[Jibeji]

Hello,
I have bought those cheap chinese 0.8mm carbide end mills for drilling and cutting PCB.

https://fr.aliexpress.com/item/Livra...670232400.html

Do you have any experience to share in terms of feed rate, plunge rate and RPM ?

[cj7hawk]

Hello,
I have bought those cheap chinese 0.8mm carbide end mills for drilling and cutting PCB.

https://fr.aliexpress.com/item/Livra...670232400.html

Do you have any experience to share in terms of feed rate, plunge rate and RPM ?

I'd love to know this as well, especially using it to route out larger holes (eg, 1mm, 2mm, 5mm etc ) - And whether it can handle it in larger holes through the entire PCB or whether it needs to go through in stages.... I tried a 1.8mm cutting tip and tried to cut through a full PCB in one operation, and the chatter was horrendous even at slow speeds. But in 7 steps, it was fine, even at full speed.

David.

[CitizenOfDreams]

I cut PCBs with carbide chip breaker bits like the one pictured below. My parameters are:
Spindle speed 8000 RPM
Feed rate 300 mm/min
Plunge rate 150 mm/min
Depth of cut 0.2mm
My CNC is just a toy hobby machine, so I have to use very shallow cuts.

[Jibeji]

Thanks for your info.
I tried with the following parameters, and it works pretty well:

12000 RPM
Feed rate: 200 mm/min
Plunge rate: 30 mm/min
Depth: 1.5 mm

[cj7hawk]

I cut PCBs with carbide chip breaker bits like the one pictured below. My parameters are:
Spindle speed 8000 RPM
Feed rate 300 mm/min
Plunge rate 150 mm/min
Depth of cut 0.2mm
My CNC is just a toy hobby machine, so I have to use very shallow cuts.

Thanks for the details - pretty close to what I do with a 1.8mm, although you're using a 2.4mm I think.

Though the .8mm ( and especially the .4mm ) look so thin and fragile. It just seems that there's no way they are going to handle 300mm/sec. Have you ever tried a sub-1mm bit? I've ordered a few to play around with but I'm not confident they will be any good.

David

[cj7hawk]

Thanks for your info.
I tried with the following parameters, and it works pretty well:

12000 RPM
Feed rate: 200 mm/min
Plunge rate: 30 mm/min
Depth: 1.5 mm

Was that with the 0.8mm?

Thanks
David

[CitizenOfDreams]

Though the .8mm ( and especially the .4mm ) look so thin and fragile. It just seems that there's no way they are going to handle 300mm/sec. Have you ever tried a sub-1mm bit?

I only use a 1.8mm chip breaker for routing PCBs. For smaller holes I use drilling (a set of carbide drill bits sized from 0.3mm to 2.0mm).

[Jibeji]

Was that with the 0.8mm?

Yes David

With 0.8mm end-mill :
12000 RPM
Feed rate: 200 mm/min
Plunge rate: 30 mm/min
Depth: 1.5 mm

[joeavaerage]

Hi,
I've used two flute endmills, 0.4mm and 0.5mm quite extensively. I use Kyocera-Tycom endmills from 'drillman' on EBay. The last lot of 0.5mm endmills I got
I paid only $2.65 each so I bought 30.

I'm making boards with really heavy copper, 12 oz or 0.42mm thick. Cutting thick copper like that requires coolant, as much to flush away the chips as anything.
In the first pass, that is 100% of diameter I cut half depth, ie 0.21mm at half normal speed 200mm/min. The second pass to depth of 0.45mm, ie the thickness of the copper
and a little bit at 100% diameter still at 200mm/min. Third and subsequent passes are with 50% step over, ie 0.25mm at full depth ie 0.45mm at full speed ie 400mm/min.
My spindle does 24000rpm and I leave it maxed out.

The wear on the endmills was a problem, I was getting only 1/4 to 1/2 hour per bit, not good enough. By using Autoleveller I reduced the amount of fiberglass in the cut stream
and doubled the endmill life. Then I started using flood coolant and that improved tool life by a factor of ten or more. I now get about 10 hours cutting per endmill.

I was that concerned about tool wear that I bought a diamond coated 0.5mm endmill from Harvey Tools at great expense. I've achieved such a dramatic improvement in tool life
by using Autoleveller and flood cooling that I haven't even tried the diamond bit....if I were to break it I'm sure you could hear me cry from here....after all you are only
8000 nautical miles away! LOL

Craig

[evalon]

@joeavaerage: Hi Craig ... If you are still here and reading this thread: Can I ask you which cooling fluid you are using? Thanks, Jesper

[joeavaerage]

Hi,
I use water and water soluable oil, nothing special. Probably just plain water would work too, its about clearing chips away from the cut zone.

