for an hour or more.

I'm still using Win 98SE on my T-Tech machine here at home and everything is manual. I have to fit each tool by hand and adjust the tool height using a vernier and test it in a spare area before running that isolation layer. Registration is done using pins/holes in the PCB rather than the modern optical system and the vacuum table. This requires a lot of experience and sympathy and skill but it does give good results. Those new automated machines look like luxury to me by comparison

[mikeselectricstuff]

I think some of the issues with the newer/better? machines may be that LPKF are trying to make the machine into a PCB vending machine rather than something that relies on operator skill. So instead of having an experienced operator setting up the equipment for registration or tool depth using manual (slow) methods they have tried to automate it all on the newer machines. Add to this the buggy nature of modern PCs and operating systems and it must be harder to get it all to behave perfectly for an hour or more.

MCUs these days are so cheap that to make a machine like that that doesn't do it all standalone is insane. cf. 3D Printers.

[tpowell1830]

hmm, with that kind of budget should you consider green tape?   

PEACE===>T

[scopeman]

I personally own two older LPKF machines that I picked up on the surplus market a 91s and a 92s. They needed a fair amount of TLC to get going. They aren't perfect but make pretty good boards.

I use standard FR4 and Chinese TiN coated 60 degree 0.1 or 0.2mm "V" bits (I have my own method of setting the initial depth as they are shorter than the LPKF) and 1mm end mills for mass copper removal and either 1mm, 1.5mm or 2mm for contour and/or inside cuts. The quality of the TiN coated Chinese bits vary widely so it took a while to find good ones.

I have a few "real" LPKF bits and routers but are reserving those for a critical job. I think that the PreciseBits brand are very good. There may be other brands as well.

Thus far I have made a few dozen boards with it from 0.031 and 0.0625 FR4. My local PCB supplier let me go thru the scrap bin and also sold me 30 3 foot X 4 foot sheets of CEM1 for five bucks a sheet so I have a lifetime supply.

I get fairly good results and have made PCB's with parts as fine as 0.020 pitch. Standard 0.050 is easy. Lines to 8 mils are not a big deal and I have done 6 with the 92s.

I even did a test board with traces as narrow as 2 mils but I would not recommend it!

I designed and had made a special T-Square to align the machine so that both sides would line up properly on double sided boards. A bit tricky to do but worth the effort.

And G0HZU is right. It takes real attention to detail to get the desired results. I wrecked a few boards and it took a while to get the hang of it but thus far I am happy with it.

Thus far (unlike the original owner of my 91s) I have not plowed the mill or universal cutter into the machine bed.

I really like the ability to design in the morning and have a proto ready by the afternoon. I know you can do this with the chemical method (been there, done that) but if I need that level of attention the board design will go out of house.

What I really like the machine for is for one of a kind boards like those used for test fixtures, card extenders, mechanical placement fixtures, test probe (spring pin) boards, small front panels made from FR4 (or even engraved plastic or aluminum panels).

With all that said I think I would take a serious look at the AccurateCNC product. They look like they are a bit more polished than my LPKF's (and a couple of decades newer).
In all fairness I do not have any experience with the newer LPKF machines. They may be easier to use but I am sure that I would be locked into their tooling and expendables which would be quite expensive for hobby/ham or occasional use.
 
If you are using these machines for commercial use you probably should figure in a budget for board material and tools from the machine manufacturer as this will keep the "finger pointing" to a minimum.

Now for the options: Microscope, Fiducal recognition camera, etc... Like buying a car (and you can buy a really nice one for the cost of these machines new!).

Best of luck,

Scopeman

P.S. I would not expect to make other than single or double-sided boards with these machines. The multi-layer process is way too complex with too many pitfalls, so unless
you are making boards for a spook government project sending the gerbers out is your best bet.

[johnyradio]

As a curiosity, has anyone tried the Othermill for doing prototype boards?

Yes. And, although there was some fiddling to get all settings correct (first board not good), my second board was sweet. I could see the need for improvements in the hardware and software, but that was 2 years ago. i'm sure they've improved. My board was a simple thru-hole board, so i can't vouch for a dense board with SMD parts.



True that mail-order PCB fab is absurdly cheap these days. Def the way to go for production runs.

But if, like me, you want to use fresh PCB's as part of your design and development process, you want to do quick prototyping at home. Then messing with Ferric Chloride can get old.

Old timers here are, i'm sure, accurate in their descriptions of bad experiences with older machines.

But i think quick one-off prototyping is a real niche, and a new generation of companies are filling that need. I think these co's understand that people want convenient, tidy, user-friendly hardware and software. The OtherMill (now named Bantam) is cute and sits on your desktop.

Another one:

- allows you to work with milling tools down to 0.178mm (7 thousands of an inch, aka 7 mil).
- makes real circuit boards in minutes.
- any surface mount component.
- High spindle speed lets you mill a 3?×5? board in minutes
- Price point is thousands less than machines of similar accuracy.
- hard to argue with their video. https://youtu.be/Tp81Aneil24
- $1,800.
https://makezine.com/2015/11/16/first-look-prometheus-desktop-pcb-mill/

Unfortunately, as noted above, the prometheus has not started shipping yet. The co just emailed me, says "slowly" shipping first batch now, short lead-times hoped for Feb '18.

OtherMill/Bantam has been shipping for a couple years now.

