PDA

View Full Version : New build – input/comments appreciated



ngundtoft
08-09-2015, 06:18 AM
Hi CNC Experts

For some time now I have planned to build a 3-axis CNC router. I will primarily use it for wood working, mainly loudspeaker building, but also for general wood working including dovetail joints. However, I would also like to use it for work in aluminium, mainly for making front and rear panels for electronic cabinets and for engraving them. I could also imagine using it as a 3D printer.
I have looked at various home pages and forums for inspiration, but MYCNCUK seems to be most in line with my own ideas. So after having learned from the projects on these pages, especially the machines described by EddyCurrent and Kingcreaky, also comments from Jazz have been very useful, I have made some drawings and a part list, and now I would really appreciate any comments on this, before I start buying parts and putting them together.

Work area: X-axis: 733mm, Y-axis: 580mm, Z-axis: 220mm

Frame made of 50 x 50mm and 50 x 100 (3mm) steel profiles. I intend to use the epoxy self-levelling support technique (30 – 40mm wide channels, 5mm thick).
Gantry made of aluminium profile.

Mechanical parts:



Ballscrews (C7): RM1610: X-axis: 1015mm x 2, Y-axis: 800mm x 1, RM1605: Z-axis: 340mm x 1
Ballscrew supports: BK12 (fixed end) x 4, BF12 (loose end) x 4
Ballnut housing: 1610 x 3, 1605 x 1
Linear rails: Hiwin profile 20mm, HGH-CA (HA, CW)), standard preload Z0. X-axis: 1000mm x 2, Y-axis: 750mm x 2, Z-axis: 416mm x 2
Pulleys: 20T, 21,8mm Dia., 8mm hole x 8
Belts: 15mm HTD, 1:1, (20T : 20T).


Electrical parts:



Stepper motors: Nema 23 - SY60STH86-300BBF, x 4 (parallel connection)
Stepper drivers; Leadshine EM806, EM860 or AM882 x 4
PSU Linear, 70V, 9A. 500W toroidal transformer (rectifier, capacitors, softstart)
Controller Smoothstepper, or Pokeys57E
BOB Machschmidt BOB v1.5, OPPBV3 (DIY CNC) both require separate spindle control board (e.g. SP5 V5 Spindle Board from DIY CNC). Alternatively Chinese board with built in spindle control: http://www.ebay.com/itm/6-Axis-CNC-M...item1c3dfdfbde (http://www.ebay.com/itm/6-Axis-CNC-MACH3-Engraving-Machine-Interface-Breakout-Board-USB-PWM-Spindle-/121299139550?hash=item1c3dfdfbde)
Spindle motor: 2.2KW with VFD and water cooling.



Software;



Control SW: Mach3 or 4.


http://www.mycncuk.com/attachment.php?attachmentid=16074&stc=1http://www.mycncuk.com/attachment.php?attachmentid=16075&stc=1http://www.mycncuk.com/attachment.php?attachmentid=16076&stc=1http://www.mycncuk.com/attachment.php?attachmentid=16077&stc=1http://www.mycncuk.com/attachment.php?attachmentid=16078&stc=1http://www.mycncuk.com/attachment.php?attachmentid=16079&stc=1http://www.mycncuk.com/attachment.php?attachmentid=16080&stc=1

Lee Roberts
08-09-2015, 07:39 PM
Nils,

Wow what a start! I think this introduction has first place position as the best example of a new comer, doing their much needed research first and then presenting to us a damn fine example of what is here on the forum learn from, be inspired from and then made into a reality.

I think if you build this machine to what is suggested in the CAD models and it finishes mechanically sound with an Electrical setup to the same standard, you'll finish with a flawless very capable machine.

The ONLY thing I think you should consider and do, is to move the Y Axis (Longest Axis, the bed) motors to outside the machine frame, I think their current location could give you a risk you don't need to take, first they could get in the way of the Z Axis accidently causing unnecessary damage, they may also be in your way if you decide to start using a coolant system as they could get wet and soiled and for woodworking, they could end up covered in dust, motors get very hot, internals could become clogged up/dirty.

Other than that I think you have a winner - GET CRACKING !

Lee

routercnc
08-09-2015, 08:00 PM
Agree with Lee - nice design. Will do just want you want it to !

Some other minor things to think about:
Z travel is quite long - I guess you need that much to do the speaker cabinets if they are partially assembled? If you want to cut aluminium then you will probably need to raise the job towards the cutter to limit the reach of the tool and keep the stiffness up. Some thick packer plates or wooden blocks should do it but you can add those when you are ready.

At the bottom of the frame I'm not sure I'd do the diamond in the base legs, may be just a diagonal across from one corner to the other, and I'd probably add adjustable feet to level it.

That's it really . . .

Blackrat
09-09-2015, 05:45 PM
looks good

i think you might struggle with alignment with the y rails running top and bottom , maybe placing on the front face will be easier .... my first machine was like that ... just a thought :D

JAZZCNC
09-09-2015, 05:55 PM
100% agree with Lee (No I'm not sucking up so you lot can shut up before you say it.!! :hopelessness:) Let this be a lesson to others on how to go about starting a CNC build.! . . . Well done Sir A-Star +.!

I do have some suggestions for you thou. Other than turning the motors outward which is a must do.!

Put some ends on the gantry it will greatly stiffen it up. I would also think about lowering the ballscrews has they are vunerable up there and some form of protection from chips etc.
This can be achieved by making the gantry narrower so it's just extends past the sides of the machine. The gantry ends then extend down past the top rail 3-5mm away from the steel and on the inside of the ballscrew. The gantry sides can then be bolt to the gantry and the bearing plates making gantry slightly narrower and a very strong structure.
This then allows you to place the ballscrews lower down out the firing line of chips protected by the top rail and damage from dropped materail etc.
The gantry side being on the inside of the screw makes alignment and setup of the ballscrew much easier. More importantly it narrows the restriction from the moving part the gantry and allows you to cover the screws easier with covers for protection. If your not sure what I mean ask and I can provide examples.

Also think about extending the top rails past the ends of the machine.? If done at both ends this can gain you back some cutting area and full access to work area by placing gantry in space using a short cantilever and at other end allow gantry to pass the front of machine so materail to be clamped on the end and edge machined or over height material.
I suggested this to Eddy at time he was designing and he didn't take up on it to only later end up cutting lumps out of his bed so could edge machine material ie: dove tails etc.

Only other change to frame I would make is to remove the back plate on the axis which moves acreoss the gantry. It's not required, it adds weight and very little strength. Can make the building much harder has the mating edges need to be perfectly 90deg otherwise it can twist and bend the bearing plates. Profiled linear bearings don't tolerate this very well and will bind.
Nice strong Drop down bracket will work just has well be much lighter, cheaper and easier to build.

Electrics then the only thing I would change is the BOB. For get the cheap chinese stuff you'll only regret it. I wouldn't use DIYCNC stuff either if your going with Smooth stepper. Honestly there really is only one BOB worth using with the ESS and that is PMDX 126.
All the others will at some time give you trouble or restrict the machine performance. Pay the extra at the begining and you won't regret I promise you that, which can't be said for the others. Trust me on this I've used most if not all of them.

