Re: Encoder/servo selection for a Robotool centroid retrofit
Quote:
Originally Posted by
Jim Dawson
Linear encoders would be my choice even with the Gecko drives. There is no downside to using them, and many advantages. You just have to match the resolution to what the Gecko drive is capable of reading.
Yea I'm researching that right now. The Centroid requires a minimum of at least 1000 ppr encoders and recommends 2000ppr/8000cpr encoders. Trying to find documentation on the Gecko to see if it supports that.
Quote:
Originally Posted by
Al_The_Man
I initially purchased the CUI BLDC encoder development system for programming the simulated hall tracks for the BLDC motor encoder version, but got anything but satisfactory results, and the tech support was no help, so it left a bad taste.
I traditionally used Renco, but then Heidenhain took them over and the line selection shrank quite a bit.
Koyo are good as well as Hengstler.
If possible I usually fit through hole type to avoid mounting and a coupler.
Your original servo's may have had tach feedback which is not used now on DC motors, as Torque mode drives are used as opposed to the DC velocity type which had a tach back to the drive.
Al.
I will look into those brands shortly. They may very well be tach feedback. From what I have read/seen, replacing the old resolver/encoder and leaving the tach in place but not hooked up should be ok?
Re: Encoder/servo selection for a Robotool centroid retrofit
What I used to do when using new drives was to remove the tach brushes at least to avoid any problems.
Al.
Re: Encoder/servo selection for a Robotool centroid retrofit
Can you guys help me understand some of the terminology here? First what are they referring to as the lock range? I'm not seeing any specific encoder resolution maximums called out in the documentation yet. That seems unusual though? I have seen minimums called out but no max.
Also from the Centroid encoder requirements
"1. Encoder Type: Centroid MPU11 based CNC hardware (Oak, Allin1DC, Acorn) requires quadrature, differential encoders which operate on 5VDC. This is the most common type of encoder. Encoders labeled "TTL type" (non differential) are unsuitable, and fortunately are quite rare."
Is the bellow spec saying the G320x requires TTL? or that its simply able to run it as well?
G320X SPECIFICATIONS
Power Supply: +18VDC to +80VDC
Quiescent Current: 30mA
Motor Current: 0 to 20 Amps
Motor Inductance: At least 1mH
Short Circuit Protect: 22A trip
Lock Range: +/- 256, 512, 1024, or 2048
Feedback: Quadrature TTL Encoder 5V Compatible
Feedback Resolution: X4 Encoder Line Count
Feedback Voltages: <1V for Logic 0 and >4V for Logic 1
Switching Frequency: 20kHz
Current Limit: 0 to 20 Amps Trimpot Adjustable
Analog PID: Proportional, Derivative, Integral Coefficient Trimpots
Step Pulse Frequency: 0 to 300kHz
Step LED “ON” Time (min): 1uS
Step LED “OFF” Time (min): 2.5uS
Size: 2.5” x 2.5” x 0.8375”
Package: Blue aluminum anodized cover, black aluminum hard anodized bottom heatsink
Weight: 3.6 oz. (100g)
Encoder Supply: +5VDC 50mA maximum
Pulse Multiplier: 1, 2, 5, or 10 times input step rate
Re: Encoder/servo selection for a Robotool centroid retrofit
Lock range seems to refer to the maximum allowed Following Error settings. Some following error is normal when in motion, excessive following error will cause accuracy problems. This is where having the loop closed at the controller rather than the drive is beneficial, the controller will try to compensate for an excessively lagging axis by making adjustments to all axes on-the-fly, thus following the correct motion profile. Closing the loop at the drives will do this to some degree and is still pretty good, but not the best.
The Gecko will only accept a TTL encoder input (A and B), the good news is that you can use a differential quadrature encoder and just not connect the /A and /B signals. Later when you want to upgrade the controller and / or drives, you will still have the full differential encoder available.
Re: Encoder/servo selection for a Robotool centroid retrofit
I could never understand that decision by Gecko to go the single ended route, when it would be so easy to add a RS485 receiver on the drive.
Al.
Re: Encoder/servo selection for a Robotool centroid retrofit
What is lost by not connecting the /A and /B signals? I was doing a little research and could not figure out if it was related to filtering out noise or its positional accuracy.
Also since the centroid Acorn board is actually connected to the drive and not the encoder, does it matter if the encoder is running in TTL if the drive is still outputting the step/dir signals? I'm also trying to figure out if the encoder requirements are actually for the spindle encoder (which is connected to the board through a DB9 port) and not the servo drives.
If losing the /A and /B signals is a significant loss though, I will scratch the Gecko as an option and look elsewhere.
Re: Encoder/servo selection for a Robotool centroid retrofit
The full differential quadrature signal from the encoder is more or less noise immune, at least much more so than single ended. The way the encoder is read by the device is a function of how the device (drive in this case) is designed. In reading the specs, it looks like the Gecko does read both the rising and falling square wave signal from the encoder channel, so you would get 4 pulses per encoder line. Two for A channel and two for B channel, which are 90° out of phase with each other.
In the interest of clarity, the Acorn outputs a step & direction signal to the drive. The drive then tries to keep up with what the Acorn is telling it to do to position the machine correctly, it does actually work.
Any encoder requirement in the Acorn manual should be related to the spindle. The Acorn has no idea what the drive is doing with the exception of any error signal between the drive and the Acorn.
