A machinist tale
Posted by
Blue
on 2005-07-15 16:08:39 UTC
Hi,
Some years ago I bought/stole for £600 an ex. technical college Harrison 280
CNC/Manual 11" lathe which was nearly new, although it dated back to the
1980's. The machine was in secondhand machinery warehouse belonging to a
friend of mine but he was acting as a middle-man for someone else. I didn't
want to test it where it was because, if I got it working, the sellar may
have changed his mind or put the price up. As it was, after I'd removed the
control box, the sellar (not my friend) tried it on by saying he'd quoted
the wrong price and he really wanted £850. My answer was to say a deal had
been struck and I wasn't interested at that price, but I'd put it back
together for him for £30/hr. and I'd make sure it took 40 hours to do it, he
caved in. I had to dismantle it to get it into my cellar but that was no
problem as all the wiring, and where it went, was labelled first. I'd also
obtained a manual from the manufacturer.
On putting it back together and testing the controls it was found the Z
stepper drive didn't function. This drive was made by Parker Controls who
are very big in machine controls for industry in this country. A phone call
to their technical department proved pointless, they wanted £250 just to
look at it and nearly £450 for a new one and refused to give me any help
even identifying some of the components. Well after all I bought the lathe
because it could be used as a manual or I could always rip out the
electronics and fit up to date bits using this list for help. Nothing
happened for a few months until out of interest I looked at it a little more
closely and noticed the side missing from a tantium bead capacitor next to a
5 volt (7805) voltage regulator. My multimeter proved there was a dead short
across the power rail and ground. I replace the regulator and the tantium
cap and it all worked. Total cost of the repair came to about £1 and the
bits came out of my over ordered components box. I'd drawn the offending
part of the circuit to help me work it out, so I sent it to a friend I
correspond with in Canada. Next to me this guy is a wizard with electronics.
Anyway the upshot was he said it was a bad design because two ceramic (10nF)
capcitors should have benn fitted across the input/output and ground of the
7805 and without them the power rails would tend to oscillate causing it to
let the smoke out. I fitted the caps to the underside of my board and also
to the X driver and they have worked for over two years. The devil in me
wouldn't let it go so I had to write to Parker Controls and tell them their
electronics design was crap and how they should be fixed, I didn't get a
reply and as far as I know they still don't retro-fit the drivers that are
sent in for repair. Nice work if you can get it!
The above is just background for the real story. I've been using the 280 on
a daily basis and have gotten quite good at programming the on-machine
computer via the keyboard. I didn't get the off-machine software and,
although it is 20 years obsolete, the software company still want silly
money for it. A few weeks ago I started to machine some ball-handles for a
Quorn Tool and Cutter Grinder, I promised to build one for a friend in the
States. I'd been looking forward to this job because I wanted to make
someone happy and it would be the first time I'd machined something using
circular interpolation and multiple tool changes. There are two sizes of
ball-handle and I made the smaller of the two first. After a few days trying
to machine proper handles and learning how to do the curved ends, I changed
the design and cut the machine time in half to 45 minutes each. On moving on
to the larger handle the problems started. I noticed the second tool, which
put on the end radius, was taking 0.026" off the end before it started the
radius and also the second radius, which forms part of the end ball, was
starting 0.026 short. This was puzzling as I was sure I set the offset
correctly. I cured the problem by adding to the offset the difference of
0.026".
The next problem then appeared with a flat between the first half of the end
ball and the second half which is machined by a parting-off tool as a third
tool change. It turn out the parting tool was losing steps as it
machined/waist the metal down to 0.125" to make room for chamfering prior to
machining and parting-off the end ball. I tried slowing down the parting cut
but it still lost steps. After thinking it through the only thing left to do
was to split the program into three programs and repositioning the tools. I
could rough out and finish machine most of the ball-handle with two tool
changes in a first program. Machine the waisting/part part-off in a second
and it doesn't matter if there are a few lost steps. The third program would
machine the back end of the end ball and finally part-off. Rather than write
all the new programs from scratch I used the edit function and copied the
bits I needed for the second and third before deleting line after line. It
was as I was doing this I noticed why the second tool needed an adjustment
to its offset. I had left a "0" off a Z G00 movement. You would not believe
how many times I'd looked at this line of code and not seen the missing
zero.
Well back to it............. Two weeks work to make 30 ball-handles just as
well I'm not trying to make a living from them.