Craig

[RCaffin]

Very high spin (how fast can you go?)
Slow traverse speed (learning exercise)
Vacuum table for flatness
Air blast to clear the copper chips away
Optional pulsed misting with kero/olive oil

Cheers
Roger

[A_Camera]

Hi,
I use water and water soluable oil, nothing special. Probably just plain water would work too, its about clearing chips away from the cut zone.

Craig

Dust shoe and vac is in my opinion better than blowing or flooding, but that's just my own experience and my experience is limited to standard 35um PCB.

[RCaffin]

I think that it is a bit more than just clearing the chips. Keeping the cutter 'wet' prevents BUE or the copper from sticking to the cutter, and that is definitely of the good. 'Flooding' - not needed (from experience).

Cheers

[joeavaerage]

Hi Roger,
I think you are right.

I haven't experimented extensively with cutting copper, 420um copper clad board is about the most serious example I have and that is done at comparatively low
surface speed by virtue of the small diameter endmills. I have come to regard copper as 'sticky', that is to say inclined to BUE. I understand pure nickel is worse.

I have experimented with aluminum more extensively and it too can be 'sticky'

It seems in absence of any cooling or lubrication there is a ceiling in surface speed beyond which BUE is inevitable. With 6061 I find without cooling I'm limited to
surface speeds of about 250m/min. With flood cooling I have gone as high as 500m/min without BUE.

My spindle is maxed at 24000 rpm. With a 0.5mm endmill that equates to a surface speed of only 37.7m/min in copper. Even without cooling I don't get BUE at that piddling
surface speed. Having said that if I don't clear the chips, especially in the first full width slotting pass, then they get recut and I get BUE. Even if I manage to avoid BUE
the cut quality tends to be poor with burrs etc. Cutting pure copper cleanly is not as easy as it looks. With flood coolant directed as a low pressure jet my results are improved
DRAMATICALLY. No BUE and no burrs with tenfold and better tool life.

I built my mill to cut steels and stainless so I equipped it with flood cooling, it is actually one of the cheaper of the different cooling options, ergo I tried it on my heavy copper boards
with great success.

I'm interested in your comments regarding air and misted kero/olive oil.

Whether that interest is sufficient to cause me to make the investment in time and equipment to try it out remains to be seen. At this stage I have three strategies (double sided tape
mounting, Autoleveller for perfect Z correction and flood cooling) which allow me to make very acceptable boards in 420um copper clad board. While I'm always keen to investigate
better ways of doing things given I already have a very acceptable solution there is little impetus from me to push the boundaries.

Craig

[RCaffin]

Best access is probably via "mql lubrication system'

What is MQL? (the basics) | Minimum Quantity Lubrication About MQL
The MQL Handbook: A great primer on Minimum Quantity Lubrication | Minimum Quantity Lubrication THE Handbook on MQL
Minimum Quantity Lubrication - MQL - Metal Cutting Fluids | Unist The big vendor

Now, you can spend several $K on this - and it might be justified in a commercial case too, but I am kinda stingey and after reading lots I decided to MYOG. And this was amazingly good and easy.

7724 shows the MYOG lash-up layout, with labels.
7725 shows the two lines going to the nozzle: air and liquid
7721 shows the nozzle

The solenoid valves are the smallest size readily available: basically on/off pneumatic valves. The big catch here is that the rubber used in a lot of these valves does NOT like kero. The kero is absorbed and the rubber swells up. Neoprene or Viton is needed, but they are slightly more $. I was using Festo (I had them), but ran out and am now using AirTac from China. I was assured by Australian reps that the AirTac brand was rubbish, but they didn't sell them, did they? A lot cheaper.

The pressure regulator is essential. I adjust the pressure to get the right spray of micro-droplets.

The honey bottle is the reservoir - just above the valve in height. I do not pressurise this, except when repriming the control valve. That uses the yellow fuel line, which is resistant to the kero. The lid is actually sealed (hey, commercial honey bottle!) only for the priming.

The air line is obvious. The tiny line fo the kero/olive oil mix is actually surgical tubing, but any PE tubing should do.

At the nozzle end the PE tube dives inside the LocLine and sticks out at the front by about 1 mm. The continuous air flow sucks liquid out as DROPLET spray by venturi action.

