My first part - and a dimple

This forum is dedicated to those Hobbyists Interested in CNC machining & 3D Printing in their home shops. (Digital Read Outs are also topical, as is CAD/CAM as it relates to CNC)

Moderator: Harold_V

User avatar
ctwo
Posts: 2996
Joined: Tue Mar 27, 2012 12:37 pm
Location: Silly Cone Valley

Re: My first part - and a dimple

Post by ctwo »

Thank you, RET.

I like how the manual machines give feedback and you can adjust quickly on the fly. I also like the flexibility of complex cutting paths of the CNC, and I can also use my other senses to adjust the feed right on the spot!

I programmed the short segments because eventually Z will be tracking a sine curve.

I am currently working on a heart shaped 3D piece.

BTW, EMs are expensive!
Standards are so important that everyone must have their own...
To measure is to know - Lord Kelvin
Disclaimer: I'm just a guy with a few machines...
RET
Posts: 960
Joined: Wed Jun 07, 2006 8:36 am
Location: Toronto, Canada

Re: My first part - and a dimple

Post by RET »

Hi ctwo,

In the picture, this is what I meant when I was talking about the almost "mirror finish" I get in 6061-6 aluminum. As you can see, the finish is quite good and if you look closely, you can see the edges are sharp, as a matter of fact, they will cut you if you aren't careful. What you see is the part as machined with no coolant or polishing. Once I break all the sharp edges, the part is finished.

In playing around with this machine, I've found that there can't be any play anywhere, even a few tenths. Unless your Bridgeport is in very good condition, the "Y" axis dovetails usually wear in an hourglass configuration resulting in some play (usually several thousandths) in the middle of the travel. I know because that's how my Bridgeport is.

In the CNC mill I built, the ball screw nut is preloaded on its screw in each axis and each ball bushing is also preloaded on its Thomson shaft (2 bushings per shaft). In turn each shaft is bolted to its support rail, so there is no slop or give anywhere in the system. Another advantage of the ball bushing system is that there is almost no friction, the only drag is the lip seals that keep the dirt out and the grease in so much smaller steppers can be used for the same working force. It takes large shafts and ball screws to get the rigidity, but as I said, the friction is minimal.

The traditional dovetail way type machine is inherently stiffer because of its design, but it is almost impossible to get ALL the play out. Each type of construction has its pluses and minuses and you can argue both sides of the subject about which one is better. I believe some commercial CNC machines are built using either method.

By the way, if you cut a sine curve around the outside of your part, the cutter will gouge in the valleys. If that isn't acceptable, you need to add an "A" axis (rotary) with that axis parallel to the "X" axis and coincident with the axis of the part. That will remove the gouging on the sine curve, but you will have to have that same rotary axis vertical for the cuts at the top of the part. That would be one reason for having a 5 axis machine where the rotary axis is automatically changed to do the different faces.

Perhaps some of this will give you a bit to think about. Just what I've picked up in my fooling around with CNC.

Richard Trounce.
Attachments
Riding Truck side finish.jpg
User avatar
ctwo
Posts: 2996
Joined: Tue Mar 27, 2012 12:37 pm
Location: Silly Cone Valley

Re: My first part - and a dimple

Post by ctwo »

Hi RET, the Bridgeport is not in the tenths range. It will half-step unreliably because I am using the original steppers. The machine is otherwise in good condition. I set the gibs during assembly and there was no binding at the ends.

I am not visualizing why there would be gouging, but I note there will be rounded corners. I had thought of the 4th axis since I have set up a stepper on an 8" HV rotary table as well. That is what the part is actually for. It's not really needed to be that way, just an idea.
Attachments
Bridgeport-ways-20160119_072417.jpg
Standards are so important that everyone must have their own...
To measure is to know - Lord Kelvin
Disclaimer: I'm just a guy with a few machines...
RET
Posts: 960
Joined: Wed Jun 07, 2006 8:36 am
Location: Toronto, Canada

Re: My first part - and a dimple

Post by RET »

Hi ctwo,

You are very lucky. From the picture, I can see your machine is in very good shape because the original frosting is still there and not worn in some spots. If there is any wear at all, it is very slight. I assume that the quill is used for the "Z" axis and it is probably in good shape too.

If you can, try and find out why the microstepping is unreliable. Over any distance, that will make a difference in machine position because control is open loop and the program doesn't know if the machine hasn't stepped. Over the length of a program, this will give an accumulated error and throw off the positioning.

I always write my programs so the machine finishes back where it started and I call that "Home." I make sure that home matches a part feature to make things easy. If the dials don't read the same when the program completes, I know something has gone wrong. My setup microsteps reliably and that's why it will run accurately to tenths. Of course, the low friction of the ball bushings is a big factor in this reliability.

You should be able to figure out what is wrong because I'm pretty sure that when the machine was new, it did step reliably. Do your driver cards match the steppers and provide full torque, is there too much friction in the ways/ball screws, are the ball screws and nuts dirty, etc. Somewhere, something isn't right.

