My First Router, Built in Steel

**Haydn** · 02-04-2007, 06:02 PM

Hi John,

Thanks for the comments. I haven't started using the machine yet. I have just been doing some motor tuning as I finally get an hour here and there to do the wiring. Initially things are looking good, far better than I had hoped for a stepper system. I currently have X (the axis I had concerns over-weight/inertia etc) running quite happily at 7000mm/min and have reached rapid max of 14500mm/min which looks impressive but i'm sure would be totaly unreliable without servos and encoders. At 15000mm/min the x axis locks and motor stalls.
I guess that each jointing method has its pros and cons. For me using epoxy, bolts and knock pins seems to be working well. I am a woodworker not a metal worker, I do a little welding from time to time but don't have the time or inclination to learn to stitch weld etc to get anywhere close to a stress free frame. This is my first machine and I kind of want to do the best I can as I can't afford the time to do it twice. By using epoxy and bolts I am able to reasonably easily correct problems I come across. I have seen people mention that epoxy joints tend to absorb vibration to some degree and I think i'm inclined to agree. One problem with the epoxy bed for rails that thankfully I didn't have to try and solve is a mis-placed hole cannot be welded and re-drilled as the epoxy will obviously melt. I guess that bolts can work loose but to me thats part of general machine maintenance. All bolts are threadlocked. Some where the joint shouldn't need to come apart just have epoxy squeeze out on the threads when they were joined. These do still unbolt but I have to put a bar on the allan key as they become very stiff-I don't think these will loosen but we will see very soon!

I looked around to see what other people were doing and what jointing systems were within my means without investing extra time and money. Seeing what people can achieve using epoxies on manufacturers web sites helped too. It seems to have worked for me but we shall see over time. The choice is varied and yours!!

Haydn

**Haydn** · 02-04-2007, 07:03 PM

Next tasks were fitting the Y axis ballscrew and making/fitting the Z motor mount. The ballscrew support bearings were mounted on the aluminium gantry uprights, a jig was made to lay a bed of epoxy on the aluminium for the bearings at equal distance from the Y rails (remember that the extusion does not have flat faces). The ball screw was fitted with a measured runout of 0.02-0.03mm. This was higher than I hoped to achieve but the screw had a small bend in it which made measurement tricky. In use the screw is very smooth with no binding or stiffening. I will check that the nut does not get warm in use, indicating an alignment problem. I then manually machined a yoke to connect the ball nut to the Z plate. You can see the epoxy bed for the bearings in the pic.

**Haydn** · 02-08-2007, 02:15 AM

[QUOTE=Haydn;249546] I managed 7500mm/min without too much trouble at 10000mm/min the motor could not overcome the inertia. I am quite happy with this as I planned to get somewhere in the region of 5000mm/min./QUOTE]

Please ignore this comment from one of my previous posts, I am getting wrong information from mach, I have discovered a timing issue with the PC which I am trying to sort!

**Haydn** · 02-08-2007, 02:26 AM

Here are a couple of pics showing where it's up to at this stage. Any questions fire away i'll do my best!

**Haydn** · 02-08-2007, 03:07 PM

OK, I sorted the timing issue and have got even better results. I have managed to run rapids at 13500mm/min and still have the gantry drag me across the floor! Sweet and smooth.

Things are looking up again.

**Mike F** · 02-08-2007, 06:27 PM

Haydn,

I don't understand how you are managing to get rapids of 13.5m/min using a ballscrew? How long and what diameter is the ballscrew? I am using a 2 metre long ballscrew, 25mm diameter an 5mm pitch and can only get rapids of 3.5-4m/min because the ballscrew is limited to an rpm of 800 to avoid whipping. I can only assume your ballscrew has a very coarse pitch.

Please enlighten me.