The really amazing thing is that the liquid valve way back up the line can deliver pulses of liquid down to about 0.3 sec long. I had some doubts at first, but it works.

Controls: air is on/off via M8. 'Mist' (really spray) is via M7.
BUT: when M7 is turned on all it does is to allow a special timing circuit to operate. One 555 timer controls how often the mist is really pulsed, by enabling a second 555 timer which controls the pulse length.
I have LEDs scattered around to tell me when things are happening.

This works on PCB copper, on softish Al, on 7075 Al, on steel and on titanium 6Al4V alloy.

Q?

Cheers
Roger

[joeavaerage]

Hi Roger,
so simple, I imagined some sort of pressure fed metering pump.

Is there a great deal of difference between the performance of flood cooling vs MQL?

I would have thought that the kero vapor would be fairly noxious. I have had no trouble with storing water/water soluable oil despite reports on
various forums about bacteria and bad smells etc. My mill sits in a steel tray thus all coolant drains back to the pump reservoir. I don't have a proper
filtering regime in place at the moment, the gauze filter is adequate if a bit messy.

Ultimately its the results which count. As I have previously posted BUE due to high surface speeds is not an issue, my spindle doesn't go fast enough for
that. What does give me grief is recutting chips in full width slotting paths, ie the first cut. I have found that half the copper depth, 0.21mm at 200mm/min
with flood cooling is sufficiently conservative to get good and repeatable results. Would MQL do any better do you think?

Craig

[RCaffin]

Simple? Yep, although you can add all sorts of ($$) complexity. Doubt you get anything extra for it though.

Flood cooling vs MQL? It's not that simple. Simple flood cooling does not really wash the chips away, so you can get lots of second cuts. Bad stuff.

I would separate out the two processes involved for consideration:air blast and MQL. They need to be treated separately.

Air blast gets you the chip clearance. You can also get this with high pressure flood coolant, but you end up paying quite a high cost. Yes, water (etc) has a lot more weight for pushing the chips away, but air blast can do this just as easily by using several bar of pressure applied close up. The chips are not that heavy after all.

The point of MQL is almost completely separate from that. The function here is to keep the cutting edges 'wet' so the metal cannot stick. For this you need mere drops of some sort of oil. You can use an organic oil like olive oil here just as easily, with the glorious virtue that olive oil is totally skin-compatible.

Major companies like Boeing and Ford are switching to this and they are finding that they get double the tool life. There are several reasons for this. The cutting edge slides across the metal better, which reduces abrasion and wear on the cutter. There is not the thermal shock you get when the cutter goes from carving off some high-tens steel to being hit by cold coolant, and this reduces micro-cracking on the cutter surface and coating. To be sure, that means you need to be using good carbide cutters which can take the heat - so?

Caveat: drilling deep holes without high-pressure coolant injected down the length of the drill may be more difficult. However, separate the two process again. The usual rationale for the coolant is to clear the swarf out of the hole, NOT for cooling. I find peck drilling with air blast very effective instead.

Kero vapour: I never notice it. The reason here is that the amount used is microscopic. You see the honey bottle in my photos? Half a bottle of kero/olive oil will easily last me 6 months. Parts come out 'dry'.

Can one avoid BUE with very high spindle speeds (HSM)? Yes. The physics changes a bit: the cutter is through the material so fast that BUE has trouble getting started. What may be less obvious is that with such high spindle speeds you need to reduce the chip thickness a bit, to limit the forces on the cutting edges. But you carve them off much faster. For real HSM you will need higher spindle power: not for the cutter but to overcome the air drag inside the motor and the bearing drag!

A real improvement for HSM, and for more conventional machining too, is the trend from simple grinding of the cutter flutes to honing of the cutter flutes. Not all companies offer this yet. Roughly ground cutters drag terribly, while honed cutters slide through better and can also have sharper cutting edges.

Copper is a pig, and so is soft aluminium. I have cut a lot of 5000-series Al with MQL. Some tuning is needed, but it works very well.

Cheers
Roger

[joeavaerage]

Hi All,

Cutting pure copper cleanly is not as easy as it looks

This is probably a little off topic but may go some way to explaining why that should be so.

