Automated XYZR Stage, The R Component

mawyatt · Post by **mawyatt** » Sat Jun 15, 2019 6:59 am

Chris,

There are a few surplus motorized goniometers on eBay, but still $180~300 and up range. If I do this I'll use two, so looking at $360 + which is above what I'm willing to spend for this capability now. Also looks as if some of the motors are the older 5 phase stepper types, which I can't control with my present setups, so possibly need to replace the motors with 2 phase types.

So now looking into a 6 axis system, XYZR and X and Y tilt!! The capability (8 axis or more) to do this is already present with the Quad Axis system

If someone wants to use these with the Trinamic or Pololu controllers I'm developing, they should work fine with the proper motor and setup.

Best,

mawyatt · Post by **mawyatt** » Sat Jun 15, 2019 7:37 am

Mentioned earlier about ordering thinner stepper motors for the rotational axis. One of these, the 14mm (actually it's less than 13mm!) thin 0.9 degree motor arrived yesterday. These have a brass belt adapter which I had to remove. Tried heat, pounding and eventually had to cut the adapter off (don't have a proper gear puller). Probably damaged the bearings, but it works fine, have another, and they were cheap so I'll order another backup.

I've mounted the motor to a small 40mm square ARCA plate with foam tape, later I'll epoxy it in place.

These small thin steppers have a high internal resistance ~18 ohms and will get very hot unless you curb the holding current, also they need a high supply voltage (~12 volts or greater) to operate properly. Remember stepper motors are current mode devices (current induced magnetic field) and with modern controllers which use pulse modulation to control the motor current, the motor internal power dissipation is computed as I^2 Rmotor, not the usual V*I, so higher voltages usually result in lower motor temperatures!!

These thin 0.9 degree motors don't have much holding torque, but otherwise work like a charm and fit nicely on a small 40mm ARCA plate!!

Best,

chris_ma · Post by **chris_ma** » Sat Jun 15, 2019 8:16 am

mawyatt wrote:For anyone that's interested in utilizing these devices I'm here to help those in this community so they won't suffer the long, arduous task I went thru!!. I've spent 1000's of hours and many many 1000 of $ to gain this knowledge, so you don't have to!!

Hi Mike,

thanks for the extremely generous offer!
I'm in the process of planning a simple motorised Z camera setup (vertical) that I'm hoping to extend later on by a XY flat XY stage later on.

I already ordered a Trinamic TMCM-1110 board and a used KR-20 rail since those seem to be good candidates from my research and hope to get all the bits and pieces during the next week.

I have some basic programming and electronic knowledge but never worked with stepper motors so far, so any help with getting a simple setup running will be much appreciated.

I'll start a new topic as soon as I get the parts as it might be also interesting for other beginners.

looking forward to your future videos, these acceleration movements have something very hypnotising

chris

mawyatt · Post by **mawyatt** » Sat Jun 15, 2019 9:14 am

Hi Chris,

I'm not familiar with this Trinamic device? I am familiar with the TMC5130, 5160, 5161 and 5072 controller/drivers though, and have developed custom Printed Circuit Boards for these and various combinations of such. These all work directly from the Raspberry Pi.

I'll try and help as much as I can.

The KR20 is certainly a good choice for the rail, with a good 0.9 degree motor they perform quite well indeed!!

Best,

mawyatt · Post by **mawyatt** » Sat Jun 15, 2019 9:29 am

Chris,

Looked up the TMCM-1110 TMCL. This is based around an older seprerate TMC429 motion control chip and a TMC262 driver chip. Also doesn't seem to have all the features of the newer chips which integrate the controller and driver on a single chip.

I'm not familiar with the TMCL language either, so may not be of much help...but will try.

If it's not too late you might consider switching to one of the Trinamic devices I mentioned, and operate from Raspberry Pi or Arduino. I can help with the RPi but not so much with the Arduino.

Best,

chris_ma · Post by **chris_ma** » Sat Jun 15, 2019 9:41 am

mawyatt wrote:I'm not familiar with this Trinamic device? I am familiar with the TMC5130, 5160, 5161 and 5072 controller/drivers though, and have developed custom Printed Circuit Boards for these and various combinations of such. These all work directly from the Raspberry Pi.

As far as I understood the 1110 can be directly programmed from a PC IDE through an USB connection so I hope it will be easier to control. here's the getting started pdf:
https://www.trinamic.com/fileadmin/asse ... tarted.pdf

I ordered this before I saw your results, but if I fail to make it work I might bug you for one of your custom boards so I can use all of the cool acceleration profiles

chris

ps: hmm, I might be still able to return the board, so I'll check that. your setup certainly looks very elegant and I do have a basic understanding of the RPi

chris_ma · Post by **chris_ma** » Sat Jun 15, 2019 10:12 am

one question:
what's the difference between your custom board and the boards of the Trinamic EVAL-KITs? (like this one):
https://www.trinamic.com/support/eval-k ... 5130-eval/
Is it mainly cost saving or does your design enable more features?
I'm mainly asking because I'm a bit worried that if I use a custom board and fry it somehow (I'm known to be clumsy

it will be hard to replace, so using a reference design would lift that worry even if it's a bit more expensive.

and one more if I may:
which of the four controller/driver chips you mentioned (TMC5130, 5160, 5161 and 5072) would you recommend for a simpel but precise stack and stitch system on 3 axes? (eventually, the first few weeks/months I'll have enough to learn with Z only).

mawyatt · Post by **mawyatt** » Sat Jun 15, 2019 10:13 am

Chris,

Looks like it should get you up and running, but doesn't show any of the advanced features of the modern Trinamic devices. This was the very reason I went with the Trinamic devices, for the advanced features.

