Fanuc Robodrill AlphaT10C -- Retrofit -- Help

andypugh wrote:

prashant wrote:

Here is what i need.
1) Servo motors ( as per your suggestions) X and y axis stall torque is 3 N-m, And Z is 6 N-m along with brake. The Z axis does not have a counter weight. (Max rapid required is 300in/m) (Screw Pitch is 5mm)

The machine already has very good servo motors and encoders installed. I would try very hard to keep them if I was you.

2) Spindle Motor direct coupled 4000 Rpm should be fine. ( current motor is servo 8000rpm, 7KW)
3) Drives for the servo.

And again, these are high quality parts, I would try to keep them unless they are definitely broken.

The first step is to see what control signals the existing drives give, and what feedback is from the motors.

It _might_ be cost-effective to sell the original Fanuc parts and replace with something else. They appear to be quite valuable as second-hand spares.

The problem is everything Fanuc touches is proprietary, so I think you will have problems
dropping somebody else's motor in their place. I have converters that make these encoders
look just like any standard brushless motor. it is also possible the motors can be used
without the commutation signals, either with BLDC in sensorless mode or with one of the
commercial sensorless drives. The motor's encoders provide normal quadrature
encoder signals.

But you need to check that it takes + / - 10V control signals if you want to stay with known configs.

Fanuc drives do NOT take +/- 10V signals, and maybe never did, ever. The earliest ones got
3-phase phase control pulses synched to the 3-phase mains, later ones sent PWM to the transistors.

Jon

Andypugh wrote:

It _might_ be cost-effective to sell the original Fanuc parts and replace with something else. They appear to be quite valuable as second-hand spares.

Could you please guide me -- where i can sell these fanuc motors and drives and what price should i quote for them. just a tentative price. I dont think it will be feasiable to sell the motors outside India. Is'nt it?

today i enquired about servo in India. Smart Servo with drive Tiwan Make is available 1.5KW-- 7N-m with brake 926$ and 1Kw--4.8N-m 618$. i have contacted some more suppliers here. lets see how things work out.

In the meen time i will post the encoder details of current motors. lets keep all options open.

Jmelson Wrote:

I have converters that make these encoders look just like any standard brushless motor. it is also possible the motors can be used without the commutation signals, either with BLDC in sensorless mode or with one of the
commercial sensorless drives. The motor's encoders provide normal quadrature encoder signals.

we will have to try the convertor to see if it works with these encoders. Unfortunately the machine is in india. i cannot buy parts which i am not sure about. it will keep on bulding the cost.

I am in for your drives though. In fact, i dont mind buying the Complete PWM Servo System for large machines. but i really dont know wether it will be cost effective because of the logistics --- import duty and stuff.

Jmelson Wrote:

I have converters that make these encoders look just like any standard brushless motor. it is also possible the motors can be used without the commutation signals, either with BLDC in sensorless mode or with one of the
commercial sensorless drives. The motor's encoders provide normal quadrature encoder signals.

I've been asking for some time to get the encoder part number from these motors. The photos you
posted are not completely readable, and don't show the encoder at all. I'm not sure what I wrote above
is quite correct, I had thought these were the older red cap encoders, but they might be serial
encoders. We have a converter for those, also, but the serial encoders do not produce any
signals that are usable directly by other controls/drives, without our converter.

Jon

Jmelson Wrote:

I've been asking for some time to get the encoder part number from these motors

Please find picture of z axis encoder attached.

Here's what is reads

Pulse coder Alpha A64
Type A860-0360-T201 / E
No. SS- 127427 Pat pending
made in Japan

prashant wrote:

Jmelson Wrote:

I've been asking for some time to get the encoder part number from these motors

Please find picture of z axis encoder attached.

Here's what is reads

Pulse coder Alpha A64
Type A860-0360-T201 / E
No. SS- 127427 Pat pending
made in Japan

OK, great. This is a serial encoder, and is compatible with our new serial Fanuc converter.
The A means it has absolute commutation data, and no backup battery is needed.
Our converter uses the absolute data at power-on, and then counts from the high-res
data after passing the index position the first time. the converter produces
differential AB and Index, plus the commutation or "Hall" signals. The only difference
is your encoders have the rectangular connectors instead of the DB-15 style like
the ones I have been using. But, I have the correct connector pinout chart
for those connectors. This encoder provides 65536 counts/rev. the encoder
converter reads the position 10,000 times a second, and converts to quadrature.
One problem with this is if the motor is running at 3000 RPM, that is 50 revs/second.
50 * 65536 = 3.28 million counts/second. You need to be sure whatever reads
the encoder on your CNC motion system can keep up with that count rate.

Jon

Jmelson Wrote:

This encoder provides 65536 counts/rev. the encoder
converter reads the position 10,000 times a second, and converts to quadrature.
One problem with this is if the motor is running at 3000 RPM, that is 50 revs/second.
50 * 65536 = 3.28 million counts/second. You need to be sure whatever reads
the encoder on your CNC motion system can keep up with that count rate

Does it mean that the converter will be able cope up only if the motor is running at 9.15 RPM or less.
I really dont mind limiting the speet to 1000RPM. For me, accuracy is more important than speed.

Can we change the encoder? Does the motor has some kind of Intelligence to recognise the encoder?

Or is it possible to make the converter to read at a higher speed?

prashant wrote:

Jmelson Wrote:

This encoder provides 65536 counts/rev. the encoder
converter reads the position 10,000 times a second, and converts to quadrature.
One problem with this is if the motor is running at 3000 RPM, that is 50 revs/second.
50 * 65536 = 3.28 million counts/second. You need to be sure whatever reads
the encoder on your CNC motion system can keep up with that count rate

Does it mean that the converter will be able cope up only if the motor is running at 9.15 RPM or less.
I really dont mind limiting the speet to 1000RPM. For me, accuracy is more important than speed.

Can we change the encoder? Does the motor has some kind of Intelligence to recognise the encoder?

Or is it possible to make the converter to read at a higher speed?

The converter will handle this just fine, it is whatever is reading the converter OUTPUT
that needs to keep up with the converter. What will you be using to interface to
the PC? Mesa hardware? No problem, their boards should be able to count the
encoder pulses just fine. The later version of the Pico Systems interfaces also can
handle this count rate. (Our earlier units could not count faster than 1 million
counts/second.)

If you don't mind machining, then you could change the encoder, but you will lose
a LOT of resolution. Typical encoders provide 4000 - 8000 counts/rev, and you
would have to align a 6-channel encoder to the motor's poles to get proper
commutation. I can't see any reason to replace one of the most advanced
encoders ever made.

Jon

andypugh wrote:

prashant wrote:

Does it mean that the converter will be able cope up only if the motor is running at 9.15 RPM or less.

No, it means that you need to use something faster than a PC parallel port to count the pulses that the Pico encoder converter sends out.
Something like the Pico servo controller seems ideal...

Right, the converter's lowest rate of quadrature output is 500 K counts/second.
So, you can't count that with software. It would not make sense to try to
count a 64K count/rev encoder at low rates, the PID would get
so far behind the actual position.

Jon