What I've suggested are just that suggestions and what you have drawn so far would work great but some of those things I've suggested I think you'd find would probably be on your bucket list for Mk2.!!

Well done and if you build with the same careful attention then you'll have nice machine.

njhussey
09-09-2015, 06:00 PM
Wow, well researched and thought out!!!

I'd definately do what Jazz suggests and extend the rails, I wish I'd done it now (mk2 will have them like that!!) not only to cut dovetails but to put a 4th axis etc. I can vouch for putting sides on the gantry and extending them down below the rails, makes aligning the ballscrews easy (make your frame as square as possible, will make aligning easier...less shimming!!)

Looking forward to seeing this build...

ngundtoft
09-09-2015, 06:00 PM
blackrat - sorry for mixing up the X and Y axes. I expect that you mean the X axis according to my new drawing? Maybe it is better, but it will also reduce the effective Y-axis work length. Anyone to comment on this?

JAZZCNC - with respect to extending the top rails (Y-axis), I am not quite sure what you mean by "placing gantry in space using a short cantilever"?
With respect to BOB - I have noticed that you are not happy with cheap Chinese boards :pride: - and very understandable so, if you have had bad experience with them. My reasoning was that the signal "processing" (Smoothstepper) is the critical part and that the BOB is really a fairly simple device which is not so likely to fail or give problems. A concern with the PMDX 126 (which is undoubtedly very good quality) is that it does not have controller for spindle and pump.

JAZZCNC
09-09-2015, 06:22 PM
blackrat - sorry for mixing up the X and Y axes. I expect that you mean the X axis according to my new drawing? Maybe it is better, but it will also reduce the effective Y-axis work length. Anyone to comment on this?
There is no right or wrong it's how you view the machine. From the front it makes more sense to call long axis Y.


JAZZCNC - with respect to extending the top rails (Y-axis), I am not quite sure what you mean by "placing gantry in space using a short cantilever"?
With respect to BOB - I have noticed that you are not happy with cheap Chinese boards :pride: - and very understandable so, if you have had bad experience with them. My reasoning was that the signal "processing" (Smoothstepper) is the critical part and that the BOB is really a fairly simple device which is not so likely to fail or give problems. A concern with the PMDX 126 (which is undoubtedly very good quality) is that it does not have controller for spindle and pump.
Argh but your assumption that the BOB is fairly simple device is mistake often made. BOB is actually the single largest cause of trouble and strange happenings with CNC machines.! Cutting corners here is recipe for trouble believe me.
The thing you have to realise is that the BOB handles all the signals and if it's low quality with slow opto's or poor circuitry then those signals get strangled, distrorted and weakened. There's no point having an external motion card which gives nice fast crisp pulses if the BOB is going to bottleneck them and distort the signals.
Like anything in life the CNC machine is only good has the weakest link and cheap BOB will be just that then some.!!

The PMDX 126 does have Spindle speed controller daughter board pmdx 107 with Relays for controling pumps etc.

With regards to Cantilever overhang then like this.
Attachment 16024 Attachment 16023

ngundtoft
10-09-2015, 07:06 PM
Can you please tell me what these profiles are called and where you can buy them. It looks like they are approx. 100mm high in the first of your pictures. I have searched the net, but I have not found the same profiles.


One more question - what distance approximately should you aim for between the tip of the spindle and the bed (steel frame), in the down position?

http://www.mycncuk.com/attachment.php?attachmentid=16081&stc=1

JAZZCNC
10-09-2015, 07:06 PM
Can you please tell me what these profiles are called and where you can buy them. It looks like they are approx. 100mm high in the first of your pictures. I have searched the net, but I have not found the same profiles.

In the first picture It's an equivalent to ITEM profile size 120x80 with 8mm slot. Called Profile 40 I-Type Slot8 from Motedis in Germany.
http://www.motedis.co.uk/shop/Slot-profiles:::999991.html?XTCsid=m34gaqdl9mvr60bnk0tf jmi0i0
You will need spacers with this profile for the BK/BF12 blocks as the slot hole spacing doesn't match.

The smaller narrower machines use 45x90 Profile B slot 10 and they do match the BK/BF12 blocks.

Other option is don't use BK/BF blocks and machine the bearings directly into the gantry sides and buy slightly longer screws.


One more question - what distance approximately should you aim for between the tip of the spindle and the bed (steel frame), in the down position?

Hard to answer because it depends on several factors like longest tool length and Max material size your likely to want to cut. Best way is to make a call on the max material thickness and the longest tool your likely to need and Add together plus 10-20mm extra.

For the amount of Z axis travel then I wouldn't go past 200mm with 150-160 giving a good balance of range and strength. The longer it has to extend the more it flexs so keeping short will be stronger but could limit you at some point with longer tools.
Too long and your weakening the machine. No point having 150-200mm travel when your only ever using tools that extend from spindle 50mm and cutting material less than 50mm. All your doing in this case is introducing risk of higher resonance which affects surface finish and accurecy.

This is why people who want a wider material cutting range opt for the adjustable bed because they can keep the strength and still have option of cutting thicker material with longer tools. The machine in Post#9 has this feature so can accept 150mm thick material and cut thru it with 150mm length for making mortice and tenions in 6" Oak beams or V-carve/relief the face of 310mm thick beams.

In normal day to day use the bed mostly stays in one middle of the road position.

kingcreaky
10-09-2015, 10:19 PM
Well done sir.. Salute.

only thing I can suggest, is, instead of driving each shoulder of the gantry with its own stepper motor, connect the two ballscrews with a long belt to one stepper. you wont regret it. That way you can calibrate it square once, and it remains mechanically fixed that way going forward.

I think a few of the boys have done it, most recently mr njhussey look at his thread

matt

ngundtoft
11-09-2015, 02:00 PM
I don't know if it is necessary, but I think the cover plate over the X-axis ballnut holder will stiffen the holder in the x direction.
http://www.mycncuk.com/attachment.php?attachmentid=16084&stc=1http://www.mycncuk.com/attachment.php?attachmentid=16085&stc=1
Kingcreakys suggestion to use only a single Y-axis motor and use a belt instead, sounds very interesting. Obviously you save a motor and a driver and you avoid any misalignment between the two gantry sides. This is a fairly heavy machine.
Do you think that one Nema 23 motor is sufficient to do the job?


Obviously the top cantilevers at the rear of the machine, after having extended them, will obstruct a belt. Extending them gave me an extra 153mm, so that the spindle could Work 70mm in front of the frame, which is a fairly small work area for dovetails. So I was thinking if I only extended the front end rails I could fit a belt and I would have 153mm length for the dovetails, which I think would be plenty.
http://www.mycncuk.com/attachment.php?attachmentid=16083&stc=1

JAZZCNC - I noticed on your Pictures, that you are using a lot of proximity sensors. I was planning on using micro-switches, but maybe the proximity sensors are more reliable? Do you prefer magnetic, capacitive or inductance types?

JAZZCNC
11-09-2015, 02:54 PM
I think a few of the boys have done it, most recently mr njhussey look at his thread
What you mean a few have done it I practicly invented it, all them other buggers just copied me. . :hysterical:

But he's right it does work well.! . . . . Not pretty but affective.!



