Topics

Can only get a single isolation pass around tracks #eagle #pcbgcode #etch

AlbertHall
 

I have ground planes on both sides of the board. Rtch tool size set to o.3mm.  As you can see below there are multiple isolation passes around the drilled hole but only a single pass around the tracks (except where they are close to each other).

How do I get multiple passes around the tracks, please?


John Johnson
 

Since your tool size is 0.3, and step size is 0.15, then 0.3+0.15=0.45, which is over the maximum. You could try upping the max to, say, 0.5 and see how that works.

Regards,
John

On Aug 19, 2018, 8:43 AM -0400, AlbertHall <jim@...>, wrote:
I have ground planes on both sides of the board. Rtch tool size set to o.3mm.  As you can see below there are multiple isolation passes around the drilled hole but only a single pass around the tracks (except where they are close to each other).

How do I get multiple passes around the tracks, please?


AlbertHall
 

That adds an extra pass around the drilled holes but the tracks still only have a single pass.

AlbertHall
 

I have attached the board file if anyone would like to have a go at this.

peterg1000
 

Took your board and erased all your tracks and polygons, then set all the tracks to 0.4mm in "net classes" and all clearances to 0.4mm in "design rules". Ran the autorouter and chose the minimum "vias" solution.  Then ran pcbgcode with my own default settings and a 0.2mm 30degree engraving tool.

Examined the results using UCCNC toolpath display, and took the attached screenshot for your info. Note that there are some nets not routed, but you get the flavour of the results.

Peter

AlbertHall
 

So I deleted the polygons on both sides and then added them back and connected them to ground again and now I get multiple passes around the tracks.
Aaargh!
I hate software.

But which software do I hate - eagle or pcb-gcode?

sdmonaco
 

Hi Albert,

You might try my single pass etch approach, which I now use for all of my boards.  I setup pcb-gcode to do a singe pass etch for all traces and shapes.  After completion, I inspect the board without removing it from the fixture.  If I see any areas with incomplete etching, I lower my z-axis zero point by 2 mils and re-etch the board with the same gcode file.  I continue with this until I have a nicely etched PCB which usually only takes about 1 to 3 passes.  This approach also corrects most leveling issues as well.  I try to keep my trace widths over 12 mils because each pass will shrink your trace width by about 1 mill for a 30 degree spade bit.

Sam

mlmcnc
 

Use "Autoleveller" it will save you a lot of time and give you consistent track thicknesses.

Rick Obel
 

My approach is very similar to Sam’s, but three differences that I believe help:

1.       Immediately before cutting a PCB, I pre-level my sacrifice board, which is MDF.  I believe that the MDF warps over time, even though I rarely remove it, or even adjust it.

2.       I don’t use angled bits (sorry I don’t know the correct terminology).  Instead, I find that conventional end mills produce a cleaner, more consistent cut.  I am still experimenting with diameters, but find 0.6mm breaks easily, while 0.9 is unnecessarily wide.  Lately, I have been using 0.7mm but will focus on 0.8 next.

3.       If the incomplete etching (is that the right word?) is localized, I will re-run just those areas at a deeper depth.

Rarely do I need three passes, and a second pass is only required about half the time.

-Rick

 

From: pcbgcode@groups.io [mailto:pcbgcode@groups.io] On Behalf Of sdmonaco via Groups.Io
Sent: Monday, August 20, 2018 7:56 AM
To: pcbgcode@groups.io
Subject: Re: [pcbgcode] Can only get a single isolation pass around tracks

 

Hi Albert,

 

You might try my single pass etch approach, which I now use for all of my boards.  I setup pcb-gcode to do a singe pass etch for all traces and shapes.  After completion, I inspect the board without removing it from the fixture.  If I see any areas with incomplete etching, I lower my z-axis zero point by 2 mils and re-etch the board with the same gcode file.  I continue with this until I have a nicely etched PCB which usually only takes about 1 to 3 passes.  This approach also corrects most leveling issues as well.  I try to keep my trace widths over 12 mils because each pass will shrink your trace width by about 1 mill for a 30 degree spade bit.