Mike

**Haydn** · 02-08-2007, 08:04 PM

Hi Mike,

My screw is 1500mm long, 25mm dia and 10mm lead, it's supported with THK BK and BF17's. The screw is direct driven and has a runout of 0.01mm after fitting. The screw itself is very stiff which was useful, making alignment much easier than the smaller Y screw. The motors I am using are 1.8 deg 640oz in, these are running at 60v 5.5A which is pretty close to their limit. This is deliberate so I can get better acceleration. At 13.5m/min the stepper would be going 1350rpm, any faster and i'm sure the torque will drop off very quickly. The screw runs true, you cant see any wobble at all. I can see a bit on the Y screw though this did have a slight bend in it before fitting.
Did you get your screw new or secondhand? How accurate is the end machining? If the end machining was a bit excentric that could cause the problem.....just a thought.

Haydn

Edit: my pitch is twice yours so were probably going to be in the same ballpark as far as running/working the machine goes.

**Mike F** · 02-09-2007, 04:07 AM

Haydn,

My screws are all brand new, THK, machined by THK and using the same fixings as you, it's just that the manufacturers data says I should not run the X axis screw any more than 800rpm which would give me 4000mm/min. You must be exceeding the recommended rpm limit for your screw if you are achieving 13500mm/min. I have to say that my servos, 430oz/in from HomeCNC, will only spin at 1600rpm max and with a 2:1 reduction, the screw turns at 800rpm and at this speed there is absolutely no deflection or any sign of whipping at all. I feel pretty confident that I could spin the screw much faster but to start with, I thought I had better remain inside the manufacturers recommended speed.

I appreciate your screw is shorter and twice the pitch and it would be interesting to know what the recommended speed should be and by how much you are exceeding it. If I find the time I will look up the data. If my calculations are right, you must be spinning your screw at 1350rpm.

Mike

**Haydn** · 02-09-2007, 05:22 AM

Hi Mike,

Thats interesting to hear, I rang the man who made the screw for me. He said I should be ok at about 1000rpm and that manufacturers specs are often below the real limits which I can understand. He did say that a 10mm screw will wip less than a 5mm screw but the maths is a nightmare! Though I can run at that speed I won't be, I don't need to. Ideally (which I didn't think I would achieve) I wanted to be able to feed at carbide spiral cutter rates which are between 5 and 10m/min. I do this manually at work and the cutters last so much longer if feed is high. So in a working environment it is likely that I probably wont rapid over 10m/min, no point in running it to such a point that ballnuts wear prematurely etc. Though the screw is great and am very happy, I am more surprised that the motors can handle it!

Thanks for asking about this, it made me clarify the point with the engineer that made the screw (and manufactures cnc machines), in his opinion it will probably be ok but, probably has a hint of doubt which could turn out to be expensive!

Thanks
Haydn

**Haydn** · 02-09-2007, 08:09 AM

Hi Mike again,

Just to clarify my ballscrew is not a THK. There are a lot of variables that determine working limits of a screw, nut rotational load, ball size etc DN and DP values are variable through different screw types, and the resonant frequency of the screw itself which may well be why you are suffering with wobble over 800rpm.

Just a thought, the maths is beyond me!

**Mike F** · 02-09-2007, 08:57 AM

Haydn,

Thanks for the reply. I am not actually experiencing any flexing of the screw but that is only because I cannot drive the screw fast enough to find out what its limit is

My limiting factor is the servos - they can only spin at around 1600. I suppose I could try a 1:1 drive and double the speed but that will be an experiment well in the future.

It would be nice to rapid at your kind of speeds, especially with a machine that is 2000mm long, but I will only be cutting at very much lower speeds, well within the current setup limits.

Thanks again,

Mike

**JerryFlyGuy** · 02-09-2007, 10:01 AM

Haydn, a couple questions and a point or two.

Questions first, what make of stepper are you using? I've got 640's as well and was wondering if you have a rough 'force' number. Do you think you are getting ~100lb's at this 1350 rpm? From your numbers I can then calculate what my machine's gonna do at that stepper speed.. and interpolate what it will do between there and 0 rpm. I've done the math from 'the ground up' but real life experiance is the real story.

Also, are you using a pre-loaded ballscrew? [Zero backlash or something?] I could see that using one of these 'tighter' ballscrews might decrease the efficency of the screw, maybe this is why Mike can't get as high a speed due to increased pre-load on his screw??

What kind of accelleration time do you have? Ie; full stop to full 13.5m/min is X seconds..