Make sure that the ball screw/nut combination is protected so that dirt, coolant etc. can't get at it. Cleanliness is important, so I would recommend way covers to protect against dirt and chips. With a CNC machine there will be lots of chips. You should see how the commercial machines throw them about.

Best of luck and I'm sure you will get it sorted out eventually. As I said before, at least the machine works and that's a great start. By the way, I don't claim to be an expert, what I'm saying is just what I've learned along the way.

Richard Trounce.
User avatar
ctwo
Posts: 2996
Joined: Tue Mar 27, 2012 12:37 pm
Location: Silly Cone Valley

Re: My first part - and a dimple

Post by ctwo »

Thanks RET. I am using the original Bridgeport Sigma stepper motors. The windings were not designed with such resolution, and magnetic fields are not powerful enough for reliable microstepping; although, microstepping does run the motors very smooth. I do not think there is an error accumulation, just that sub-thousand steps are not consistent. For example, if I half step through an inch, every half may land +/- 3 tenths or so, but each full step lands pretty much on the mark. I am using both Mach3 and Linux CNC with Gecko drivers, but modern electronics cannot completely compensate for the older stepper motors. I am running open loop.
Standards are so important that everyone must have their own...
To measure is to know - Lord Kelvin
Disclaimer: I'm just a guy with a few machines...
RET
Posts: 960
Joined: Wed Jun 07, 2006 8:36 am
Location: Toronto, Canada

Re: My first part - and a dimple

Post by RET »

Hi ctwo,

At first, from your description I thought that maybe the number of poles on the steppers were limiting the resolution and changing the steppers might make a difference. Instead, it could be the "digital resolution." The steppers and driver board I'm using (Sherline) microstep at eight steps per thousandth of an inch not ten, but that's close enough to give a resolution in tenths. Stepping in tenths might be possible, but instead of 3 bits (two to the third power is eight), it would take at least four bits, perhaps more.

Between playing with the steppers and the driver board (perhaps also the software), you should be able to have the setup microstep reliably at eight steps per thousandth like my setup does, so you get a resolution in tenths.

One way to check for errors would be to put a dial indicator on the table with the dial on the spindle so that when you run the x axis manually it moves the dial. That way you can tell the X axis to move 937 thousandths and see if the dial indicator matches.

Another possibility would be to write a little program that runs the table through a number of directions in both X and Y and then returns to the start (include Z as well if you want to). If the dials read the same at the finish as they do at the start, then you should be OK.

If you make all your programs end at the same X,Y and Z co-ordinates as they start the way I do (for me, that's always 0, 0, 0), that is also a good way to verify the accuracy of the system. If the dials don't read the same at the finish as they did at the start, you have a problem. Its also a good idea to record the dial readings at "Home" (program start) so if something goes wrong, a crash, -power failure etc., you can get back to the starting point manually.

Keep on playing with it and you'll get there.

Richard Trounce.
User avatar
ctwo
Posts: 2996
Joined: Tue Mar 27, 2012 12:37 pm
Location: Silly Cone Valley

Re: My first part - and a dimple

Post by ctwo »

The Brigeport BOSS3 (and several versions) have counters on the X and Y, and a simple dial on the Z. The counters are much like the old thumb counters, with gears and wheels inside, and driven directly from the screws. They are set up to show 0000 after homing to machine 0.

When I used a tenths indicator and step in 5 tenths increments, every half step would land somewhere between .3 and .8, and every full step was within a tenth, sometimes two. But that was over the short 8 thou range of the indicator I was using. I will try it again with a 2000 tenths indicator I have. I have not measured long travels to fine tune the steps per inch setting. But I am sure of the results that even half stepping will only be accurate to around a half tenth. That is just the expected performance of the 35 year-old step motors. Plus there is some play in the belts and ways - it all adds up.

I was considering whether to limit my moves to full tenths since the motors step much more reliably in that range. I'll have to test more. I'm thinking of starting at 0, then moving out an inch and back 0.0001, out and back in to 0.0002, etc... I'd plot the results along different segments of the screw too. Then do it again in full step.

I tried setting backlash by moving forward, and then back a thou. That was usually less than 5 tenths, but when I set backlash from 5 tenths or less, or enabling it all, I would get steps in my circles at the four quadrant transitions, so I've just turned that off. I think what I was seeing was more play in the ways and slack in the belts.

I have modest expectations for the machine. I think that if I can continue to run it at 100 ipm rapids and land on full steps within half a thou, that would be exceptional performance for what we have here. I think I would need scales and go closed loop to get better.
Standards are so important that everyone must have their own...
To measure is to know - Lord Kelvin
Disclaimer: I'm just a guy with a few machines...
RET
Posts: 960
Joined: Wed Jun 07, 2006 8:36 am
Location: Toronto, Canada

Re: My first part - and a dimple

Post by RET »

Hi ctwo,

If I were playing with your machine, one of the first things I would definitely try is to lock the table with the power off to the steppers and see what happens when I try to move the table with the handwheel. It would also be a good idea to have a dial indicator on the table to make sure it didn't move at all when you are trying this. Do this with all the axes to see what you've got.