 

Sam

peterg1000
 

My approach is to use a small vacuum table with a sacrificial MDF top, leveled in-situ. No need to use an autoleveler - the vacuum ensures the FR4 is absolutely flat and tight to the surface and has no spring whatsoever.

Can consistently mill accurate 0.3mm tracks and isolation if required using a 30degree tool with a 0.2 mm spade tip. Cutting depth is set in pcbgcode to 0.2mm.

Peter

joaobatista
 

can you post image of your vacuo system?

mlmcnc
 

A vacuum table with a leveled top is indeed a good approach. However, that only ensures that the BOTTOM of the board is level. In my experience, not all FR4 is consistent in thickness. Remember that 1oz copper is only 0.0347mm thick. Autolevelling is the only way to track the top surface. If you do this you could use a smaller depth of cut with correspondingly narrower kerf with your V bit.

Bernardc
 

Assume that you not use SMD on your PCB as etching with 0.9 mm….

Autoleveler is perfect to correct the copper default. I use 2 flutes V type 60° dia 0.127 mm with 0.05 mm depth and it is not necessary to make further pass.  Mini = 0.0254, Max = 0.762, Step = 0.0508, Feedrate = 400, Turn = 12000 T/mn. I can etch for SSOP. (The problem remains welding ;-)..)

 

Bernard  

 

De : pcbgcode@groups.io [mailto:pcbgcode@groups.io] De la part de Rick Obel
Envoyé : lundi 20 août 2018 17:32
À : pcbgcode@groups.io
Objet : Re: [pcbgcode] Can only get a single isolation pass around tracks

 

My approach is very similar to Sam’s, but three differences that I believe help:

1.       Immediately before cutting a PCB, I pre-level my sacrifice board, which is MDF.  I believe that the MDF warps over time, even though I rarely remove it, or even adjust it.

2.       I don’t use angled bits (sorry I don’t know the correct terminology).  Instead, I find that conventional end mills produce a cleaner, more consistent cut.  I am still experimenting with diameters, but find 0.6mm breaks easily, while 0.9 is unnecessarily wide.  Lately, I have been using 0.7mm but will focus on 0.8 next.

3.       If the incomplete etching (is that the right word?) is localized, I will re-run just those areas at a deeper depth.

Rarely do I need three passes, and a second pass is only required about half the time.

-Rick

 

From: pcbgcode@groups.io [mailto:pcbgcode@groups.io] On Behalf Of sdmonaco via Groups.Io
Sent: Monday, August 20, 2018 7:56 AM
To: pcbgcode@groups.io
Subject: Re: [pcbgcode] Can only get a single isolation pass around tracks

 

Hi Albert,

 

You might try my single pass etch approach, which I now use for all of my boards.  I setup pcb-gcode to do a singe pass etch for all traces and shapes.  After completion, I inspect the board without removing it from the fixture.  If I see any areas with incomplete etching, I lower my z-axis zero point by 2 mils and re-etch the board with the same gcode file.  I continue with this until I have a nicely etched PCB which usually only takes about 1 to 3 passes.  This approach also corrects most leveling issues as well.  I try to keep my trace widths over 12 mils because each pass will shrink your trace width by about 1 mill for a 30 degree spade bit.

 

Sam

Rick Obel
 

I tried autoleveler and concluded (maybe wrongly) that the PCB board high spots were the result of very minor warpage, not inconsistencies in the materials’ thickness.  As a result, I reasoned, when the autoleveler measures the height, it does so without the benefit of any pressure (mere electrical contact triggers height measurement), but when cutting, a little pressure will force the PCB’s high spots down, with the result that they are not fully cut in those areas.  To remedy that problem, I bought a thick piece of Plexiglas, a vacuum pump and other materials to make my own vacuum board, but then realized that I could get good results, cutting not more than 0.10 mm in two passes of 0.05mm each, by simply leveling my MDF in advance of a days’ cutting.  (I don’t know if it moisture or thermal expansion/contraction, or otherwise, but I find that more than after a few days, the MDF takes on a warp of up to 0.10mm. )

 

I would be very interested to see others’ vacuum boards.  I saw Peter’s earlier comment about success with his vacuum board (table) and his good results cutting 0.2mm  -- he must have meant 0.20 mm, right?  Because 0.2mm is 4 passes at 0.05, and I typically need only 1 pass, and rarely more than two.  Am I missing something here?  The boards I am cutting are usually 4”x6”, and are much more problematic than 2”x3”.