I find it interesting that you can get to 1350rpm w/ that large of a stepper. I assume its a 34 frame motor? I'd always 'guessed' that they wouldn't really get over 1000rpm or so. I used to always calculate my stepper stuff from the torque, however Mariss has been kind enough to clarify that we should be calculating the 'power' not just the torque [Torque*rpm/ some number] will give power, power is a real indication of performance over just the torque value.

Thanks for the great thread and pictures!

Jerry

**Haydn** · 02-10-2007, 08:16 AM

Hi Jerry,

I have the Rex RHT640 motors. I don't have a torque curve for these and had a look at the HomeShop site where I got them from and it looks like these have been replaced with a 740oz version. I don't have any means to measure the thrust at these speeds, all I can say is I cant stop by leaning on the gantry! Please remember that these are only rough tests. I am supplying the motors 60V at 5.5A which is pretty well at their limits. I may well find that if the motors were run for any time and these speeds they may get overly hot due to iron losses. I have the acceleration set to 300mm/sec/sec. If your motors have a high detent torque you probably wont get decent acceleration and speed, a problem with triple stack and 42 size motors. I think running at these speeds in a working situation may cause the gantry to sway if acceleration/deceleration was too high. Hopefully I will have the real life experience soon! I didn't do the maths, I looked to see what was working for other people and took it from there - other peoples real life experiences, I found too many variables in the theory and measurements I couldn't make. I remember reading an article somewhere about steppers performing well around the 1000-1200rpm mark and performance dropping off sharply after that, wish I could remember where it was.

Here is an email quote from homeshop :

"I would recommend 60-65 volts as this will give you more high speed torque. The motors will run quite warm but that's OK. Maris likes to see people use his drivers to run motors on the cool side but I've pressed him on this before and he will concede that steppers are made to run hot and that the losses that occure are not enough to result in lower torque that if run cool. Many people run out 640's at 68 volts w/Geckos for the extra torque."

We shall see how hot they get! 68V would be over 25 times the rated voltage. I use a 180K current limit resistor, I see some people have opted not to which can give higgher high speed torque but linearity is lost as the motors could draw up to 7A.

My ballscrew has adjustable preload/backlash. At present this is set for a light preload.

Haven't calculated the power, but do remember Mariss telling me that over 200w servo's would be more appropriate.

Hope this helps

Haydn

**JerryFlyGuy** · 02-10-2007, 10:34 AM

Thanks Haydn, I've purchased different motors than yourself. They are 640's and I think they are triples but can't remember for sure. Anyway, to really know a comparable performance I'd need some type of force number but thats ok. Really, the only way to know what is going to happen is to try it and see, which will be happening fairly soon I hope

Thanks for your time Haydn

Jerry

**Haydn** · 02-10-2007, 07:13 PM

Hope things work out as you plan Jerry.

Haydn

**Haydn** · 02-10-2007, 07:23 PM

Here you can see the Y axis motor mount. The mount was machined manually, I think there is a pic of part of it being machined in a previous post. The counter bore for the motor frame was done with a jig and a hand router at low speeds and fine cuts. In the second pic you can see Y and Z motors fitted, the Z mount is a 20mm thick aluminium plate. The X axis ballscrew is assembled and waits to be fitted.

**Haydn** · 02-11-2007, 05:59 PM

I had done most of the electronics by the time I got to this stage. Here is a picture of the controller. You have left to right, main on, start and e-stop motor ground x, y, z.

On the back bottom left to top left, 5A fuses Geckos, machine e-stop out, x,z , motor out, limits x, y, z parallel port. On the right ther is mains in and a spindle in/out and extractor in/out controlled by SSR's.

**bp092** · 02-11-2007, 07:37 PM

you did a really nice job on the controller, any pics of the insides

?

**Haydn** · 02-12-2007, 05:13 AM

Hi BP,

Here are some pics of the inside of the controller, still a couple of bits to wire at this stage, two SSR's for spindle and vac control, the remote machine mounted e-stop circuit and the limit switch connections from jack sockets to breakout board. You can see the two SSR's in the last pic.

Haydn