Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: rpseguin on May 23, 2018, 03:24:27 AM
-
https://photos.app.goo.gl/uy1vg5d0iz33WUFn1
I just picked up a Sharp CNC knee mill (Sharp HCV) with a Mitutoyo Millstar CNC control with Mitutoyo linear scales on X, Y and on the quill. I don’t know what generation of Millstar this is. 1, 2 or 3? Anybody? Anybody have manuals and documentation for the control?
The Millstar controller is an old Pentium S PC with a Galil DMC-1020 motion card, a whopping 4 megabytes of RAM in a case with monitor, does power on, but there’s a problem with the IDE drive failing to boot. Bad IDE hard drive would be my guess.
The PC has PCI (but I think no PCIe) slots.
It has Glentek brushed DC servo drives for X and Y axes and I'd like to keep/use them if I can.
I'd also like to keep/use the Mitutoyo linear scales.
Recommendations on hardware to get to adapt/use the existing Glentek drives and Mitutoyo scales with Mach 3?
Also, they replaced the head with a 3HP Gromax variable speed head. I don’t know anything about Gromax. Anybody have manuals they can share?
My kids have been helping me clean off some of the grit and dirt and rust with Scotchbrite and WD-40. The Y axis definitely shows some wear and a bit of scoring on the ways, but the overall machine is pretty solid.
Anyone have manuals for the Sharp mill?
Docs for Mitutoyo Millstar?
Docs for Gromax 3HP variable speed head?
Recommendations for way oil? Spindle oil? ...
Anybody have a Millstar hard drive that I could clone?
Is it worth it to try to resurrect the Millstar controller?
Or, would it be better/faster/easier/cheaper to try to convert it to Mach 3 CNC?
How to use the existing Glentek brushed DC servo drives with Mach 3? I'm guessing they take 0-10V input, but not sure.
How to use the existing Mitutoyo linear scales with Mach 3 CNC?
-
Hi,
Mach3 and in fact Mach4 are native step/direction controllers. They can with suitable motion cards be used to drive analogue servo drives. The motion card usually completes the loop. I would
suspect that is what the Galil card does. Galil is a very good brand but expensive, if you can reuse it by all means do, I suspect you wont buy a new one no matter how good they are.
Another possibility is to use the step/direction output from Mach/controller to step/direction input DC servo drives with your existing servos. Gecko 320s have as good rep as do CNCdrives servo drives.
DC servos are yesterdays technology, if you have to start to replace components, particularly if you have to mix and match between manufacturers it can become expensive really quick. The combination
becomes vunerable to obsolescence.You might consider AC servos and matching drives. Delta is a trusted Twainese brand and about $500 for a 750W AC servo and drive which will outperform any DC servo of the
same power. The resolution and control solutions of a modern AC servo so far exceed anything that was ever even attempted with DC servos and AC servos are now cheaper as well!
I would suspect that the Mitutoyo scales are separate to the CNC control. Can you provide some pcs and model numbers? If they are separate or could be made so that would be easy. Some people like the idea
of trying to incorporate them into a closed loop control setup. I have no doubt its possible but if you can do it I suspect JPL and /or NASA will be wanting to hire you. If that's the case play with their mill! LOL
Don't know anything about the spindle motor. Most probably an asynchronous AC motor. May I suggest work out how to control the axes first...see whats possible and within your budget and THEN start getting
serious with the spindle motor.
Craig
-
The pictures in the link I posted show the scale (Mitutoyo models AT-102 and AT-111) and you can see one of the brushed DC servo motors.
First, I'd like to see if I can resurrect the Mitutoyo Millstar CNC control. The computer appears to work, but since I can't get it to boot, I can't really check it out. I've never used a Millstar, so I don't know what the GUI is like or how easy/hard it is to use.
However, I do want something reliable and good.
Brushless (AC) servo sounds nice, but that'll run a lot of money unless I can score a good deal.
I thought about using Geckodrives.
I'm just looking for manual/docs for things for completeness. I don't think there's anything wrong with the Gromax head or anything.
-
Hi,
by far and away the cheapest alternative is to use the Galil and existing servos and drives.
If you find you need to replace the drives you are in the territory where replacement with a DC drive like
a Gecko looks cheapest but you are adding a layer of complexity which may mean that the apparently
more expensive option, AC servos, actually works out cheaper, they are a buy once and use for decades
type solution.
If you wish to resurrect the existing controller you will need Mitutoyo support, and what's the bet they say it obsolete.
It becomes not a project in Mach or even CNC but a vintage electronics restoration project.