Regards,
Terry
Some years ago I bought/stole for £600 an ex. technical college Harrison 280
CNC/Manual 11" lathe which was nearly new, although it dated back to the
1980's. The machine was in secondhand machinery warehouse belonging to a
friend of mine but he was acting as a middle-man for someone else. I didn't
want to test it where it was because, if I got it working, the sellar may
have changed his mind or put the price up. As it was, after I'd removed the
control box, the sellar (not my friend) tried it on by saying he'd quoted
the wrong price and he really wanted £850. My answer was to say a deal had
been struck and I wasn't interested at that price, but I'd put it back
together for him for £30/hr. and I'd make sure it took 40 hours to do it, he
caved in. I had to dismantle it to get it into my cellar but that was no
problem as all the wiring, and where it went, was labelled first. I'd also
obtained a manual from the manufacturer.
On putting it back together and testing the controls it was found the Z
stepper drive didn't function. This drive was made by Parker Controls who
are very big in machine controls for industry in this country. A phone call
to their technical department proved pointless, they wanted £250 just to
look at it and nearly £450 for a new one and refused to give me any help
even identifying some of the components. Well after all I bought the lathe
because it could be used as a manual or I could always rip out the
electronics and fit up to date bits using this list for help. Nothing
happened for a few months until out of interest I looked at it a little more
closely and noticed the side missing from a tantium bead capacitor next to a
5 volt (7805) voltage regulator. My multimeter proved there was a dead short
across the power rail and ground. I replace the regulator and the tantium
cap and it all worked. Total cost of the repair came to about £1 and the
bits came out of my over ordered components box. I'd drawn the offending
part of the circuit to help me work it out, so I sent it to a friend I
correspond with in Canada. Next to me this guy is a wizard with electronics.
Anyway the upshot was he said it was a bad design because two ceramic (10nF)
capcitors should have benn fitted across the input/output and ground of the
7805 and without them the power rails would tend to oscillate causing it to
let the smoke out. I fitted the caps to the underside of my board and also
to the X driver and they have worked for over two years. The devil in me
wouldn't let it go so I had to write to Parker Controls and tell them their
electronics design was crap and how they should be fixed, I didn't get a
reply and as far as I know they still don't retro-fit the drivers that are
sent in for repair. Nice work if you can get it!
The above is just background for the real story. I've been using the 280 on
a daily basis and have gotten quite good at programming the on-machine
computer via the keyboard. I didn't get the off-machine software and,
although it is 20 years obsolete, the software company still want silly
money for it. A few weeks ago I started to machine some ball-handles for a
Quorn Tool and Cutter Grinder, I promised to build one for a friend in the
States. I'd been looking forward to this job because I wanted to make
someone happy and it would be the first time I'd machined something using
circular interpolation and multiple tool changes. There are two sizes of
ball-handle and I made the smaller of the two first. After a few days trying
to machine proper handles and learning how to do the curved ends, I changed
the design and cut the machine time in half to 45 minutes each. On moving on
to the larger handle the problems started. I noticed the second tool, which
put on the end radius, was taking 0.026" off the end before it started the
radius and also the second radius, which forms part of the end ball, was
starting 0.026 short. This was puzzling as I was sure I set the offset
correctly. I cured the problem by adding to the offset the difference of
0.026".
The next problem then appeared with a flat between the first half of the end
ball and the second half which is machined by a parting-off tool as a third
tool change. It turn out the parting tool was losing steps as it
machined/waist the metal down to 0.125" to make room for chamfering prior to
machining and parting-off the end ball. I tried slowing down the parting cut
but it still lost steps. After thinking it through the only thing left to do
was to split the program into three programs and repositioning the tools. I
could rough out and finish machine most of the ball-handle with two tool
changes in a first program. Machine the waisting/part part-off in a second
and it doesn't matter if there are a few lost steps. The third program would
machine the back end of the end ball and finally part-off. Rather than write
all the new programs from scratch I used the edit function and copied the
bits I needed for the second and third before deleting line after line. It
was as I was doing this I noticed why the second tool needed an adjustment
to its offset. I had left a "0" off a Z G00 movement. You would not believe
how many times I'd looked at this line of code and not seen the missing
zero.
Well back to it............. Two weeks work to make 30 ball-handles just as
well I'm not trying to make a living from them.
Regards,
Terry