I don't know if it is necessary, but I think the cover plate over the X-axis ballnut holder will stiffen the holder in the x direction.

Yes it's very necessary and I'd suggest something a little stiffer than that thin plate. Look a pic #2 and you can see the kind of stiffness i'm talking about.




Kingcreakys suggestion to use only a single Y-axis motor and use a belt instead, sounds very interesting. Obviously you save a motor and a driver and you avoid any misalignment between the two gantry sides. This is a fairly heavy machine.
Do you think that one Nema 23 motor is sufficient to do the job?

No for gantry this weight you'll really need nema 34 around 6nm with high voltage drives. The machine in Pic 2 uses twin screws joined with belts running a nema 23 on 70Vdc and while it very works well this is about the size limit of machine I'd fit it on.




Obviously the top cantilevers at the rear of the machine, after having extended them, will obstruct a belt. Extending them gave me an extra 153mm, so that the spindle could Work 70mm in front of the frame, which is a fairly small work area for dovetails. So I was thinking if I only extended the front end rails I could fit a belt and I would have 153mm length for the dovetails, which I think would be plenty.

Personally I would have a little at both ends. Think about the future or upgrades like 4th axis.? Passing the front with enough length will allow a very easy and neat 4th axis to built into the machine without impacting the cutting area.
The extra cost of longer rails or ballscrews won't be that much and usabilty of the machine is much higher. The extra functionality will return the cost 2 fold when you come to sell it or pay for it's self within a few jobs.




JAZZCNC - I noticed on your Pictures, that you are using a lot of proximity sensors. I was planning on using micro-switches, but maybe the proximity sensors are more reliable? Do you prefer magnetic, capacitive or inductance types?

Prox switches I use are inductive and they are much much better than low/mid end mechanical switches. Cheap as chips and very accurate.
I posted a video showing how repeatable they are.! . . Excuse my horrible voice!!


https://www.youtube.com/watch?v=A11Zvi3nAFE

njhussey
11-09-2015, 10:42 PM
What you mean a few have done it I practicly invented it, all them other buggers just copied me. . :hysterical:

But he's right it does work well.! . . . . Not pretty but affective.!

Unashamed copy in my case and proud of it.....I can vouch that it's not pretty but it will do until the machine cuts it's own replacement part [emoji3]

ngundtoft
11-09-2015, 11:50 PM
I have stiffened up the X-axis ballscrew nut mounting by increasing material of bracket to 30mm, increasing cover plate to 10mm and inserting a 20mm triangular support.

http://www.mycncuk.com/attachment.php?attachmentid=16086&stc=1
I feel somewhat uncertain with respect to using a single motor and belt setup for the X-axis. I remember reading that the Nema 34 motor should normally be avoided because acceleration is much slower than the 23. So do the advantages (better alignment, single motor) make up for the drawbacks (lower acceleration, long complicated belt)?

Maybe a long belt could just go below the top cantilevers (and the bed) by using some extra idler pulleys?
http://www.mycncuk.com/attachment.php?attachmentid=16087&stc=1

njhussey
12-09-2015, 12:03 AM
I have stiffened up the X-axis ballscrew nut mounting by increasing material of bracket to 30mm, increasing cover plate to 10mm and inserting a 20mm triangular support.

Image: http://www.mycncuk.com/attachment.php?attachmentid=16043&stc=1

I feel somewhat uncertain with respect to using a single motor and belt setup for the X-axis. I remember reading that the Nema 34 motor should normally be avoided because acceleration is much slower than the 23. So do the advantages (better alignment, single motor) make up for the drawbacks (lower acceleration, long complicated belt)?

Maybe a long belt could just go below the top cantilevers (and the bed) by using some extra pulleys?

If you look in my build log you'll see me jogging my long X axis (your Y) at 10m/min. I had it up at 15m/min and it was fine so I've n9 qualms about using the 6Nm Nema 34 with the 230V driver, it's plenty quick enough. My belt is 2525mm long and again I've no problem with it (so far!!!)

JAZZCNC
12-09-2015, 03:45 PM
Neil I don't mind in the slightest and consider it a compliment with anything taken from what I offer or show. Also wasn't mean your machine wasn't pretty but the fact long belts in general are messy.

njhussey
12-09-2015, 05:50 PM
Neil I don't mind in the slightest and consider it a compliment with anything taken from what I offer or show. Also wasn't mean your machine wasn't pretty but the fact long belts in general are messy.

It's messy, I don't mind admitting it, as I've a just got a temporary plate and pulleys on it to get it going.....

Neale
13-09-2015, 11:26 AM
..and once the machine is set up and working, you'll use it to remake components to look better? That's what I keep telling myself!

Clive S
13-09-2015, 02:18 PM
I think you will need more teeth in mesh on the motor there are several examples of how to do it on the build logs.

JAZZCNC
13-09-2015, 07:13 PM
I'm Going to contradict myself now..!!! . . . If your not going straight across with the belts then I wouldn't bother has it's very complicated and messy with belt that long winding around the machine. Keep it simple else stay with Slaved Motors.

Yes the belt approach is much more reliable and gives great piece of mind your always in sync. But not if the setup is overly complex it's needing constant attention and with belt this long threading around the back I fear it may cause trouble.?

So I'd say either change the design and keep it more direct or keep with slaved motors which if properly tuned and connected to good drives don't give too much trouble.

CharlieRam
13-09-2015, 09:54 PM
I don't know if you've seen my long running build thread, Which ive only just started :-) ut this is what I am planning on doing with the motor, It will clear the gantry and bed just.

The first image is showing the belt and bearings, the second is showing it with a 50x25 channel housing the belt and also holding the motor and idler bearings. I may have to use a larger channel but I wont know until I get to that bit.

Cheers, Charlie

Lee Roberts
16-09-2015, 03:28 PM
All fixed!

.Me

ngundtoft
17-09-2015, 10:57 AM
I am back.
I have not been able to post mails, because my thread suddenly disappeared. Lee has been very helpful and has recreated most of it. However, there are still some drawings/pictures missing. JAZZ can you please mail the very instructive gantry pictures again and CharlieRam can you also please send your nice drawings again?
Anyway, I agree with JAZZ, I was not too happy with the very long and complicated belt in my (now missing) drawing. The drawing from CharlieRam is a much better Construction, from a mechanical point of view, but, (as JAZZ also points out) I do not like that it blocks the end of the bed if you need to route long structures.

Previous drawing with long, complicated belt:
http://www.mycncuk.com/attachment.php?attachmentid=16071&stc=1
Drawing where I have put back the extended cantilevers at the top of the machine and moved the now longer, ballscrews, and crossed the belt over directly between them. The bracket serves as cover for the belt, and the motor/idler-pulley assembly goes on top of it, so that it doesn't block the end of the bed.
http://www.mycncuk.com/attachment.php?attachmentid=16073&stc=1


http://www.mycncuk.com/attachment.php?attachmentid=16072&stc=1

Boyan Silyavski
18-09-2015, 11:20 PM
I would go even further and make it so that the belt is tightened by some screw that moves the motor up and down. Additional plate that slides or sth similar.