My guess is that the linear scales are free standing.
Craig
-
I think the first step in resurrecting the Mitutoyo Millstar CNC control is for me to clone someone else’s disk, since this one doesn’t look it it wants to boot.
When I look at the BIOS screen, it doesn’t show the hard drive, even though I hear it spin up.
Although I’m not sure how many different motion control boards, sero drives, encoders, ... they used or if they stuck with the Galil DMC 1020/10x0 series, Glentek, ...
Then I’ll need tuning and other parameters, so maybe it will be a lot of work to try to get an old control working.
-
Hi,
this forum is dedicated to Mach CNC software. It seems unlikely then that this forum will provide much information about a Mitutoyo control.
CNCZone has a wide readership on a very diverse range of equipment and may produce results for you.
Have you emailed Mitutoyo?
Craig
-
Yes, I’ve already emailed them.
I’m also planning on the strong possibility that I won’t be able to get the Mitutoyo control going, in which case I’ll be trying to retrofit it with Mach 3.
-
Hi,
if your thinking that a Mach retrofit is the way to go may I suggest getting as much information about the Galil board you have, my guess is that you will be able to interface Mach
to the Galil and the Galil will close the control loop of your servo amps and servos.
Galil are still very much in business and would also guess that theres plenty of info about your board.
Craig
-
Hi,
saw this:
Controller Introduced Estimated End-of-Life New Equivalent
DMC-1010 - DMC-1040 October 1992 1 to 2 years DMC-18x6
So despite the Galils age its still current just! If I read the model interpretation correctly it sounds like a
two axis controller.
Craig
-
Hi,
DMC-1020.........................2-axis ISA......
You'll need ISA slots...starting to look pretty dated and you still need another axis. A two axis controller
would give you 2.5D capability at best.
Would you post the details off your existing servos, preferably either a power rating or torque/speed, something,
anything to work out what power they are. Looking at the dated controller I'm thinking you should look really closely
at what you can do for a given sum in the AC servo field. DMM for instance do 200 and 400W units for $200-300.
Mach3 is $175 or Mach4 for $200. A Ethernet SmoothStepper (my choice) is $185 running on a Latte Panda
board, $150 including Windows 10 Enterprise. That combination will kick ANUS! Whats more it will do so for years
and years to come and have the capability to grow and expand as you become more demanding.
Craig
-
Hi,
https://www.lattepanda.com/products/3.html (https://www.lattepanda.com/products/3.html)
Craig
-
Hi,
I have a couple of corrcections.
The LattePanda 4/64G is $149 without Windows activation, $209 with activation.
DMM 400W 60V servo $147
DMM DYN2 drive to suit $138 for a total of $285
for the high voltage setup, ie 230V input, no power supplies etc
DMM 400W 200V servo $177
DMM DYN4 200V drive $218 for a total of $395
You'd also need a couple of BoBs so posibly as much as another $100 for good ones.
Craig
-
Yeah.
Once I looked up the Galil DMC-1020, it was pretty clear that it’s really old and only 2 axis.
Here’s a better pic of one of the Glentek brushed DC servos:
https://photos.app.goo.gl/YVFRGjpOjOThiQvH3
Model: GM4020-39-02006301-6
Torque: 23 In-Lb/368 oz-in (not huge for such large motors)
Resistance: 1.3 ohms
V/1000RPM: 7
Max RPM: 3500
-
Hi,
rough calculations:
23/12= 1.92 ft.lb approx 2.2Nm
3500/60 =58 revs/sec
58 X 2 x 3.141 = 364 radians/sec
POWER= torque x speed
= 2.2 X 364
= 801W
Hey, those servos have got some serious grunt.... you'll want 750W units to match them, a bit
more than I allowed for. You could run 400W units but may not have quite the sting as the
current configuration. Plenty of blokes dream of having 400W servos on their axes...750W will
really make that mill sing...probably well more than is required for hobby purposes.
Craig
-
Thanks!
I just realized that i may have 2 or 3 G320 Geckodrives
-
Hi,
those servos are going to suck current like no ones business, I came to 32A at rated power.
I think you are buying into trouble....ditch the DC servos and get the current thing, they will last the life of the
machine. Additionally they have control and reporting modes that DC servos have never had.
I would go for 400W 200V DMM servos and matching drives...$1200 and you're out of there! No shagging
around with old s*********t!
Craig
-
Ok. Thanks!
I assume those drives are digital with step/direction control capabilities.
Have any opinions on pulse generators or how to interface with more modern computers since parallel ports are old and not been on PCs for some time? Or just try to buy an okd PC with a parallel port?