Another option:
- shipping now
- plus paste applicator
- plus pick-and-place
- €2,000.00 (2,000 Euros = ~ $2,400 USD)
https://youtu.be/VWi9GUzY1fY
https://cirqoid.com/

[lectrohamlincoln]

Hi Scopeman,

I was hoping you could elaborate a bit more on your Protomat setup (drill height setting and double sided registration). I picked up an old Protomat 93s and have had pretty good success milling singled sided boards with a 0.015" end mill. I've found it difficult to set the tool depth correctly and I haven't had any luck with "V" bits. Sadly, this setup does not currently allow me to mill boards with QFN 0.5mm pitch. My Protomat did not come with the alignment strip so I haven't tried double sided boards yet either.

Also Scopeman or G0HZU, how do you calculate the feeds/speed for "V" bit and micro end mills?

In case anyone was wondering, I've had no problems running the Protomat 93s from a modern Windows 10 laptop with a usb->serial adapter. I use Kicad to generate the GERBERs, LPKF CircuitPro PM 2.3 to generate the isolation paths, and BoardMaster 5.1.214 to control it. 

[scopeman]

My ramblings on the LPKF:

On the Protomat settings.

The drill height setting is limited by the mechanics. I check this by installing a 38mm (standard length, verify with a caliper) drill, set the depth limiter all the way to the top, and move the drill head over the platen (Drill motor off!, you don't want to drill the table) and watch as you manually press the motor down. The tip of the drill must clear the platen (table surface) by 0.5mm. You can use a feeler gauge or a piece of brass stock to verify this. If you have to adjust the motor height there is a locknut on the top of the assembly that allows you to adjust this.

Normally when drilling the depth limiter is set all the way to the top. When I am processing thin material I sometimes set the depth limiter so that it will limit the penetration.The smallest drill I have been able to use with my unit is 0.45mm (#77 IIRC). My units have the 20K RPM fixed speed spindle so I am limited as to the drill speed and sizes I can use.

Make sure that you first run the "Marking Drills" run as the first cut after milling cutter (V bit) milling cutter setup. This will insure that the drills will drill properly centered. I have my BoardMaster set to use the same cutter that I use for milling both sides. Since this is a quick process it does not detract from the life of the cutter very much, and there is no need to change tools.

Milling setup

I set the milling depth by running the depth limiter down (while in the tool change position) so that the tool bit tip is not exposed. Then I use a 1 inch square or scrap of PCB board held tight against the bottom of the depth limiter and I loosen the allen screw to allow the tool to drop onto the 1 inch square Cu. This sets the tip of the tool even with the edge of the depth limiter. This works really well with 60 degree TiN coated chinese tools as the length is not the same as the LPKF tools. I then rotate the ring on the depth limiter 8 clicks (1mil) and make a test cut and check it with a microscope. I then move a click at a time until I get a depth that gets me 0.1mm width cuts.

The trick with the depth setting is to make sure that you are only removing the copper and not cutting into the FR4 (if so only a fraction of a mil, a click or so). FR4 dulls the milling tools. If the tool acts like it is plowing rather than producing smooth cuts, chances are that you are cutting too deep. Once a bit has been run in this condition it is usually toast and will never cut a fine burr free line. I suspect that could be the issue with you Chinese tools. You should be able to slide your finger over the cut line, it should feel smooth. Inspect it with a microscope. Taylor Tools has some nice inexpensive pocket ones on eBay.

I always clean the PCB material after it is taped down with a white (very fine) Scotchbrite(R) pad first and vacuum away any dust and then wipe the board down.

I have also used 3-in-One dry lube (coat the side and remove all excess with a cotton ball or better yet a microfiber cloth, Harbor Freight or WalMart) for some stubborn FR4 material. I use the dry-lube as it will not collect chips. I use the same dry lube on the depth limiter threads and on the linear bearing rods.

It goes without saying that you will need to run a vacuum anytime the spindle is cutting as chaff from the process will get in the way of producing a good board.

I do my boards out of the normal sequence in that I mill one side then I flip and mill the opposite side so that I do not have to reset the depth setting or change the tool. I also start with a fresh tool so that I won't have to change one mid project.

When I mill I will do a few test milling of an area to confirm my depth setting as sometimes it may take a click or two more to get consistent results. Depending on how your drilling data is set up you may have to flip the board a last time to drill and route.

Pay attention to the colors on the screen to make sure that you are drilling/routing the right side! Mark the panel with a sharpie as to the side you are working on. Also when you start and when you flip make sure that the red strip is tight against the reference pin in the slot in the table.

You will also have to do the machine home setup to make sure that when you flip the board and drill that the holes will line up. This is all in the manual. Probably the most tedious setup of the unit. Really critical for small via hole front to back alignment.

Now with all of this being said, most of the time I use milled boards for one off protos or boards for test fixtures or mechanical setups. I once made some internal guide plate parts for locks for a friend out of FR4 and some huge FR4 fender washers for my shower to keep the inside rail from cracking through the plastic wall. They will never rust and I can have an infinite supply. The machine is very accurate for milled routed cuts so you can easily make things like solder together RF tight boxes out of double sided FR4.

If I need any quantity of complex PCB's it is cheaper to go to PCBWAY or a similar source unless you absolutely
must have something that day as you will spend the greater part of a day knocking out that PCB. This is especially true if you do not run the machine on a frequent basis as it is easy to make an error or omit a step.

Sam
W3OHM