As for how far the Z should extend is suggest you contemplate this screen snip and read post #8 (http://www.mycncuk.com/threads/8778-Driving-the-gantry-on-Z-axis?p=70040#post70040) where i am trying to explain the benefits of properly designed Z axis


The idea is to achieve maximum Z extension with minimum effort and maximum rigidity as a result, using the spindle body to make Z super stiff.
is more than enough for more


My advice is to understand this/about the Z/ and check your design backwards from there. 150mm Z travel is more than enough for all purposes.

The Z on the picture will extend 200mm with 3kw spindle , but that would seriously question a beginners design. As a precaution the plate should be 20mm steel or 30mm aluminum and reinforced with side braces /against front back deflection/ exactly where the black line is especially at the yellow area. Of course another design problems will arise from such an extension so you have to make sure left/ right you dont have any weeak element on gantry, gantry sides, ball screw mounts, etc.




http://www.mycncuk.com/attachment.php?attachmentid=15500&d=1433957691

JAZZCNC
18-09-2015, 11:43 PM
The Z on the picture will extend 200mm with 3kw spindle , but that would seriously question a beginners design. As a precaution the plate should be 20mm steel or 30mm aluminum and reinforced with side braces /against front back deflection/ exactly where the black line is especially at the yellow area. Of course another design problems will arise from such an extension so you have to make sure left/ right you dont have any weeak element on gantry, gantry sides, ball screw mounts, etc.

Boyan you are going Way over the Top with your idea of a Z axis.! . . . You need to read the man requirements.!!
Using 20mm Steel isn't required for cutting Wood and will actually lower the machines performance due to having high inertia and affecting accleration/velocity. Or increase costs by requireing larger motors or Servo's.

The Z axis has drawn will be perfectly fine for cutting Wood even with 3Kw spindle on it.

Boyan Silyavski
19-09-2015, 09:34 AM
Boyan you are going Way over the Top with your idea of a Z axis.! . . . You need to read the man requirements.!!
Using 20mm Steel isn't required for cutting Wood and will actually lower the machines performance due to having high inertia and affecting accleration/velocity. Or increase costs by requireing larger motors or Servo's.

The Z axis has drawn will be perfectly fine for cutting Wood even with 3Kw spindle on it.


Ok. But...:hysterical:


The way i have drawn it is from 40mm aluminum that weights 6.6kg

Simplified plate 150x500x20mm aluminum is 4kg, steel is 12kg. Not so much weight, for what will bring as rigidity. A couple of welded short bars more at the raised sides and as i like to say- will reap through aluminum like butter even fully extended.

Ok, i will stop drinking Monster Reaper. Too much energy drinks...

JAZZCNC
19-09-2015, 10:51 AM
Ok. But...:hysterical:
Simplified plate 150x500x20mm aluminum is 4kg, steel is 12kg. Not so much weight, for what will bring as rigidity.

That's a cost 3 x weight increase for ridgidty gain that isn't needed in the first place. Then you add in the hidden cost's of lower acceleration which for a Z axis, esp if doing 3D or relief work is quite important. Then if you really need that acceleration it's a move upto Servo's and the extra cost there.!!

I agree 200m travel is excessive in most cases and 150mm covers 98% of work and tools used. But Reaping thru Aluminium at full extension like you say would be foolish thing to do and the cost for the less than 1% time you really have no choice is very high.

It's a simple fact that you can't have ridgidty without it costing something.!! . . . . Money or performance. Wood's etc needs higher performance so can live with lower ridgidty.

ngundtoft
19-09-2015, 12:16 PM
Great input - thank you.
Silyavski your idea, with some kind of threaded screw to adjust the belt tensioning, sounds good. I will find a way to implement that.
As for Z-axis extension (and stiffness) I think I will stick with my 220 mm. I agree with JAZZ that you can do with less by moving the bed up or Down, and I have decided to implement some way to do that. But I will mainly be cutting wood, and for that I don't think it will be a big problem. When cutting aluminium, I think it is ok to raise the bed to a position where the spindle plate is moving around its center position and that will be sufficient. Spindle motor will be 2,2kW.
I am not quite certain about the limit and home switches.
The limit switches/sensors are there to define the absolute outer limits for the gantry/spindle.
I expect that means that you must have limit switches in both sides – 2 for X-axis, 2 for Y-axis and presumable just 1 for the upper Z-axis?
The home switches define a central point on the bed, including a middle position on the Z-axis?
That is a total of 8 switches/sensors if I am correct?

routercnc
19-09-2015, 02:10 PM
Don't really need to do it like that. You need limit on X and Y on furthest away point only. For the other end the limit switch function can be shared with the home switch ( so only one near side limit / home switch needed per axis)
That makes 5 in total. You can add a Z lower limit if you want but not everyone does that.
As you can gather from that your home position is then in the nearest corner (0,0) on a graph paper not the middle. All movement and cutting is done at positive X and Y coordinates and since Z zero is at the top home then cutting is done in the negative Z

JAZZCNC
19-09-2015, 02:33 PM
That makes 5 in total.

I'll name that tune in 3 switches.!! . . . Lol (Ref to Uk Tv program called name that Tune for those out uk)

You can share the Home and Limit function using just one switch. Place that switch on the moving parts and trigger with a Target and you can bring the switches down 3 + 1 for home if slaved motor.

This is my prefered way has it saves wiring and less switches to go wrong.

Also you can have the Home Zero Position a set distance away from the physical switch trigger point. So if you want Home in the middle of the table you still can. Why you would I don't know but you can if required.

Boyan Silyavski
20-09-2015, 06:04 PM
I'll name that tune in 3 switches.!! . . . Lol (Ref to Uk Tv program called name that Tune for those out uk)

You can share the Home and Limit function using just one switch. Place that switch on the moving parts and trigger with a Target and you can bring the switches down 3 + 1 for home if slaved motor.

This is my prefered way has it saves wiring and less switches to go wrong.

Also you can have the Home Zero Position a set distance away from the physical switch trigger point. So if you want Home in the middle of the table you still can. Why you would I don't know but you can if required.

I always forget that. Have to write it down. Again i have bought a ton of switches for my new machine and was just at the point of wiring them.

So you define start and end trigger or you define just one trigger at one end of travel?

Thats sth i don't understand. Have to read again the mach 3 manual obviously.

ngundtoft
20-09-2015, 11:01 PM
I am confused - apparently I am lacking some basic knowledge about the limit/home switching concept (but I am new to all this, so trying to learn).
Basically I can follow the mail from routercnc, which corresponds with my understanding. But when JAZZ limits the number of switches to 3 :chuncky:, I get confused and have some questions:


If we agree that the purpose of limit switches is safety and prevents the machine from exceeding its working area - why then is it not required to have limit switches in both ends of each of the axes? (ok Z is a little different).
I did not really mean that the home position should be in the centre of the bed, it was mainly to distinguish it from the limit switch position. If you always work from the nearest corner (0,0) - as described by routercnc - then the limit switch position and the home position is the same, and what would you need the home position for? Or rather - it only makes sense to use the same switch for both purposes.
Soft limits are set up in the SW as I understand it. But what are the reference points for these soft limits? Asking because you sometimes get the impression that you can almost manage with these soft limits and omit the mechanical limit switches.
I can see that there is not much purpose for a normal limit switch on the lower Z-axis, since the tools will vary in length. But how does the router know the cutting depth? I think I saw a video (Obiwan?) where the spindle was moved to a certain (X,Y) position and then lowered (Z), until the tool made contact to a reference point (or activated a switch), to give a reference position. Is this the way to do it - or are there other ways?

