Acorn vs Oak pulse frequency and servo motor resolution

[adthelad]

Hi,

I am doing a comparison of Acorn vs Oak for a gantry style mill. I am looking at using Delta servo drives and motors. Typically these motors are provided with 17 bit incremental encoders, with 20 bit as an option. My question is about the usability of these higher encoder resolutions on a Centroid system.

If my math is correct, Acorn has a max 400kHz output pulse frequency, so a 5mm pitch ballscrew could not move faster than approx 15.26 mm/s using the full resolution of the 17 bit encoder without any electronic gearing. With 20 bit, this is much lower, 1.9 mm/s.

So naively, I see not much advantage to having a higher resolution encoder on the servo motor with Acorn as I will be forced anyway to use some electronic gearing anyway.

I was trying to dig deeper into the Oak controller and understand how it communicates with the Delta servo drives. The controller comparison page says that that Oak has 'Position/Precision & Velocity modes'. Would Oak also suffer from limits on how fast pulses can be delivered to the servo drive for higher resolution encoders or is this mode of operation totally different?

Thanks!

[eng199]

Electronic gearing must be used to bring the encoder counts per revolution down to something manageable, like 8192 to 32768 counts per revolution. If the axis ratios (turns per inch) are reasonable on the machine, this has no effect on actual accuracy.

The protocols involved are not suitable for extremely high pulse rates. Step and direction and quadrature both have relatively low limits because they make a pulse for each count. This is why all high resolution encoders that rotate at significant RPM use a serial protocol that sends a binary count value.

The advantage of high resolution is realized in the servo drive. It is doing the position control. The extra counts per revolution give higher precision in control loop calculations for smoother motion. Also, position error can sometimes be kept below the resolution that the drive is reporting via electronic gearing. This results in the motors following control (ACORN / Oak) commands down to the last bit of control resolution setting. Managing following error is the job of the servo drive in this configuration. There isn't much advantage to having the control resolution so fine that it is in the following error range.

[adthelad]

In the case that the servo drive is doing the position control calculation, is it true that there is no difference in position precision/resolution achievable through Acorn versus Oak?

Looking at the pinout between Oak and a Delta servo drive, the control method still seems to be a pulse train, similar to Acorn, but the extra features on Oak are encoder feedback from the drive to Oak as well as analog feedback. Is Oak still limited at 400kHz?

[eng199]

In the case that the servo drive is doing the position control calculation, is it true that there is no difference in position precision/resolution achievable through Acorn versus Oak?

On the typical machine, there is no difference. The maximum speed, axis ratios, machine mechanical accuracy, and other factors would have to be considered to determine if there will be a real difference in a particular configuration. Accuracy comes into play as well as precision/resolution. For example, 20 bit encoders used on motors aren't accurate to 1 /1048576 of a revolution. The resolution is there, but not accuracy. Usually the overall goal is to make accurate parts. Resolution greatly in excess of accuracy is not a benefit in the overall picture.

Looking at the pinout between Oak and a Delta servo drive, the control method still seems to be a pulse train, similar to Acorn, but the extra features on Oak are encoder feedback from the drive to Oak as well as analog feedback. Is Oak still limited at 400kHz?

ACORN is not necessarily limited to 400kHz. The open collector outputs will limit speed. The limit will also depend on external circuitry and cable length. The logic level outputs can go faster, depending on what they interface with and if 5V levels are an option for the drive. The rate setting is (1.2MHz / (parameter 968)). This gives a theoretical limit of 1.2MHz, which is not practical for any step and direction drive interface.

Oak is also limited to 1.2MHz (1,200,000 encoder counts per second). However, Oak outputs quadrature A/B signals just like a typical quadrature encoder. It is reasonable to run quadrature up to this limit.

[adthelad]

Thanks for the comments. Personally I prefer line driver outputs rather than the open collector outputs from an electrical/noise immunity point of view, but this is probably a mintor detail most for use cases that can be managed with careful enclosure layout. I do see that this is planned as a future expansion for Acorn six which now tempts me.

For a gantry router, I would like output for five servo drives, X, 2xY, Z and option for fourth axis. With Oak, it looks like I would need to use the OpticDirect 120112 board to expand to more than four axes. Is it possible to have four drives hooked directly to Oak over the H3, H4, H5 and H6 connectors plus one OpticDirect 120112 board, or would I need five OpticDirect 120112 boards? I know I could hardware pair the Y drives, but I would like to leave the option of software pairing on the table. It also looks like Centroid do not sell the OpticDirect 120112 board on their website?

Otherwise Acorn six may be the better option for me to get around the extra enclosure space/cost of the OpticDirect 120112 boards.With possible quadrature drivers in the future, my concerns about Acorn seem to go away. Acorn six really looks like it covers almost all of the same specifications as Oak - are there any important details to consider that I am missing?

[eng199]

Personally I prefer line driver outputs rather than the open collector outputs from an electrical/noise immunity point of view,

I agree.

I do see that this is planned as a future expansion for Acorn six which now tempts me.

There is a board with differential line driver outputs in testing phase. Acorn Six can output quadrature or step and direction.

Is it possible to have four drives hooked directly to Oak over the H3, H4, H5 and H6 connectors plus one OpticDirect 120112 board

That is how it's normally done. You will probably get OpticDirect 220323, which is functionally equivalent.

Acorn six really looks like it covers almost all of the same specifications as Oak - are there any important details to consider that I am missing?

Oak takes encoder feedback from all the drives. When everything is working right, this isn't a major benefit. If a problem communicating position to a drive occurs, the control doesn't know and can continue indefinitely with position errors. Feedback is also useful to see actual error on the Centroid live tuning screen during setup.
Oak uses drive specific cables with all communication signals included in one cable. For drives that Oak cables support, wiring is much quicker and simpler. Acorn Six uses a more generalized approach, which will be easier if there isn't a compatible Oak cable.

[adthelad]

With AcornSix, is it possible to take linear encoder feedback from four motors (X, Y1, Y2, Z)? I see that it says AcornSix is compatible with the encoder expansion board which I think I would need for feedback from four linear encoders. For a gantry style machine with two motors on the Y axis, is it possible to have two independent linear encoders for feedback on each side of the gantry?

[cnckeith]

With AcornSix, is it possible to take linear encoder feedback from four motors (X, Y1, Y2, Z)? I see that it says AcornSix is compatible with the encoder expansion board which I think I would need for feedback from four linear encoders. For a gantry style machine with two motors on the Y axis, is it possible to have two independent linear encoders for feedback on each side of the gantry?

i assume you would be software pairing Y1 and Y2?

[adthelad]

With AcornSix, is it possible to take linear encoder feedback from four motors (X, Y1, Y2, Z)? I see that it says AcornSix is compatible with the encoder expansion board which I think I would need for feedback from four linear encoders. For a gantry style machine with two motors on the Y axis, is it possible to have two independent linear encoders for feedback on each side of the gantry?

i assume you would be software pairing Y1 and Y2?

Yes, I would prefer to software pair the axes if possible.

[adthelad]

With AcornSix, is it possible to take linear encoder feedback from four motors (X, Y1, Y2, Z)? I see that it says AcornSix is compatible with the encoder expansion board which I think I would need for feedback from four linear encoders. For a gantry style machine with two motors on the Y axis, is it possible to have two independent linear encoders for feedback on each side of the gantry?

i assume you would be software pairing Y1 and Y2?

If I software pair the axes, would this be possible? My main concern is that with a gantry layout machine, is it possible to have indepedent encoders for both sides of the gantry? i.e. two y motors each with their own encoder