|
Before looking for servo
drives I need to know what specification the drive had
to live up to regarding voltage and current. The
datasheet for the motors specifies the voltage constant
as 0.47Vs/rad but what does that mean... Well, basically
it means that it takes 0.47V for the motor to turn one
radian in one second.
1 radian is roughly 57.3°
so 57.3°/0.47 = 121.9°/V/s or expressed even easier
20.32rpm/V (121.9 * 60 / 360 = 20.32). The datasheet
furter states that the maximum useful speed is 2000rpm
so in order to reach that I'd need a supply voltage of
2000/20.32 = 98V.
Electrically the motor
armature can be thought of as a resistor, an inductor
and the actual armature in series. The resistor is
representing the resistnace in the wire that the
armature winding is made of and the inductor the
inductance in the winding. To reach 2000rpm the
armature needs to "see" 98V, not the connectors on
the motor. When current is pushed thru the motor
armature there will be a voltage drop across the
resistance, how large this drop is depends on the
armature resistnace and the amount of current.
My X and Y axis motors
have an armature resistance of 0.24 ohm and the maximum
continous current is 19A, that means that when we push
19A thru the windind we drop 4.6V across the resistance,
this means that to be sure we can reach 2000rpm even at
full continous load we need an aditional 4.6V on top of
the previously calculated 98V. Now on these particular
motors this is pretty marginal and can almost be
neglected but on smaller motors the armature resistance
can easily be 1ohm or more.
Next thing to account for
is the maximum PWM dutycycle of the servodrive. A modern
servo drive, using PWM technology, usually can't
modulate to 100%. This is because the internal bootstrap
capacitors for the highside MOSFET drivers need to be
recharged. What this means is that if the drives maximum
PWM dutycycle is 90% the maximum voltage that the motor
will see wil be 90% of what you feed the drive plus,
depending on drive design, you might drop another volt
or two across the actual switching elements and current
sense resisotor in the drive.
Different drives has
different maximum PWM dutycycle but most seems to be in
the 85-95% span so my choise was to err on the safe side
and went for 85% for my calculations.
So to be sure I could
reach 2000rpm I need a supply voltage of 98V + 15% +
4.6V = 117V
First drive that springs
to mind is the G320 from
Geckodrive,
I've had great success with their step-motordrive so I
wouldn't hesitiate but the 80V maximum voltage was way
to low. Geckodrive doesn't specify the maximum PWM
dutycycle but even if it's 95% I would only be able to
reach 1550rpm no load - so the G320 was a no-go.
Next I looked at (and
actually got one to try) the Mammut from
CNCDrive. It's
rated at 180V and 40A peak so it looked like a perfect
match - unfortunately I had severe problems with their
tuning software so I simply couldn't make the drive work.
Support emails suggested I had malware in my computers
which of course was a possibillity but not very likely.
Lately it was discovered to be a compatibility issue
between different versions of the .net framework but at
that point I had already moved on. (I'd like to point
out though that I have since tried some of CNCdrives
newer product and have had zero problems hard- or
software wise this far.)
I also looked at the
Viper200 from
Larken. This looked like a nice drive as well but at
that time they claimed to have features IN the drive
which wasn't actually available yet (feed forward
control). As far as I can see feedforward is now
availble and I've read a couple of success stories from
people using them so I'd really like to try them out
some time!
Another option obviously
is Rutex, but due to very bad reputation regarding
support and product specifiactions I didn't dare to go
there - YMMV.
Then I found the Finish
company Granite
Devices and their VSD-A drive. This is drive speced
at 200V and 10A continously, 15A peak for 0.5seconds.
10A was a little on the low side for my motors, rated
19A and 24A continously but 10A would still give me
4.7Nm of torque - continously and with the 2:1 belt
reduction that's 9.4Nm at the screw. I actually bought
three of these drives (now discontinued in favour of the
VSD-E and VSD-XE) and let me tell you, these are some
really nice drives. True torque control, differential
encoder inteface capable of tracking encoder speeds up
in the MHz range, pulsemultiplier setable to any
arbitrary value, input smoothing filter, etc etc. I
really liked these drives, unfortunately it turned out
that although 10A continous proably would have been
enough to run the machine I had trouble getting the
motors to accelerate in at an acceptable rate even with
the motors unloaded. The 15A peak of VSD-A simply wasn't
enough.
So, what I finally ended
up using was the HP-UHU drive. This is a kit based drive
that started as project on the CNCZone. I decided to get
three kits even though I was quite sceptical at the
beginning but I thought that if I could make them work I
could always fall back on the nice VSD-A's, living with
the limited acceleration. Although I did have some
problems once that was sorted the HP-UHU has actually
been working quite nice, so far no I haven't done any
machining with them but I have run them hours and hours
with out problems.
More about the HP-UHU
drive on next page.
<Back>
<Next>
|
