And that is why I am at my wits end with it at the moment. I think the drives are too cold & unresponsive. The Heidenhain control keeps going on strike because the slideway movement does not corelate to the expected movement by some miniscule amount. Thermal one way or another I reckon.

Hi,

This machine has feedback from rotary encoders on the motor shaft which is backed up by linear scales =Closed loop in my understanding.

Note quite. All servos have a rotary encoder so that the servo drive, or PID controller, knows the servos angular position and can compare that to the commanded
angular position, and drive the servo to close the error between the two.  But we don't specify angular position, we specify '6.342mm from there' or '12.687mm absolute'.
In an ideal world you could convert any linear measurement into and angular displacement of a ballscrew. This relies on the ballscrew being lineraly accurate and no
backlash. Much expense is encountered to ensure that is the case, but inevitably there will be some variance between angular position and linear position, and that variance
will be the basic accuracy of the machine.

If however you have good linear scales then you can compare linear position to angular position, with usually linear position being the more accurate, and demand that the servo
alter position to minimise the linear error.

All Feild-Oriented-Control servos, normally called AC servos, and the current state of the art in servos, require a very accurate and immediate knowledge of the angular position of the servo
in order to operate, and thus MUST have a rotary encoder. What I have elluded to is that you can have dual loop control, where the rotary encoder is used in the torque and velocity loops, but
the load sensing encoder/linear scale is used to control the position loop. As I say this is the last word in machine position control. If you can have it ....you want it!!!
This scheme would maximise the excellent mechanical properties that your machine has.

Over the course of time it is servos and particularly servo drives that have advanced and result in the machines we have today. Your machine is well built and rigid and would compare favourably
to any machine made today in terms of mechanical accuracy and rigidity. What a modern machine has is a much more sophisticated CNC control....and less, surprisingly enough, about the controller
or PC, but the smarts built into the servo drives. In this example we are talking about the servo drive, a Delta A2 series, offering this dual loop feature. The PC/controller can and is in fact fairly
basic, its the servo drive that adds the state of the art feature and the increased resolution and accuracy that comes with it.

Now I'm not saying you could not get your existing analog servos (with rotary encoder) AND the linear scales to operate in the same manner with LinuxCNC, but unless you are an experienced realtime
programmer it would not be easy, whereas using suitable AC servos then its (comparatively) easy.

Craig

Craig

Hi,
depending on the age of your servos the situation could be different and worse than you suppose.

In the early days,and to be honest, I would think predates your machine, the servo did not have an encoder, but it had a tachogenerator, or a speed sensor, and that was
feedback to the servo amp and it assisted in the closing of the velocity loop and the controller (PID controller) took encoder, or more likely linear scale feedback and
used it to close the position loop. Then in later years the tachogenerator was replaced by a resolver, which serves a similar purpose to an encoder, but encoders were at that
time incredibly expensive and fragile.

When I went to University and studied Control Engineering, many of the systems I studied and very extensively experimented with, and was tested, TESTED, and TESTED
on were of this type. We could at that time really only dream of Field-Oriented-Control being practical.That was in the early 80's. By the 90's however encoders and microporcessors
had advanced that much that Field-Oriented-Control was feasible and by the late 90's was de-rigour in the servo industry.

Advances in microprocessors, encoders, fast power electronic devices has only ever increased since.

I can buy a microprocessor that can do  floating point single-cycle -multiply-and-add at 90Mhz, with a whole suite of advanced peripherals for rotating machinery control for $20!!!!
A few years ago that was unheard of. Likewise I can buy a 1200V 20A MOSFET for $5.00 that can switch all day at 100kHz...amazing.

If you do buy new servos you are in effect buying into this evolution of electromechanical devices, and while not for free, at prices which only ten years ago would have been
considered impossible.

Craig

I have been a bad lad & spent most of the day putting this back together.
X & Z axis motors are 4 1/2" square with 95mm square hole pattern. About 8" long including encoder cover. Don't know the shaft size. I buried the other motor before I remembered to measure. I think that makes the small axis motors NEMA45. Plate & mounting are in the attachments.
Ratio of X is 20:92 on a 6mm single start screw & backlash is approx 1/2 a pitch on the primary pulley (9 deg) which works out at 0.0175mm lost motion. So obviously some of the polish has come off the machine in 30 years.

That is the spindle motor plate in the previous post. Here is the X axis & Z axis motor

Hi,
OK, the top name plate, the spindle , looks to be an asynchronous (induction) motor of 5.5kW, quite a substantial motor.
It would need a 230VAC three phase input VFD. Looks like the speed range is 1500 to 8000 rpm. It would make sense to reuse this
if its in working order. Is there a VFD/driver built in for this motor?. If there is and its working then by all means retain it.

The bottom name plate is less clear.

Maybe some German readers are following this thread and can help.
Md.konst =0.44Nm            motor constant 0.44Nm/A
Dauer 2.2                        guessing this is rated torque 2.2Nm
Dauerstorm 5.0                 translates to constant current 5.0A
n_max 4000       rated rpm 4000

This suggests a power of (4000/60)x 2PI x 2.2 =921W and is consistent with the size of the servo.
Given that the servos are not direct coupled suggests that fitting metric motor would be possible if a little fiddly.
It would require you make a new plate to take the new servo, but should be able to screw the plate to the existing threaded holes.
If I were you I would be looking for some 750W servos, or if you want a little more then 1kW servos. I think 750W would be more
than enough, and as they are a common size can be had very competitively priced.

If you don't want load sensing servos then 750W Delta B2s are perfectly fine and cheapest being Deltas entry level.
If you want load sensing, either immediately or for the future, then 750W Delta A2 series.

B2 for $438USD:
https://www.fasttobuy.com/flange-80mm-239nm-ac-motor-driver-kits-with-3m-cable-220v-075kw-cnc-servo-motor-for-cnc-router_p28084.html

A2 for $634USD:
https://www.fasttobuy.com/ecmac10807rsasda20721f-delta-220v-750w-239nm-3000rmin-80mm-dmcnet-ac-servo-motor-drive-kits-with-3m-cable_p27766.html

Note that I have seen advertised elsewhere these A2 servo kits for only $50 more the the B2 kits....it may pay to do a little more searching,
although I have faith in this supplier.

Note this same company has their own brand of servos, ToAuto, which are very well priced. I have no reason to doubt
their quality or performance but they wont have load sensing and they don't have set-up and tuning software.

If you are new to servos then having good (free) set-up and tuning software is to my mind essential. This forum and CNCZone is littered with posts
about people who've bought el-cheapo Chinese servos and then struggle to get them to work. Its not that they won't work, but if you've never
set up a servo from scratch before it is a challenge. I would not recommend you start your learning curve there, buy either DMM (Canadian brand, made
in China) or Delta (Taiwanese brand, made in China), both are good quality, good backup and most importantly set-up and tuning software.

Craig