 

 

From: pcbgcode@groups.io [mailto:pcbgcode@groups.io] On Behalf Of Bernardc
Sent: Monday, August 20, 2018 10:41 AM
To: pcbgcode@groups.io
Subject: Re: [pcbgcode] Can only get a single isolation pass around tracks

 

Assume that you not use SMD on your PCB as etching with 0.9 mm….

Autoleveler is perfect to correct the copper default. I use 2 flutes V type 60° dia 0.127 mm with 0.05 mm depth and it is not necessary to make further pass.  Mini = 0.0254, Max = 0.762, Step = 0.0508, Feedrate = 400, Turn = 12000 T/mn. I can etch for SSOP. (The problem remains welding ;-)..)

 

Bernard  

 

De : pcbgcode@groups.io [mailto:pcbgcode@groups.io] De la part de Rick Obel
Envoyé : lundi 20 août 2018 17:32
À : pcbgcode@groups.io
Objet : Re: [pcbgcode] Can only get a single isolation pass around tracks

 

My approach is very similar to Sam’s, but three differences that I believe help:

1.       Immediately before cutting a PCB, I pre-level my sacrifice board, which is MDF.  I believe that the MDF warps over time, even though I rarely remove it, or even adjust it.

2.       I don’t use angled bits (sorry I don’t know the correct terminology).  Instead, I find that conventional end mills produce a cleaner, more consistent cut.  I am still experimenting with diameters, but find 0.6mm breaks easily, while 0.9 is unnecessarily wide.  Lately, I have been using 0.7mm but will focus on 0.8 next.

3.       If the incomplete etching (is that the right word?) is localized, I will re-run just those areas at a deeper depth.

Rarely do I need three passes, and a second pass is only required about half the time.

-Rick

 

From: pcbgcode@groups.io [mailto:pcbgcode@groups.io] On Behalf Of sdmonaco via Groups.Io
Sent: Monday, August 20, 2018 7:56 AM
To: pcbgcode@groups.io
Subject: Re: [pcbgcode] Can only get a single isolation pass around tracks

 

Hi Albert,

 

You might try my single pass etch approach, which I now use for all of my boards.  I setup pcb-gcode to do a singe pass etch for all traces and shapes.  After completion, I inspect the board without removing it from the fixture.  If I see any areas with incomplete etching, I lower my z-axis zero point by 2 mils and re-etch the board with the same gcode file.  I continue with this until I have a nicely etched PCB which usually only takes about 1 to 3 passes.  This approach also corrects most leveling issues as well.  I try to keep my trace widths over 12 mils because each pass will shrink your trace width by about 1 mill for a 30 degree spade bit.

 

Sam

Rick Obel
 

I do not use smd. 

 

From: Rick Obel
Sent: Monday, August 20, 2018 11:05 AM
To: pcbgcode@groups.io
Subject: RE: [pcbgcode] Can only get a single isolation pass around tracks

 

I tried autoleveler and concluded (maybe wrongly) that the PCB board high spots were the result of very minor warpage, not inconsistencies in the materials’ thickness.  As a result, I reasoned, when the autoleveler measures the height, it does so without the benefit of any pressure (mere electrical contact triggers height measurement), but when cutting, a little pressure will force the PCB’s high spots down, with the result that they are not fully cut in those areas.  To remedy that problem, I bought a thick piece of Plexiglas, a vacuum pump and other materials to make my own vacuum board, but then realized that I could get good results, cutting not more than 0.10 mm in two passes of 0.05mm each, by simply leveling my MDF in advance of a days’ cutting.  (I don’t know if it moisture or thermal expansion/contraction, or otherwise, but I find that more than after a few days, the MDF takes on a warp of up to 0.10mm. )

 

I would be very interested to see others’ vacuum boards.  I saw Peter’s earlier comment about success with his vacuum board (table) and his good results cutting 0.2mm  -- he must have meant 0.20 mm, right?  Because 0.2mm is 4 passes at 0.05, and I typically need only 1 pass, and rarely more than two.  Am I missing something here?  The boards I am cutting are usually 4”x6”, and are much more problematic than 2”x3”.