If the ball screws are OK (adequately preloaded) there should be no motion possible in the handwheel. If you can feel anything at all, the ball nuts (and possibly the screws too) need to be attended to (cleaned, fixed or replaced). The timing belts between the steppers and the ball screws are easy to adjust to take out any slack.

If I had that machine, I would play with it and make it into a really good machine. If nothing is broken and the ways are good, you have 90% of what you need and changing stepper motors and electronics is relatively cheap and easy. If I can do it, you can too.

Hey, even as it is, it works! With a little tender loving care it can be even better.

Richard Trounce.
User avatar
ctwo
Posts: 2996
Joined: Tue Mar 27, 2012 12:37 pm
Location: Silly Cone Valley

Re: My first part - and a dimple

Post by ctwo »

The Boss CNC's did not come with any hand wheels or locks for the XYZ axis, just a crank and locks for the knee.

This picture represents the identical machine.
http://www.practicalmachinist.com/vb/at ... m_2424.jpg


The ball screws are zero backlash. According to the gib adjustment procedure, you would place an indicator on knee to saddle, tighten the gib until pushing/pulling the table results in no more than 5 tenths movement, less spring back. I.e., I pull the table and the indicator will show a thou, then spring back to 5 tenths.

I think a good servo system is a bit expensive, $2800 for 3-axis with AC drivers. That's cheap if you can make money from the machine.
Standards are so important that everyone must have their own...
To measure is to know - Lord Kelvin
Disclaimer: I'm just a guy with a few machines...
RET
Posts: 960
Joined: Wed Jun 07, 2006 8:36 am
Location: Toronto, Canada

Re: My first part - and a dimple

Post by RET »

Hi ctwo,

$2,800.00 isn't bad, I spent $2,500.00 for each axis (one machine tool quality ball screw & nut, four ball bushings, two Thomson shafts and two support rails per axis) but I'm happy with the results. I get that almost mirror finish because the cutter can't move any way but the way the program tells it to.

One thing you should try. Set up your tenths dial indicator on the table with it touching the spindle to check the "X" axis. Make sure the timing belts are tight. With the power on to the steppers so they can't move, try and push the table back and forth as hard as you can. There should be no motion at all on the dial indicator, or at least no more than a total of 5 tenths (the "slop" in the gib setup). If there is more than that, the ball screws need attending to or replacing. Do the same check for the "Y" axis.

I don't like what I see on your part. That's why I think something isn't right. You have a good machine which should handle 1/2" and 3/4" dia. end mills (and even 1") if you don't push the machine too hard.

I would keep on until it works the way it should.

Richard Trounce.
User avatar
ctwo
Posts: 2996
Joined: Tue Mar 27, 2012 12:37 pm
Location: Silly Cone Valley

Re: My first part - and a dimple

Post by ctwo »

Hi RET, If I push axially, there is no movement, but there will be movement if I push and pull on the end of the table, about a thou that springs back to 1/2 thou from starting point.

I think there may be several things that could be going into that dimple as has been mentioned. I am also now not sure about the cutter. It was bought used and I broke it since and have lost the tip to inspect it. I also recall hearing the part/cutter chattering and seeing chips vibrate on top of the part (I was cutting dry, and with a few inches sticking up above the vise jaws). I've also noticed my Kurt vise movable jaw makes a clink when I loosen it, like it is lifting and dropping back down against its ways. I bought that used too and have not disassembled it. The guy said something about resurfacing it, so I need to take a look at it. It looks good to the eyeball.

Anyway, you also saw my G-Code love part, and I know that was done with a half dull cutter. The inside pocket had some chip smearing after the half-way point and you can see tears on the bossed heart. I'll get the right formula soon. Right now, I am working on a "noise" problem giving false e-stop or limit triggers. They don't seem to happen until I am cutting a part...of course!
Standards are so important that everyone must have their own...
To measure is to know - Lord Kelvin
Disclaimer: I'm just a guy with a few machines...
RET
Posts: 960
Joined: Wed Jun 07, 2006 8:36 am
Location: Toronto, Canada

Re: My first part - and a dimple

Post by RET »

Hi ctwo,

Sounds as though the machine may be OK. I would suggest that you buy a new end mill.They aren't expensive, about $10 to $15 dollars each and then try your part again with less sticking out of the vise. No more than an inch, preferably less. Make sure the vise is tight. Try a 4 flute, center cutting end mill.

Also, try rewriting your code so that you use climb milling. Start off with a feed rate of 1" per minute and if it likes that you can increase it. Also start off with a depth of cut of .020." You should be able to get away with quite a bit more, but start off light.

That's what I do with a new part. I start off slow to see how it works and then increase rates and depths until it looks close. Remember, with your machine, it will be very easy to break a 1/8" end mill or to cause the work to move in the vise. CNC doesn't give you the "feel" you get from manual machines.

Richard Trounce
Post Reply