I'm quoting from:
Materials and Processes in Manufacturing, fifth ed. by Paul DeGarmo table 17-5 page 532.
The context of this chapter of the book is the physics of chip formation of metals. It is an involved and technically challenging/interesting
subject. The table I refer to occurs near the end of the discourse and is titled:
'Specific power of metal removal in machining various metals'

I will condense the data therein somewhat for the sake of brevity. The numeric figure is the required power to generate unit volume of chips or the material concerned
under ideal conditions. As such you might call it a figure of merit how easy/hard it is to machine a given metal.
Aluminum (various grades) 0.41 hp/cubic inch
Brass (various grades) 0.43 hp/cubic inch
Copper, annealed 0.72 hp/cubic inch
Steel, mild 0.59 hp/cubic inch
Steel, high carbon 0.70 hp/cubic inch
Nickel, pure 0.89 hp/cubic inch

Note how annealed copper requires MORE power than a hard and tough steel to produce a given volume of chips, and pure nickel requires
even more. The explanation given in the text is that copper (and nickel) while being very much softer than steel are very much more ductile.
Thus to form a chip the tool has to deform the metal to the limit of its ductility and thereby break a chip free of the base metal. The energy required
to produce a chip is the product of the force required times the distance its required to act to form a chip.

So while steel requires a great deal more force to form a chip it only has to deform or strain it a small percentage before the chip breaks. Whereas
copper (and nickel) have a lower cutting force but require several times the deformation/strain before the chip separates.

This result was a surprise to me when I fist read it nearly 40 years ago. I have referred to this text on a number of occasions since as it goes
someway to explain the seeming arcane rules that have developed around the machining of metals.

In this instance where I am machining heavy copper PCB it is essentially a metal cutting process, albeit with a modicum of abrasive fiberglass in the
chip stream just to add another wrinkle/dimension to the equation! The upshot is that while copper is soft it takes a great deal of determination to
form a clean chip. This has a number of implications.

The first is that as we are feeling our way into cutting a new or unfamiliar metal we are inclined to 'take it easy'. For metals that strain harden like copper, nickel
and especially austentitc stainless (300 series) that is a mistake. 'Taking it easy' on such a metal is to risk 'giving it a good rub' instead of cutting. The surface
deformation that happens when you 'give it a rub' hardens the surface and will make subsequent passes of the tool tooth even less likely to form a chip.

My recommendation therefore is to approach these metals fairly aggressively, as aggressively as the rigidity of your machine and the strength of your tool allows.
This represents a challenge for small diameter tools as I use. I work on a basis of the chipload per tooth being 1% of the tool diameter.
Thus for my 0.5mm tool the chipload per tooth would be 0.5 X 0.01 =0.005 mm or 5um. For the two flutes the chip load per revolution is 10um. For my spindle at
24000 rpm the cut speed is 24000 x 0.01=240mm/min.

As you can see, and from my previous posts I honor this figure in abeyance, I slot at 200mm/min and with a 50% step over I cut at 400mm/min. The calculated figure
gives a ball park estimation of where you should be.

For tools of about 1.5mm diameter I allow a chipload of 2%, the tool being stronger.
For tools of about 3mm and upward I allow a chipload of between 3 and 5%, as the tool diameter goes up the rigidity of my machine and/or spindle power become
the limiting factors.

I have often quoted 'the value of a hobby is the stuff that you learn in the pursuit of it'. As you can see I have had to learn, or rather relearn, about chip formation.
Learning is character building but that is a problem....I've got no character and don't want to start now!'

Craig

[joeavaerage]

Hi Roger,

Kero vapour: I never notice it. The reason here is that the amount used is microscopic. You see the honey bottle in my photos? Half a bottle of kero/olive oil will easily last me 6 months. Parts come out 'dry'.

That's it...I think you got me....damn....I was quite happy sailing along thinking everything is great and now a whole new thing has cropped up!

In truth I'm down to the last square foot or so of my heavy copper board and I have to weigh whether I want or need more. My original purpose was to make an AC servo drive
with about 15A continuous and 48A overload. I have done so and the thick copper has made the circuit boards easy to design and build, not withstanding the learning
involved in isolation routing such PCB.

Most commercial high current boards make extensive use of electroplating and plated through holes to achieve low resistance. With this heavy copper board I can have a trace
only 2.3mm wide having a conductive area of 1mm square. I assure you it is very seductive to be able to make traces that can handle high currents in such narrow widths.
As a result of that capability I have found lots of other uses for it. A PWM wiredrive PCB I made for a customer at work was only 5A output but that can be handled by traces
1mm wide, you can carry current in traces that you would only call signal traces otherwise.

Craig