Clive S
20-09-2015, 11:18 PM
If we agree that the purpose of limit switches is safety and prevents the machine from exceeding its working area - why then is it not required to have limit switches in both ends of each of the axes? (ok Z is a little different). This is because they use a target at both ends ie the switch will travel with the gantry and so will hit the target at the end of travel in either direction.
.
Re point 3 It is the homing that give the ref points for the soft limits.
.
re point 4 After you have homed the Z you then move Z down to a ref point (say top of work piece) and then Zero it on the dro so now the machine or should I say the controller knows where zero is and can work out the depth of cut.

Neale
20-09-2015, 11:31 PM
.
Re point 3 It is the homing that give the ref points for the soft limits.


It is really useful to have reliable, accurate, homing switches so that you can quickly set up reference positions when you switch on each time. Picking up accurate positions from one session to the next is difficult without this, whether you use hard limit switches or not. I don't have home switches on my current machine and it's a pain.

njhussey
21-09-2015, 08:02 AM
It is really useful to have reliable, accurate, homing switches so that you can quickly set up reference positions when you switch on each time. Picking up accurate positions from one session to the next is difficult without this, whether you use hard limit switches or not. I don't have home switches on my current machine and it's a pain.
I'll second it being a pain not having homing switches....

ngundtoft
21-09-2015, 10:01 AM
Thanks for clarifying this Clive. Obviously (now that I understand it) you can use the same switches for both limit positions. I see now that it is also what JAZZ writes - just wasn't clear to me.
So, since I am not using slaved motors, I should have 3 limit switches.
So even though JAZZ only uses 3 (+1) switches, it sounds like it is a very nice feature to also have the home switches - or can you also use the limit switches for this purpose? I sounds like it is the same position (0,0) + Z?
I can see the advantage in using the top of the work piece as reference, so that you don't have to measure the difference between a fixed reference point and the work piece surface. How do you do it in practice - do you need a manual rotary control or can you do it from the keyboard (without damaging anything)?

routercnc
21-09-2015, 01:41 PM
This is often a point of confusion for beginners so don't worry. Keep asking until you get it. Once you have the machine and software in front of you it makes even more sense.

Jazz only has 3+1 (roving) switches but still has the home and limit function. They are just shared from the same switch. In terms of Mach3 (or whatever software) you are either homing an axis (so a switch detection means you are at the home point) or you are manually moving / gcode cutting (when a switch detection trips the e-stop for example). You are never doing both at the same time.

I think from what you are saying in your posts you also need to understand the difference between 'machine coordinates' and 'work coordinates'. This also trips up beginners so make sure that also makes sense. Machine coordinates are where the machine is in absolute space and are set to 0,0,0 by going to (referencing to) the home switches. When you are ready to cut you need to go over and down to the corner edge of the workpiece and set the local work coordinates to 0,0,0. This will be at some machine coordinate absolute location such as 100,100,-50. You can switch between looking at machine and work coords in mach3 on the DROs to see where you are on each axis system. The work coordinate locations can be set using a touch probe, or a piece of paper between flute and workpiece using the jog function (e.g. 0.1mm increment movements) until the paper is just trapped.

Now the machine knows where 0,0,0 home position is (in case the job goes wrong and you need to reset, or you need to stop the job to go home). It also knows where the workpiece 0,0,0 is so all the cutting can take place where the work is on the bed.

Have a read of the manual, ask more questions and eventually this will start to make sense.

Christian Knuell
21-09-2015, 02:43 PM
Hi,





If we agree that the purpose of limit switches is safety and prevents the machine from exceeding its working area - why then is it not required to have limit switches in both ends of each of the axes? (ok Z is a little different).



Seeing it from this perspective you'd be right - but my personal opinion is that limit switches are not only unnecessary but even give a deceiving sense of security they cannot provide:



If a machine can be damaged by chrashing into its axis limits it is just a crap machine...
(OK - not talking about professional high speed machining centers where you have tons of material accelerated to very high speeds - but even there a limit switch is just the last line of defense - position feedback will stop those machines way before this could happen)
A machinist not noting the workpiece will require machining outside its limits is just a crap machinist... ;-)
If the machine looses steps the workpiece will most likely be damaged anyways...
And limit switches do not detect step losses - a machine may ruin anything including tool changers or sensors within its work area without the limit switches taking any note of it. Which is also why a machine should never ever run unattended.
So basically: if you really need security you need position feedback.



What sometimes makes sense is reference switches to automatically home the machine. But even though they can make life a bit easier in some situations they are rarely necessary:


Usually you clamp your workpiece just anywhere on your machine and need to touch it off anyways because your machine has no eyes.

In this case knowing machine zero is absolutely pointless - it is just a set of numbers without any meaning for the actual machining process.
Only in case of step losses it can safe you a little time to get back to the correct workpiece zero - if your workpiece isn't destroyed anyways which will most likely be the case... And a machine losing steps regularly enough to make you want homing switches is again just crap or badly set up ;-)
But even in this case you can usually just touch off again (if the reference edges are still available)...


Reference switches are only necessary if you have things like tool changers, tool length sensors or fixed clamping devices on your machine.
And: if the switch isn't precise (or a chip gets in between) machine home will also be imprecise accordingly. Many cheap homing switches are much worse than touching off the workpiece. I program CAM Software and I came across quite some people unnecessarily referencing after every toolchange and wonder why there are visible marks on the finished workpiece that wouldn't be there if they just kept X and Y (or at least used proper limit switches with good repeatability instead of the cheap stuff they bought).


Christian

JAZZCNC
21-09-2015, 04:07 PM
Machine coordinates are where the machine is in absolute space and are set to 0,0,0 by going to (referencing to) the home switches.

Every thing neil says is correct but just want to clear up something that could confuse or be taken wrong by a beginner. Neil says "Absolute" space but this is very different to absolute positioning mode G90 used with-in G-code.
Absolute positioning mode doesn't refer to absolute space but rather it moves an absolute distance in inside the WORK coordinate system. Compared to Incremental Mode G91 which increments by a commanded distance.
ie: G90 X150 would move 150 units away from Work zero. G91 X150 would Increment 150mm from the absolute position so if your at G90 X150 then with G91 X150 you'll end up at X300 absolute work coordinate.

One other slight point I'll make for new folks which I know confuses them is that the Work Zero doesn't have to be on the corner of the Work piece or material. It can be any where you want it to be on the material. Where you located the Zero coordinate in the CAM softare in relation to the part will mostly dictate where on the material you actually place the WORK zero.
ie: For instance if your part was circular like a Wheel for instance then you'll most likely locate Zero at the centre in CAM which means you can't loctate WORK zero on the Corner of material.
In practice often you'll be cutting out of much larger piece of material than the actual parts size so will job to any location that the part will fit within and knowing the Programmed Zero point on the part choose point in that area.