The Pololu Tic-500 devices I was using (and still do) work extremely well and was getting excellent results with. However, I wanted to try deal with the motor/rail induced vibration issues to get to 1~2 second image spans for massive S&S efforts. The Tic-500s don't have the ability to control all the various motor, timing and current waveform parameters like the Trinamic devices, so I use the Tic-500 for nominal XY and even could use for R axis control, but reserve the Trinamic for Z where the advanced features can be beneficial.

Believe me, I have no fondness for Trinamic, nor their Customer Service, unlike the Pololu Customer Service which is exceptional!! The Trinamic chips are very powerful though, but also very complex to use. There wasn't much on using these Trinamic devices on the RPi either, and almost nothing on using Python, so for me this was a venture into the unknown!!

Best,

mawyatt · Post by **mawyatt** » Sat Jun 15, 2019 10:34 am

chris_ma wrote:one question:
what's the difference between your custom board and the boards of the Trinamic EVAL-KITs? (like this one):
https://www.trinamic.com/support/eval-k ... 5130-eval/
Is it mainly cost saving or does your design enable more features?
I'm mainly asking because I'm a bit worried that if I use a custom board and fry it somehow (I'm known to be clumsy it will be hard to replace, so using a reference design would lift that worry even if it's a bit more expensive.

and one more if I may:
which of the four controller/driver chips you mentioned (TMC5130, 5160, 5161 and 5072) would you recommend for a simpel but precise stack and stitch system on 3 axes? (eventually, the first few weeks/months I'll have enough to learn with Z only).

Chris,

These "Eval Boards" are all good and useful, I almost went that route myself. However, you will need not only motor control but camera and strobe triggers which these boards don't have...so you'll need to add these features to the "Eval Boards" in some way, thus another board or something (also wanted the ability to use a Homing sensor for Z axis homing).

Ended up just doing everything myself, including the PCBs. They fit within the RPi without the need for excess cables and such by using the RPi 40 pin connector directly. Also include the ability to power the RPi from the Motor supply with a dedicated DC to DC Converter since I didn't want an extra supply (for RPi) sometimes. Also induced the ability to "plug in" the Trinamic devices, so they could be replaced or used with another custom PCB if desired, everything was done with total flexabiliy in mind from the start, not an afterthought!

When I got the RPi VNC working, now things work really well from an operations standpoint, doesn't require a keyboard, mouse nor display with the RPI. No USB connection required from the remote computer to RPi either, all WiFi enabled wireless and it works extremely well indeed!!

Honestly the only issue I'm up against is the sudo pigpiod (sets up RPi GPIO ports), which causes a "no handles available" error message often enough to be annoying (maybe every 10~15 sessions). This is not uncommon with RPi and Python, evidently caused by not closing all the pigpio ports. I'm sure it will get resolved by someone soon, but for now just an annoyance!

I would recommend the TMC5130A chip for Z, and dual TIC-500s for X and Y. I'm working on other configurations but not ready to recommend them for "Open Use" just yet...need much more testing and evaluation. The TMC5130A and Tic-500s have had extensive testing.

Best,

Chris S. · Post by **Chris S.** » Sat Jun 15, 2019 9:31 pm

mawyatt wrote:There are a few surplus motorized goniometers on eBay, but still $180~300 and up range. If I do this I'll use two, so looking at $360 + which is above what I'm willing to spend for this capability now.

Mike,

I'm pretty sure you know this, but just in case: With goniometric stages, you don't want two identical goniometers; you want a pair of complementary goniometers, with the top one having a lower point of rotation than the bottom one, such that the goniometer pair, when stacked, has a common point of rotation. This can be seen here.

The problem with most goniometers on eBay is that only the bottom one or top one--or multiple copies of either--is on offer. Complementary pairs are much less common, and when they appear, tend to command high prices. (Or do now, since we started using them here.)

Of course, with the number of axes in your system, you can probably correct for the drift induced by using a pair of identical goniometers, but this would add programming complexity.

This is probably old news to you, or if not, then something you haven't gotten into the weeds on because it's not on your immediate to-do list.

--Chris S.

mawyatt · Post by **mawyatt** » Sun Jun 16, 2019 5:42 am

Chris,

Yes I am aware of this, thanks for the reminder though. I was looking at one with a 75mm rotation point and another with 50mm rotation point for the top goniometer, which was about 25mm thick, so the overall rotation point should be about right. The cost was too much for the benefit for my simple uses though. If I could find some at a lower cost I might reconsider since the added electronics are almost trivial, maybe not so with the software though, but I've got that well sorted out so not a massive effort to add this X and Y tilting feature.

Best,