 

 

From: pcbgcode@groups.io [mailto:pcbgcode@groups.io] On Behalf Of Bernardc
Sent: Monday, August 20, 2018 10:41 AM
To: pcbgcode@groups.io
Subject: Re: [pcbgcode] Can only get a single isolation pass around tracks

 

Assume that you not use SMD on your PCB as etching with 0.9 mm….

Autoleveler is perfect to correct the copper default. I use 2 flutes V type 60° dia 0.127 mm with 0.05 mm depth and it is not necessary to make further pass.  Mini = 0.0254, Max = 0.762, Step = 0.0508, Feedrate = 400, Turn = 12000 T/mn. I can etch for SSOP. (The problem remains welding ;-)..)

 

Bernard  

 

De : pcbgcode@groups.io [mailto:pcbgcode@groups.io] De la part de Rick Obel
Envoyé : lundi 20 août 2018 17:32
À : pcbgcode@groups.io
Objet : Re: [pcbgcode] Can only get a single isolation pass around tracks

 

My approach is very similar to Sam’s, but three differences that I believe help:

1.       Immediately before cutting a PCB, I pre-level my sacrifice board, which is MDF.  I believe that the MDF warps over time, even though I rarely remove it, or even adjust it.

2.       I don’t use angled bits (sorry I don’t know the correct terminology).  Instead, I find that conventional end mills produce a cleaner, more consistent cut.  I am still experimenting with diameters, but find 0.6mm breaks easily, while 0.9 is unnecessarily wide.  Lately, I have been using 0.7mm but will focus on 0.8 next.

3.       If the incomplete etching (is that the right word?) is localized, I will re-run just those areas at a deeper depth.

Rarely do I need three passes, and a second pass is only required about half the time.

-Rick

 

From: pcbgcode@groups.io [mailto:pcbgcode@groups.io] On Behalf Of sdmonaco via Groups.Io
Sent: Monday, August 20, 2018 7:56 AM
To: pcbgcode@groups.io
Subject: Re: [pcbgcode] Can only get a single isolation pass around tracks

 

Hi Albert,

 

You might try my single pass etch approach, which I now use for all of my boards.  I setup pcb-gcode to do a singe pass etch for all traces and shapes.  After completion, I inspect the board without removing it from the fixture.  If I see any areas with incomplete etching, I lower my z-axis zero point by 2 mils and re-etch the board with the same gcode file.  I continue with this until I have a nicely etched PCB which usually only takes about 1 to 3 passes.  This approach also corrects most leveling issues as well.  I try to keep my trace widths over 12 mils because each pass will shrink your trace width by about 1 mill for a 30 degree spade bit.

 

Sam

Rick Obel
 

Sorry for the multiple responses.  I appreciate your suggestions and will give your approach a try.

 

From: Rick Obel
Sent: Monday, August 20, 2018 11:06 AM
To: 'pcbgcode@groups.io'
Subject: RE: [pcbgcode] Can only get a single isolation pass around tracks

 

I do not use smd. 

 

From: Rick Obel
Sent: Monday, August 20, 2018 11:05 AM
To: pcbgcode@groups.io
Subject: RE: [pcbgcode] Can only get a single isolation pass around tracks

 

I tried autoleveler and concluded (maybe wrongly) that the PCB board high spots were the result of very minor warpage, not inconsistencies in the materials’ thickness.  As a result, I reasoned, when the autoleveler measures the height, it does so without the benefit of any pressure (mere electrical contact triggers height measurement), but when cutting, a little pressure will force the PCB’s high spots down, with the result that they are not fully cut in those areas.  To remedy that problem, I bought a thick piece of Plexiglas, a vacuum pump and other materials to make my own vacuum board, but then realized that I could get good results, cutting not more than 0.10 mm in two passes of 0.05mm each, by simply leveling my MDF in advance of a days’ cutting.  (I don’t know if it moisture or thermal expansion/contraction, or otherwise, but I find that more than after a few days, the MDF takes on a warp of up to 0.10mm. )