I wonder if I could sell off the existing components to offset the cost of the new components.
My inclination is to leave the Mitutoyo linear scales on the machine and either get a DRO or find a way to display it on same PC so I can have a DRO for manual work.
I also would like to retrofit the quill with a motor/drive to make it a 3rd axis.
Thanks!
-
Hi,
I like the Ethernet SmoothStepper (ESS) by Warp9. Its not the only choice but it is a good one amongst other good ones!
It will give you up to six co-ordinated axes....enough?
One, or if you want some more inputs, two BoBs, the old parallel port types plugged into the ESS and your set to go.
They can pulse at 4Mhz, 160 times faster than a parallel port and it allows you to use a late model PC like the LattePanda
I linked to.
I think the Mitutoyo scales and counter unit can be run as a separate system. A little bit of electronic trickery might be
required to get the scale to talk to Mach but is possible.
I personally think Mach4 is appropriate for new projects. Mach3 afficionados will disagree (i used to be one myself) but its so much
more flexible and stable.
-
Hi,
once you get away from the parallel port and use an external motion controller like the ESS or a UC300 then
you can use just about ANY PC you like....you know the one that the kids have left in the dust because its
three years old....just perfect.
It does have to be a PC, that is a Windows operating system. ARM processors wont do it and at this stage
Linux is not a go either....although according to smurph it may happen yet!
The little LattePanda I linked to is about 4 inches square, has an Arduino built in and Windows 10 Enterprise
pre-installed and activated for $209.
Craig
-
Is there an upgrade discount for people who bought Mach a long time ago?
Breakout board recommendations?
I don’t see a need for tons of inputs (famous last words?) on this mill.
The servo drive box has a pretty beefy looking transformer in it to step the voltage down (from 120V) for the existing Glentek servo drives.
-
That ZlattePanda looks pretty cool!
-
Hi,
it does at that! I've got a dual core Atom MiniITX board so not as kool (or powerful) as the LattePanda
but it does the business on my machine good-oh despite being such a low powered PC.
If you want a bit more grunt look at the UDOO series, they allow you to fit your own SSD or hard drive.
https://www.lattepanda.com/products/3.html (https://www.lattepanda.com/products/3.html)
Craig
-
Hi,
sorry...wrong link:
https://www.udoo.org/udoo-x86/ (https://www.udoo.org/udoo-x86/)
Craig
-
Hi,
I've been doing some more calculations and have come to the conclusion that 800W is more likely peak power rather than continuous.
My reasoning goes like this:
The voltage feedback constant is 7V/1000 rpm.
The back EMF at rated speed is 7/1000 x 3500 =24.5V
The back EMF is in effect the electrical reaction to the motor delivering rated power to a mechanical load, therefore the current is:
power = voltage x current
800W =24.5V x current
current= 32.6A
The total terminal voltage when producing 800W into a mechanical load will be the sum of the back EMF and the resistive loss of the armature.
Vtot = 24.5 + (32.6 X 1.3Ohm)
=66.9V
The total electrical power delivered to the servo is:
Ptot=66.9 x 32.6
=2180 W
Of which 800W is deliverd to the load so the remaining power is dissipated in the resistance of the armature:
Pwaste=2180-800
=1381W
Your servo would cook up BIGTIME if this were continuous. My guess is that the torque figure quoted is peak torque not continuous rated torque.
As peak torque is typically three to four times rated I would guess the true CONTINOUS RATED torque is 0.5 ft.lb and rated power is about 200W.
That would bring the dissipation in the servo down to 79W. That is a lot more realistic.
If my reasoning is correct then 200W servos could be used to good advantage on your machine and that may bring the price down somewhat.
I have heard CNCers say that very large and heavy and fast machines might have axis motors of 1 hp or 750W. It seems more likely that a smaller machine
like yours would have axis motors somewhat less than that, 200W is an estimate.
The power figures I quoted earlier from the DMM website are continuous power. It makes sense to compare your existing servos continuous rating
to the continuous rating of DMM's AC servos.
If you follow the proposal to replace the DC servos with AC servos then the biggest expense in the retrofit will be the servos and drives themselves. As you have seen
the DMM 400W 200V servos and drives are about $400 a set, or $1200 for the three axes. A 200W 200V servo saves you only $24 per axis, it makes sense then to get
the 400W units.
You may note that DMM list both 60V and 200V servos and drives in these power ranges. The 60V servos and drives are cheaper however you would then require high current
power supplies. The 200V units however are fed with 230V AC line voltage, no power supplies required. All up its a wash, they are both near identical in price given the power
supply. The 200V units require much less current and therefore are going to run cooler at elevated loads so the higher voltage units are preferred.