Now to use one switch for Both homes and Limits you need to turn off Home SW safety option in genreal Config. This will tell Mach3 to ignore the Limts while the Machine is Homing.
Your also better to turn on Auto Limit over ride in the setting screen. This will allow you to reset Mach3 and drive off the switch. If you don't turn this on you will need to include a Limit override Switch in your Wiring to knock out the limits while you reverse off the switch.

JAZZCNC
21-09-2015, 04:41 PM
Seeing it from this perspective you'd be right - but my personal opinion is that limit switches are not only unnecessary but even give a deceiving sense of security they cannot provide:

I agree upto a point but here are some cases they are required. Servo's for instance will just take off at full speed if there positioning system fails or gets interupted. So in this case they are very much required.

Also with small fast router machines it's very easy to jog from one side to the other which can result in banging into end stops etc often. Esp when first learning. Again I agree a machine that breaks with just a small limit crash is Crap but enough of this banging and eventually damage will occur so best avoided and for just a few sheckels more is it really worth not having.??




What sometimes makes sense is reference switches to automatically home the machine. But even though they can make life a bit easier in some situations they are rarely necessary:


Usually you clamp your workpiece just anywhere on your machine and need to touch it off anyways because your machine has no eyes.
[LIST]
In this case knowing machine zero is absolutely pointless - it is just a set of numbers without any meaning for the actual machining process.
Only in case of step losses it can safe you a little time to get back to the correct workpiece zero - if your workpiece isn't destroyed anyways which will most likely be the case... And a machine losing steps regularly enough to make you want homing switches is again just crap or badly set up ;-)
But even in this case you can usually just touch off again (if the reference edges are still available)...


Reference switches are only necessary if you have things like tool changers, tool length sensors or fixed clamping devices on your machine.
And: if the switch isn't precise (or a chip gets in between) machine home will also be imprecise accordingly. Many cheap homing switches are much worse than touching off the workpiece.

This I mostly Disagree with.!!

Homing switches are THE most useful thing you can put on your machine. I'll give examples to why.

Only yesterday I was cutting parts in large expensive sheet of aluminium when the cutter broke. The part wasn't damaged but it was my last cutter so the Job was stopped. The corner of the material where I choose Work Zero had been cut away. So how without Home switches do I get back into absolute position when my reference point is gone.?
Yes I could go back into Cam and choose some hole to re-reference from etc but this all costs time and effort. Home switches provided resonable quality remove all this hassle.

Example #2 Same large piece of material and Large Job with multiple parts nested into material. So large I don't have time to do in one sitting. Again Zero ref has been cutaway. Home switches allow me to pause the Job or break the Job into smaller jobs and still quickly get back into position.

There are other examples like power cuts or PC crashes etc that back the need for Home swithces but again for the small cost it's silly not having them.
The only thing is that if you need high accurecy homing then you need Good Home Switches. I find even your Cheap Chinese Proximity switches are more than good enough for 99% of Routers or DIY'ers using Milling machines.

So for the Small cost and the fact Limits and Homes can share the same switch it's Stupid not to have them.!!

routercnc
21-09-2015, 07:33 PM
Every thing neil says is correct .

Was me not Neil :rolleyes: but agree with you on use of the term 'absolute'.

I ran without homes and limits for several months, then added some switches (but never got round to wiring them up) and one day jogged into one and smashed it to pieces and banged into the end stop. Makes you realise how powerful these machines are and to keep your hands out of the way.

Just got round to adding limits and homes so should be OK now.

JAZZCNC
21-09-2015, 08:41 PM
Was me not Neil :rolleyes: but agree with you on use of the term 'absolute'.

Doh I blame coolant fumes.!!


I ran without homes and limits for several months, then added some switches (but never got round to wiring them up) and one day jogged into one and smashed it to pieces and banged into the end stop. Makes you realise how powerful these machines are and to keep your hands out of the way.

Ye try snapping a 12mm Carbide cutter by hand.!! . . . Even a small machine will snap one like a carrot.!!

ngundtoft
24-09-2015, 11:45 AM
Thanks - lots of really useful info. I know a lot more about limit/home switches now.
So 3 switches it is. Based on the above, I cannot really see why you would want to use separate home switches - unless you want some very special home position?
And let us switch tune:playful:. I have updated my model with a single Nema 34 motor for the Y-axis and an adjustable mounting flange, as suggested by Silyavski. Is 20T pulleys/idler pulleys ok for this, or should you choose larger ones?
http://www.mycncuk.com/attachment.php?attachmentid=16168&stc=1

JAZZCNC
24-09-2015, 02:26 PM
Is 20T pulleys/idler pulleys ok for this, or should you choose larger ones?

Yes they will be fine arranged like this with nearly half the teeth engage. Don't want to go too large other wise the extra interia starts to affect things.

ngundtoft
24-09-2015, 09:40 PM
I have been thinking about how to design a holder so that you can route dove tails at the end of boards.
This is not a function I will be using every day, so it should be simple and it should not compromise the machine in other ways.
I read the thread by EddyCurrent and how he has done it. It looks good and stable and I am sure it works very well.
However, I don’t want this function to interfere with the precision of the bed, so I am not too happy with the removable plate. Also, for handling, I find it somewhat impractical to have this opening in the middle of the machine. All that is needed is a jig that will hold the work piece in an accurate and repeatable position. So I was thinking of placing it at the end of the machine, as shown in the model. You simply fasten the work piece with one or two clamps. I may have to extend the cantilevers/ rails to get sufficient overhang for the spindle or alternatively reduce the length of the frame, since this machine is already getting fairly large.
http://www.mycncuk.com/attachment.php?attachmentid=16169&stc=1

njhussey
24-09-2015, 10:04 PM
I remember Jazz showing (it had a red frame I think) a design that did just this, it could take panels etc on the front to rout the ends. This is one thing that I'd on my list for mk2.......

JAZZCNC
25-09-2015, 12:54 AM
Something like this works well for holding boards on end.

Slots with Pockets on backside.

16170

ngundtoft
25-09-2015, 09:47 AM
Thanks JAZZ. This is actually a brilliant design. It is simple to make, it can be removed from the machine when you are not using it and the width it can accomodate is adjustable.

The only thing that may not be optimum is the thickness of the material you can clamp. I remember using this type of clamps (although much smaller) and the reach is not so big. I tried to find the reach for similar types. I only found this specification for a single one and it was 3". I am not sure if that means it will go from 0 to 3" or if it is just the maximum. Maybe you can use this type of lever clamp instead (although they are quite expensive):
http://www.mycncuk.com/attachment.php?attachmentid=16171&stc=1

JAZZCNC
25-09-2015, 02:11 PM
There are many ways to do this and many types of clamps to help but the simple and easy way's are often the most affective. In this case it would be a simple spreader bar which is bowed so when clamping wide boards pressure is applied in the centre.

I was only really offering simple example of how to do it but not really being specific regards clamping etc.

hanermo2
26-09-2015, 01:52 PM
Excellent start !

One point.
A tiny nema 23 stepper, on my machine, moves my bed mill, effortlessly.
The bed is 1600 mm wide, 60 mm thick on ribs, tool steel.
Mass is 200 kg, with upto 200 kg of stuff on it.

The "right" way to do it is drive the moving bit at 1:3 with the stepper with HTD belts.