 

I would be very interested to see others’ vacuum boards.  I saw Peter’s earlier comment about success with his vacuum board (table) and his good results cutting 0.2mm  -- he must have meant 0.20 mm, right?  Because 0.2mm is 4 passes at 0.05, and I typically need only 1 pass, and rarely more than two.  Am I missing something here?  The boards I am cutting are usually 4”x6”, and are much more problematic than 2”x3”.

 

 

From: pcbgcode@groups.io [mailto:pcbgcode@groups.io] On Behalf Of Bernardc
Sent: Monday, August 20, 2018 10:41 AM
To: pcbgcode@groups.io
Subject: Re: [pcbgcode] Can only get a single isolation pass around tracks

 

Assume that you not use SMD on your PCB as etching with 0.9 mm….

Autoleveler is perfect to correct the copper default. I use 2 flutes V type 60° dia 0.127 mm with 0.05 mm depth and it is not necessary to make further pass.  Mini = 0.0254, Max = 0.762, Step = 0.0508, Feedrate = 400, Turn = 12000 T/mn. I can etch for SSOP. (The problem remains welding ;-)..)

 

Bernard  

 

De : pcbgcode@groups.io [mailto:pcbgcode@groups.io] De la part de Rick Obel
Envoyé : lundi 20 août 2018 17:32
À : pcbgcode@groups.io
Objet : Re: [pcbgcode] Can only get a single isolation pass around tracks

 

My approach is very similar to Sam’s, but three differences that I believe help:

1.       Immediately before cutting a PCB, I pre-level my sacrifice board, which is MDF.  I believe that the MDF warps over time, even though I rarely remove it, or even adjust it.

2.       I don’t use angled bits (sorry I don’t know the correct terminology).  Instead, I find that conventional end mills produce a cleaner, more consistent cut.  I am still experimenting with diameters, but find 0.6mm breaks easily, while 0.9 is unnecessarily wide.  Lately, I have been using 0.7mm but will focus on 0.8 next.

3.       If the incomplete etching (is that the right word?) is localized, I will re-run just those areas at a deeper depth.

Rarely do I need three passes, and a second pass is only required about half the time.

-Rick

 

From: pcbgcode@groups.io [mailto:pcbgcode@groups.io] On Behalf Of sdmonaco via Groups.Io
Sent: Monday, August 20, 2018 7:56 AM
To: pcbgcode@groups.io
Subject: Re: [pcbgcode] Can only get a single isolation pass around tracks

 

Hi Albert,

 

You might try my single pass etch approach, which I now use for all of my boards.  I setup pcb-gcode to do a singe pass etch for all traces and shapes.  After completion, I inspect the board without removing it from the fixture.  If I see any areas with incomplete etching, I lower my z-axis zero point by 2 mils and re-etch the board with the same gcode file.  I continue with this until I have a nicely etched PCB which usually only takes about 1 to 3 passes.  This approach also corrects most leveling issues as well.  I try to keep my trace widths over 12 mils because each pass will shrink your trace width by about 1 mill for a 30 degree spade bit.

 

Sam

peterg1000
 

No problems milling SMD boards so far (picture attached).  Just very tedious to place components by hand and get the right amount of solder.  I use solder past thinned with liquid flux - this is applied using a fine wire pickup loop.  If "vias" are needed I use 1/32" x 1/4" brass rivets, pre - cleaned before inserting in the board.  Solder rivet head side first, then crop to nearly flush and solder this side second.

All tools are zeroed with calibrated touch probe on copper - this is accurate to better than 25 microns (0.001"). Cutting depth fixed at -0.2mm.  Spindle speed set at 18000rpm on 30degree engraving cutter with 0.2mm (.008") wide cutting tip.

Vacuum fixture is milled from hardwood and uses a vane type vacuum pump to pull about - .75bar in normal operation.  More pics attached for interest.

Peter