Brother TC-211

I found milling on my machine has to be extremely light. I actually had the table move pit of the way on a.030 side cuti cut .250 deep in aluminum with a .375 carbide endmill. I think the table is set up like a Jig bore. Precise positioning but very little side loading support

JohnW wrote: What sort of angular positioning accuracy are the original (Toshiba?) resolvers likely to achieve?

Resolvers are analogue, so given a long enough sampling period you can read them to infinite precision.

But, in practice, you tend to get something along the lines of what Mesa quote for the 7i49
"The 7I49 has approximately 14 bits of resolution and 12 bits of absolute accuracy."

So, call it equivalent to a 4096 count encoder.

JohnW wrote: I still have the motors. They need a good clean but turn smoothly (ie bearings are OK at least). X and Y are 100w 3.25kg/cm servo motors, Z is 200w 6.5kg/cm according to the original sales brochures for the machine. One thing I have noticed is that the ball spindle pitch on the machine seems more coarse that I would have expected.

that is to achieve the desired feedrates without incurring whip of the ballscrews.

It all looks standard, and I wondered how they manage to meet the original precision spec of 0.01mm. What sort of angular positioning accuracy are the original (Toshiba?) resolvers likely to achieve?

Well, my resolver converters get 4096 quadrature counts/rev, but they might have had a higher res converter originally.
Fanuc uses 1 million count/rev encoders on some machines for this reason. .01mm is not really all that fine, it works out to .0003937"

Yes, the Z axis is quite heavy! It has the original a 1.5KW servo spindle motor on it. I was thinking of a counter weight if I converted the machine to manual only.

With regard to side loads, I was more concerned with the actual vertical ways not being rigid enough for milling, or getting damaged from side loads. From pictures on the web the spindle tube appears about 6 or 7 inches long, with 4 ball bearing races, two at each end... The spec says the spindle motor is a 1.5KW Servo motor. Is it a servo because for the tapping operations, you need to synchronise the exact number of rotations with plunge depth for thread cutting?

At least with LinuxCNC, the Z axis syncs to the spindle rotation, so it doesn't need a precise spindle control. But, there is a video of one of these machines doing tapping, and the spindle is reversing REALLY rapidly to do a peck tapping cycle.
They may also need spindle orient to transfer tools from the carousel.

Jon

mblaszkiewicz wrote: I found milling on my machine has to be extremely light. I actually had the table move pit of the way on a.030 side cuti cut .250 deep in aluminum with a .375 carbide endmill. I think the table is set up like a Jig bore. Precise positioning but very little side loading support

Was this because the side forces from milling were rotating the servo motors? Adding a brake is no good if you want to mill a circle or curve, with X and Y smoothly changing. I wonder was that some sort of motor drive issue that could be corrected? Or is it these Toshiba servo motors by nature of their design (assuming you have the standard motors)?

No it was actually the table lifting and moving. The machine was designed as a high speed drill tap. All forces down.

mblaszkiewicz wrote: No it was actually the table lifting and moving. The machine was designed as a high speed drill tap. All forces down.

Very useful to know! So it is not going to be my general purpose milling machine! My revised plan now then, to just get going, is to use a Closed Loop 12nm NEMA 34 Stepper for the Z, and two open loop NEMA 23 Stepper motors I have already to move X and Y, and a parallel card. This should get me in to Linux CNC, at least engraving perspex, hopefully without getting crushed by the Z axis.

Is there any way I can get the original spindle motor, a Toshiba RA10M AC Servo motor, 1.5KW, 3K rpm 6K max running off 3 phase without a VFD? Can I just wire it up to the 3 phases and run it open loop at least for a short time, say 10 minutes? Will it be going at 6K then?