This provides better overhead, much better acceleration, better accuracy, and about the same speeds.

If nema 23 steppers dont work, get ac brushless servos, at 130€ each, or 290€ for servo motor, driver, cables, encoder, everything.
They cost == the same as 34 size steppers, and
-5x more accurate
-5x better acceleration (more than you need)
-3-5x top speed (too much. You wont want all the speed).

The top speed is not important,
the cutting speed is very important,
accuracy is vital,
and rigidity is everything.

Just my experience.
Modern brushless servo motors are
-not noisy, (IMPORTANT),
-dont backdrive and activate with back-emf (VERY IMPORTANT),
-wont take off if the cables get cut or disconnected,
-have anti-jitter circuits,
-can be used with differential connections (less sensitive to noise)
-have a disable-input on them.

Best,
hanermo

Building cnc stuff since 2002, mostly full-time.

ngundtoft
26-09-2015, 02:10 PM
Yes, and thanks for great input JAZZ, it is just me, whenever I am working with something I always start thinking if it can somehow be improved - even if it is great to begin with.
Actually, I think it will be very rare that I need to clamp very thick material, and if so I could clamp it to the frame with normal wooden clamps, or even make some kind of spreader bar as you suggest. So I will use the toggle clamps - and they are cheap too.
http://www.mycncuk.com/attachment.php?attachmentid=16173&stc=1

ngundtoft
26-09-2015, 02:28 PM
Thank you hanermo.
I am really impressed to hear that a Neme 23 can move a 200kg gantry - even with some gearing. So my measly 60kg gantry should not be any problem?
I have not really considered servo motors because most DIY people seem to be using stepper motors, and it is my empression that servo motors are more complicated to use and much more expensive. Also, I think you can get a Nema 34 and driver cheaper than 290 Euro (cannot find that symbol).
I am convinced that brushless servo motors are excellent - maybe I should look more into that.

hanermo2
26-09-2015, 03:16 PM
Euro symbol is along with dollar and pound, one more to the right.
€ alt-5 on most keyboards (.fi),
€ alt/4 on .en
alt-5 on .es

The nr1 imporant piece of gear is the motion controller.
The Pokeys for example, with hardware MPGs ! and thats crucial, move my machine with an MPG as well as a 100 grand Haas.

The MPG is wired directly into pokeys, no USB.
MPG wheels cost about 30€ each, and I put one on every axis of every machine.

Pokeys supports 3 at once with Mach3, and will likely support more (they sort of tole me they can).
CSLabs also told me that they can support dedicated MPGs.

Trust me, dedicated wheels like on a lathe, where the same MPG, in the same position, always moves the x axis, is the dogs danglers.
Just like manual, but clean, fast, 10x more accurate, perfectly repeatble.

Important point !
A usb MPG cannot be good as .. ! Because the USB has variable latencies (to 1 ssecond 100.000 less).

On a Pokeys, the engine runs at 125 kHz,
so every 0.01 ms, it checks if it needs to move an axis.

This makes it "feel" like its solidly connected to the machine.
Huge difference.

ngundtoft
26-09-2015, 03:48 PM
Ah - thanks for help with € - it is on the 5 key, but not marked.
Did not know that motion controller (MPG) is so important. Is it so much better to have separate ones for each axis? It seems that many of them can be used for all axes. What does "100 grand Haas" mean?.
I did look at the Pokeys Electronics (controller and BOB) - what is your impression with that?
With respect to servo motors - what specifications should you be looking for if you want to replace Nema 23 or 34?

JAZZCNC
26-09-2015, 04:15 PM
MPG is nice but it's not required for general use and on a machine used mostly for wood ie: router then individual MPG's are wasted money to me. On lathe or Milling machien with lots of setting up etc then yes I'd probably consider them but not on wood router.! One is enough and then it's more of a luxury than needed.!

Same goes with servos for me with Small/Med router based machines like this. The price difference is 3 x that of Decent steppers. AM882 or Em806 Digital drives with 3-4Nm steppers will be less than £100 per axis and they easily push 60Kg around at 10,000mm/min or more with 10mm pitch screw which is plenty fast enough for any DIY machine.

hanermo2
26-09-2015, 06:29 PM
The good part about dedicated MPGs is that they work like a manual machine.
Ie you are never switching on the pendant to a given MPG.

It makes the machine extremely productive for one-offs, for manual use, and for setups.
Most faster than any manual every made.

The 2 best controllers I know of are;
1. CSMIO. Very good, expensive full-on (1000€ and up, industrial).
2. Pokeys. Very good, cheap (150 € ish).

Both have excellent drivers, excellent support, and good features.
Use ethernet !

It may be instructive to note, I am not mentioning anything else, I have stopped using 3 other solutions.
The others that may be good are machmotion and something else, .. no personal experience.
Use ethernet !

Ethernet is galvanically isolated and has very low latency.

hanermo2
26-09-2015, 06:32 PM
??
I found an Em808 at 115 €.
Add 130 for motor, 30 for proper cables and connectors, == 275 €.

With the servo at 290€, ..
Seems about the same price to me. And you need to make the fussy cables.
No ?


The price difference is 3 x that of Decent steppers. AM882 or Em806 Digital drives with 3-4Nm steppers will be less than £100 per axis ...

hanermo2
26-09-2015, 06:42 PM
The average cost of the Haas machines I worked with, and sold, were 73k€, approx 100.000 $ at the time.
100 grand.
Haas is a large, very succesful, builder of large industrial machine tools. Nr in the the USA, nr 1 by nr of machines, in the world.

Disclosure: I used to be their sales manager for Spain.


What does "100 grand Haas" mean?.

ngundtoft
26-09-2015, 08:57 PM
Hi Hanermo2. Thank you for your input, I will consider an MPG - it sounds like a nice feature. But you are way out my league. I am building this machine because I cannot afford to buy a similar machine (and because I think it is fun). Obviously for professionals and companies a difference of a few hundred € per axis is no big deal, but for for my little private project it makes a difference. Also buying one or three MPG's makes a difference.

ngundtoft
26-09-2015, 09:52 PM
A small update to the dovetail jig. Since you can easily remove the vertical braces, I figured that you could just clamp thicker things directly to the base plate with clamps - same system as used in the Festool worktable (you can use the same clamps). If you are normally routing items that are not so wide, you can also have one of the vertical braces fixed tightly as a fence to ensure that the work items are perpendicular to the bed.
http://www.mycncuk.com/attachment.php?attachmentid=16174&stc=1

JAZZCNC
27-09-2015, 12:13 AM
??
I found an Em808 at 115 €.
Add 130 for motor, 30 for proper cables and connectors, == 275 €.

Hannu I didn't say EM808 I said EM806 and the AM882 which is pretty much the same drive can be bought for £50 add £25 for 3Nm nema 23 and that's well under £100. I fit these all the time to routers and they easily handle what is required for router machine this size.

JAZZCNC
27-09-2015, 12:25 AM
A small update to the dovetail jig. Since you can easily remove the vertical braces, I figured that you could just clamp thicker things directly to the base plate with clamps - same system as used in the Festool worktable (you can use the same clamps). If you are normally routing items that are not so wide, you can also have one of the vertical braces fixed tightly as a fence to ensure that the work items are perpendicular to the bed.