Re: Encoder/servo selection for a Robotool centroid retrofit
What I have done in the past where I had no choice and had inputs that were single ended with a long run in a noisy environment was to use one of the RS485 IC's 26LS32 for receiving and 26LS31 for transmitter, if needed.
As Jim said using the 4 edges of a A & B multiplies the resolution x4, some controllers also allow options, natural count, or plus x2 (2 edges) or plus x4 (four edges).
Al.
Re: Encoder/servo selection for a Robotool centroid retrofit
Ok so my take away is that the Gecko would be ok except in a very "noisy" environment?
Would it be worth looking at something like CNCdrive DG4S 08020 drive? Reading over its manual it does seem to have the capability to read the full differential quadrature signal through an RJ45 connector. Any opinions on that drive? I have never heard of the brand before.
https://cnc4pc.com/dg4s-08020-dc-servo-drive.html
Also thanks for the help, you guys have been amazing. Was really just planning on trying to put together a semi plug and play system (DMM) but this stuff is cool.
Re: Encoder/servo selection for a Robotool centroid retrofit
I have only heard about them, but know nothing personally. A cursory look indicates it is worth taking the time to study it a bit.
Re: Encoder/servo selection for a Robotool centroid retrofit
So I pulled the cap off the top of one of the servos and found something kinda interesting about the encoders on this machine. Looks like they are Broadcom HEDS-5500 A06. Resolution is a little low at 500 CPR. That means I don't have to worry about tacks though I assume.
HEDS-5500#A06 - BROADCOM LIMITED - Incremental Rotary Encoder, Optical, 2 Channel, 3000rpm, 500 Cycles, 1/4" Shaft Diameter | Newark element14
I emailed Marcus Freimanis about the CUI encoders and he said they are actually what they use to test the drives, so definitely a good fit. He also said that the G320X can support a line count up to 5000, though it becomes increasingly difficult to tune them above 2500 due to the gain settings on the drive.
Re: Encoder/servo selection for a Robotool centroid retrofit
The MassMind non-contact encoders would work nicely in that setup. 4000cpr, TTL AB, and easy to mount since they just use a little magnet.
ENC2b: AS5047P High Resolution Magnetic Encoder. massmind
Re: Encoder/servo selection for a Robotool centroid retrofit
Quote:
Originally Posted by
WesM
I emailed Marcus Freimanis about the CUI encoders and he said they are actually what they use to test the drives, so definitely a good fit.
It is not that they don't work, just not ideal in a noisy environment, IOW more susceptible to interference on a long cable run.
Al.
Re: Encoder/servo selection for a Robotool centroid retrofit
Quote:
Originally Posted by
James Newton
Those are cool little encoders, I will have to look into them for future projects. At this point I have a few AMT113Q-V encoders coming in I plan on testing out. They should be drop in (hole spacing) replacements for the encoders I have right now
Quote:
Originally Posted by
Al_The_Man
It is not that they don't work, just not ideal in a noisy environment, IOW more susceptible to interference on a long cable run.
Al.
Thats what I assumed. Now what constitutes an environment that is too noisy would be an interesting discussion. Also what components would cause a levels of noise that would interfere with the encoders?
Re: Encoder/servo selection for a Robotool centroid retrofit
An extreme example is a large plasma gantry where the torch and power cables share the same cable chain to the gantry..
Al.
4 Attachment(s)
Re: Encoder/servo selection for a Robotool centroid retrofit
So i'm trying to figure out how I want to power the servos on this retrofit. The machine currently seems to use a transformer (Western Magnetics NX-189) and then rectifier. The from the manual the Robo power requirements are 10 amp 3ph 220v and 15 amp 120 volt. Would it be safe to assume that they are putting the 120v into the transformer?
So my thought is, get rid of the motor starter/coil system, reuse or replace the fuses and use the transformer with rectifier/filters to power my servo system. The Acorn will have a separate 5v power source (which comes with the Acorn). Does that seem doable?
As a side, it seems like they are using two transformers in the robo system? The first one has numbers on it, which i assume is the voltage output, the second one has no markings. The label shows the smaller transformer as a Western Magnetics WX-190, 012-0306 REV J. I cant find any info on the company though, must less the transformer.
With these kinds of systems/retrofits is it worth trying to use the old power supply/transformer system? Or would you guys just buy a new PSU?
I just noticed I didn't get pics/specs of some of the other rectifiers, will have to try and see what those are when I get the machine on Monday.
Re: Encoder/servo selection for a Robotool centroid retrofit
I generally use the old transformers etc, these supplies are usually very rugged, and superior to SMPS in many cases.
The pic's show the primary taps selection on the transformers, set for 120v
If you have a spare contactor, it can always be used to pick up the main power or the spindle when in E-stop.
Al.
Re: Encoder/servo selection for a Robotool centroid retrofit
Is it reasonable to assume the old transformer and rectifier are outputting the approximately 63 vDC required to run the drives/servos? or did the old servo amps/power board alter the voltage in any way?
Re: Encoder/servo selection for a Robotool centroid retrofit
Well I finally got some time to start looking at this again. Got out to check on the transformer and it seems to have two taps that output 36v and 72v respectively. The way I understand it that would be about 51v and 102v roughly once they are put through a full wave rectifier? Does 51v seem kinda low for servos that are rated at 63v? I checked my outlet and its putting out a solid 121.8v.