This proposal would cost you $1200 for the servos and drives and about $695 for Mach4, an ESS, two BoBs and a Lattepanda. This assumes you have a monitor and keyboard that you
could re-purpose. Not far off $2000.
Is this within your means? What would your machine be worth once its done?
Craig
-
Hi,
I see the DMM 400W 200V servos are $177 whereas the DHT 750W servo is only $194, an extra $17....
Craig
-
Wow!!
That’s a good deal!
Also, what are the requirements on interfacing an MPG?
Any good, I expensive pendants or MPGs to recommend?
-
Hi,
I was looking at the pics you posted and I think the data you posted about the existing servos is wrong.
V/1000RPM: 7
In the pic it shows the BEMF/1000RPM as 39V
Which is correct?
I found this:
https://www.motortong.com/tseries/24248/GM4020.html (https://www.motortong.com/tseries/24248/GM4020.html)
which suggests 955W output. If that is the case you'd either have to consider gear reduction to use 400W servos or use 750W servos. Either way this
is likely to alter the cost of conversion.
Craig
-
I don't know. I guess I would say go with the plate on the motor?
I'll try to get a picture of the other motor. Hopefully it is readable.
-
Hi,
the spec, BEMF/1000RPM=39V, seems pretty plain. This is known as the Back Electro Motive Force constant, hence the acronym BEMF.
If the motor were driven by another motor at 1000rpm there would be 39V at the terminals. This is an important figure of merit for a servo.
Doing the calculation with a figure of 7V per 1000 rpm, which I suspect is in fact the tacho feedback device, then the shaft power of 800W could only be sustained
for a short period of time. With a figure of 39V per 1000 rpm that output could be sustained continuously. This is a more realistic and more probable than my previous
reasoning. I think it wise to consider the existing servos to be in the region of 800W continuous.
That has implications for your rebuild. You would ideally replace them with something of the same power but that is going to get expensive quickly. If you were doing
this for commercial purposes it would still be pretty cheap but if you have to convince your wife that you really really need it she might not believe you! You could use smaller
servos and have either a gear or belt reduction, that would increase the torque to that necessary to overcome friction and inertia but be a little slower. That would probably be
a good compromise for a hobby machine.
As I've pointed out the medium inertia 750W DMM servos and drives are about $412 an axis. They could be relied upon to be a direct replacement for your existing servos.
You could save a bit by getting 400W servo and drives, $385 per axis. You'd probably need a belt reduction of 2;1 to match your existing servos, but that would be very doable.
If you choose to retain the two X and Y servos there are some choices to make. Given that the Gailil is only two axis its no good as a controller. If you chose a HiCon board from
Vital Systems AND paid the premium to get the analogue features enabled the you could drive your two servos and still have the capacity to add another or even more than
one for the Z axis. A HiCon with activations is about $1200-$1300. Note this would allow you to use the two servos AND your original drives.
If you replaced the drives with Geckos or something similar then you could use a much cheaper controller like an ESS or a UC300. Either controller would allow you to feed the two
Geckos with step/direction pulses and still have plenty of capacity left over for the Z axis. You'd still need a servo and drive for the Z axis, say $400 and two Geckos say another $300
for the pair, for about $700 total.
I think you'd be very impressed with AC servos but despite them becoming ever more affordable they are still somewhat of a premium. If you were to try to buy NEW DC
servos, then I suspect they would be even more expensive.
I was given a 100W DC servo off a customers plasma table, this table is $150,000 or more and the whole business of thirty or more people rely on its output to stay
working. It has a faulty line driver in the encoder. I can probably jury-rig a new line driver, this servo is now only a spare. But no, we want a new encoder, I found the
OEM and priced getting a new encoder made to suit, $585USD, prepaid, one month leadtime plus freight to New Zealand, and we want a markup....Funny that they didn't
seem that keen once the numbers were put in front of them. I could get TWO NEW AC servos and matching drives with more power than they could get one encoder!
Even more, the business is about 4 doors down from the local distributor of Omron/Yaskawa. I could walk in get what I need off the shelf and Yaskawa is serious good quality
but they cant get their head around AC servos......they insist that servos have to be DC, how else can they be closed loop they ask. I've given up trying to explain to them how they
work....When the table craps out we'll see how amenable to alternate solutions when they're out of action!
Sorry about my rant. The point is that DC servos are rapidly going obsolete.
Craig