If it was me I'd have a combination of 3 things.
#1 Clamps as shown.
#2 Matrix Grid of Threaded holes for clamping Odd shapes with mill type clamps.
#3 several larger holes for Festo type clamps for deep items.

Oh and I forgot to say I'd get lift those bottom rails up about 75-100mm to create short stumpy feet for levleing and lose 2 of the diagonals wich are not required. Less welding, less chance of heat distortion.

ngundtoft
27-09-2015, 10:58 AM
Yes, good idea JAZZ. Some holes with T-nuts would increase the possibilities for fastening things. I actually thought that I extended the legs under the lower frame, but apparently I didn't. Also eliminated the lower beam I inserted for mounting the dovetail jig. I removed 2 of the diagonals since they are not required.
http://www.mycncuk.com/attachment.php?attachmentid=16175&stc=1

hanermo2
27-09-2015, 12:05 PM
Yes, I agee on the costs.
Many of my choices would be different if I did not do things for commercial uses.
Haas has the lowest costs in the industry.

A servo amplifier (ie driver) (just electronics) costs 500€.
The same from yakawa or siemens costs 2500€.
The ones I use cost about 350 for the same size, circa 1kW.

Like I said, the MPG wheel is only 30 or 40.

ngundtoft
27-09-2015, 12:10 PM
If money was no object I would go for the servo motors.

Do you have a link to a good MPG for 30 or 40€? The ones I found were approx. twice that.

JAZZCNC
27-09-2015, 12:45 PM
Yes, good idea JAZZ. Some holes with T-nuts would increase the possibilities for fastening things.

One of the things you'll probably use in vertical position is lower bracket or shelf like setup to help support and locate material when clamping. This makes repeat work like dovetails much easier and the Thread holes come in use full for things like this so I'd take them all the way to bottom.

JAZZCNC
27-09-2015, 12:51 PM
Do you have a link to a good MPG for 30 or 40€? The ones I found were approx. twice that.

Depends if your just talking about the MPG pulse generator like hannu probably is or what people often mistakenly call MPG but should call Pendant with all the other controls on it.?

Here's 2 of the MPG pulse generators. 25 pulses per Rev and 100 P/R. 100 is most commonly used I think.!

http://www.aliexpress.com/store/product/50-8mm-Hand-Wheel-Pulse-Encoder-DC-5V-25P-R-For-Cnc-Machine-Tool-Mill-Router/314742_540014466.html

http://www.aliexpress.com/store/product/50-8mm-Manual-Pulse-Generator-electronic-hand/314742_506063450.html

ngundtoft
27-09-2015, 02:50 PM
Hi JAZZ - yes absolutely, some way of fitting an end stop is very practical. I also realised that the "overhang", or working distance from the face of the dovetail jig, is only 91mm. So maybe I should extend the Y-axis rails by 50mm or so.
Thanks for the links. If this is the type that hanermo2 uses, I guess that you will need 3 - or maybe the Pendant type is just as good?

ngundtoft
29-09-2015, 02:30 PM
It seems I am getting close to having a mechanical design that will fulfill my requirements. However, there is still one point on my list that I need to look into. I would like to have some kind of adjustable bed. I will mainly be working with wood, so I think an MDF plate will be sufficient for that, and also with wood the fairly long Z-axis will not be a problem. But for aluminium I think it will be good to lift the bed for better stability. This bed could be smaller and also have provision for clamping the work material.
Something like this:
http://www.mycncuk.com/attachment.php?attachmentid=16184&stc=1http://www.mycncuk.com/attachment.php?attachmentid=16185&stc=1
If anyone has come up with a mechanically good construction for raising/lowering the bed, I would love to have some info/links.

hanermo2
29-09-2015, 02:51 PM
Right, ok, now I get it. I misunderstood.
Spot on.

I have a 10-pack of new 2M542 stepper drives, cost 45€ each. Very good drives, I removed the gecko (251) drives from service. Less noise, no back emf, differential (I used single-ended).

These pair well with the cheap 23 size steppers, and are, as you said, ok for lots of smaller machines.
Well under 100€, as you said.
For small machines with nema 23 motors, that type of stuff is good, and cheap.

Its only once you go to larger steppers and higher voltage/current drivers the cost goes up, and is then about the same as servos.



Hannu I didn't say EM808 I said EM806 and the AM882 which is pretty much the same drive can be bought for £50 add £25 for 3Nm nema 23 and that's well under £100. I fit these all the time to routers and they easily handle what is required for router machine this size.

ngundtoft
04-10-2015, 11:13 PM
I have started looking at the electrical side of the project. I noticed that GTJim switched to the PMDX-126 BOB, and has a setup very similar to my own. Since the Pilz safety relay PNOZ X9 seems to be available from eBay for reasonable prices, and I do not have much experience with these relays, I have based my diagram on the very nice drawings that GTJim has made.
I have a few questions though:


You can connect the stepper drivers to the BOB as open collector (common anode) or PNP (common cathode). Has anyone noticed any differences (with eg. noise) by using one or the other type?
The stepper drivers have Enable signal terminals – which output terminals from PMDX-126 should you use?
The stepper drivers also have alarm signal terminals – if you would like to use this function how should it be implemented? Should the relay cut power or activate the K pin (machine hold function) on the BOB?
GTJim has implemented a limit override switch – what is the practical use for this?
Can you use the K1 or K2 relay (in PMDX-126 for controlling a vacuum cleaner?
I also noticed that GTJIm uses a separate time delay relay for the power supply (1 to 10s). But this is not really required. I noticed that the Pilz safety relay has an on-delay of approx. 150ms (PNOZ 10) which is perfect for this use. 1 to 10 s is way too long time and only results in heating up the resistor. I could not find a data sheet/wiring diagram for Pilz PNOZ X9. Does anyone have a link for this?
I am not sure if I should use single or 3-phase connection to spindle motor. 3-phase operation can be implemented by using a contactor controlled from the Pilz safety relay.
Mains power indicator – since the other power supplies have indicators I think that this one is not really required

JAZZCNC
05-10-2015, 12:43 AM
#1 No difference really take you pick.!
#2 Any spare output you want to use. Then set Mach3 Enable output to that pin number.
#3 Use the Fault Input.
#4 Because when you trip a limit then you need to temp override that limit to drive off the switch.
#5 Yes
#6 Google is your friend try this.! http://www.is-com.ru/files/PNOZ_X9_en.pdf
#7 Can only use three phase if it's a Three phase Input which it won't be unless you asked for 3 phase input VFD. If single phase Just use R & S or R & T both work.
#8 This so you can see it's powered up on the outside of the control box. Personal choice but I prefer to see if box is powered so it doesn't get left on.

ngundtoft
05-10-2015, 12:38 PM
Great - thanks a lot JAZZ. I have made changes according to your input.
1. Will keep as is.
6. Google must be a better friend of yours than mine :encouragement: - I could not find it, so thank you for the link.
7. From the datasheet for the Huanyang HY02D2223B it looks like it can be used both for connection to single phase and 3-phase.
8. Yes, I was not so clear here. I intend to put the LEDs for the three power supplies on the front of the control box and I think that would really be sufficient