Atlas FLI reparation

Presenation

The Atlas FLI

The Atlas FLI is a wonder of a focusser. 3" or so in diameter, It can carry up to 11kg with a precision in the order of one micron! The mechanicals are absolutely beautiful and stunning... The material is top notch... but so is the price in the 4000€ (or more) range...

However the one that we use for the RC600 at Mars gave up the ghost. Something on the PCB blew up.

To summ things up, the FLI company was not able to help, and with the focusser declared "unrecoverable" by the owner, I proposed to try to gerry rig something to make it work again (the telescope and focusser bellong to the local townhall, so it's "publicly own" and as such has to follow lots of rules for capital material).

Mechanical design

The FLI is build around a moving center part gliding on 4 fixed posts and moving up and down through 4 belt driven threaded rods with an incredibely fine pitch (around 0.12mm per turn!). The belt seems to be a 280 teeth MXL belt with a 3.5mm width. At a 2.032 pitch, this is 568.96mm long. Each post has a 44 teeth pulley.  The driving force is a 40mm nema8 motor (8H2A40406) with 0.6A, 3.5V per phase rating and a 2.8N/cm holding torque. It has a 45° 15 teeth cog meshing with what looks like a 37 cog 2nd gear. giving a 2.466 reduction ratio. This drives a 15 teeth pulley which drives the belt.

This leads to roughly 7.5 engine turns per 1 pole turn or 17µm per engine turn!

With a 8mm run or so, this is 470 engine turn for a full run! Hence the incredible precision of the thing!!!

We will note that all moving parts are on bearings.

Original PCB

Here is the original PCB, the one the blew up. We will not the USB-B connector, the 12V power supply a homing switch on the top right, just behind the capacirot on the bottom right is a connector that links this PCB with a mini PCB holding the 2nd homing switch and the connector to the stepper engine lead.

Finaly, on the left, a red lead to the temperature probe and what is most likely the programming port for the CPU.

2 chips on the PCB are the CPU and the stepper motor driver.

Part numbers

I was given the following part numbers... I have not checked them

Motor part number is 8H2A40406A7 “Modified”.

The timing belt pn A 6G16-280012 (Note 0.080P)  from sdp-si.com

Also from sdp

A 1M 4-Y48015  15 teeth miter gear

A 6T16M044DF6005 2.03mm pitch (XML) 44 teeth polycarbonate pulley

A 6M16-015DF2504 0.080” pitch (XML) 15 teeth pulley

Fixing the HW

Theory

In theory, replacing the electronics is easy. You need a CPU that will have a USB connection, get order from the computer and drive the stepper engine (a stepper engine is a motor that will move by one "notch", typically 1.8° (200 notches per turn) each time you tell it to).

Arduinos are PERFECT for these type of operations...

On the left is a picture of my first prototpye. White USB cable connects PC to an ATTiny85 CPU, which in turn controls a stepper driver (a chip which will translate CPU level impulsion to high power motor drive) and a stepper motor on the top left).

The problem is that this does NOT fit in the FLI mechanicals! and the original FLI PCB is minuscule

Step 2: a custom made PCB (Version 1)

So, off to easyEDA, an online tool that let you create HW schematics and which is linked to JlcPcb, a company that will manufacture PCB and ship them to you (at a relatively low cost), all out of china.

My first version of a PCB was designed to be useable for other purposes as well and used the standard Arduino stepper shields. However, the ATTiny chip was designed to be directly on the PCB.

PCB design/manufacturing process

Layout

Once the schematics are crated, you need to move on to the layout, the physical placement of component and size of the PCB.

This entails PRECISE measurements of the original PCB, placement of the components and routing of the electrical traces

PCB ordering

Once your PCB is designed, the process in easyEda/JlcPcb is incredibly easy.

You see a visualisation of your PCB

Component selection

You select which components you want to put (or not) on the PCB

Component visualisation

They show you which components will, be there and where...

Billing and shipping

It's not all fun and games, things are not free :-( the cost was around 50€

Reception

And a couple of weeks later, you get your boards (5 because there is a minimum order quantity)

HW design walk through

At the time I started this project, I was working a lot with the ATTiny85 chip. This is a VERY small cpu with only 8 pins. 2 for power and 6 "free" ones. Furthermore, it can handle USB. Abeight in a weired way as the USB is done 100% in SW, meaning that there is no HW assist. But it is very small, cheap and is handled by the standard arduino system, which makes it incredibely easy to work with.

So, 2 pins for power, 2 for USB, 3 pins for the stepper (1 enable, 1 direction and 1 step). This leaves 1 free pin which will have to be used for the homing buttons. The design "just fits"... (with the homing button being handled in an analog way and a divider bridge).

My original design was also done so that I could reuse the board for other project, hence I added space for 4 push buttons (2 being electroncaly identical to the homing buttons) and connection pins.

Also, the design made use of standard arduino stepper drv sheilds.

In theory, it worked... However, the design has a number of issues:

1) It mechanicaly does not fit in the FLI box, so I had to whip out the dremmel and remove some aluminium to make it fit :-(

2) There is an electrical issue with the USB and power (more about this later).

PCB assembly

As stated above, I am using the arduino system for this. But, if I had asked JlcPcb to solder the ATTiny on the board, it would not have come with the Arduino bootloader, which would have been a pain. In addition, their cost of the chip was very high (much higher than a purchase from ali express).

So, I ordered ATTiny from ali express (with the arduino bootloader). And unsolder the chip from board 1 to board 2.

I also had to unsolder 2 parts from the original FLI board. The homing switch and the connection for the 2nd homming switch (because the connector is longer that standard).

I also needed to add the capacitor (I did not have the right one in JlcPcb) and the headers for the stepper driver.

Finally, I learned the hard way that I needed to put some glue around the USB connectors as they were WAY to fragile...

Top of the picture is the ATTiny donnor board, then the raw version of the new board and finally a nearly completed final board (the homming switch is missing)

The software part

The firmware part

The firmware is the part of the software which resides on the ATTiny chip. As stated above all developped as an arduino project.

The USB part is done using a modified V-USB library.

Then, the stepper engine accelerations/deceleration had to be created. Handeling of the ADC for the homming switches was also a pain. But it all ended up working together.

Of course, there are a lot of other parameters to handle such as the max position, current position and speed, which have to be saved in the eeprom to avoid having to redo a homing cycle each time the system is powered up...

And of course, I had to device some type of protocol over USB to allow the device and the PC to talk together.

The ASCOM driver

This gets us to the ASCOM driver. Now, I had never done anything ASCOM, I barely knew what ASCOM was... So I had to learn all of it at the same time. The good thing is that the ASCOM people have done a great job at setting up example projects to get one started...

The bad thing is that it is all in C# and using COM, which are 2 things that I have never done! so I had to learn a new programming language, system and bunch of concepts at the same time. Ho, and the ASCOM example are all serial port based, but I was going USB!

Anyhow, I did end up getting it working! And I now have a CPU independent focusser driver for my FLI.

All the files are available for download here.

Conclusion of part 1

The FLI is back to working, over 4000€ where saved, but at the cost of around 60h+ of work. But it was a very interesting experience with lots of learnings...

Happy with my work, I posted all of this on the club website and put a link from cloudy nights... My mistake :-)

ATLAS FLI reparation: the gift that keeps on giving!

You know what they say about the reward you get from doing a good work: more work!

Well, it did not take long before I started getting contacts from all over the world by people having issues with their focusser. Apprently, I am the only one that is trying to fix them. I have now had 6 people contacting me with different issues. 4 of them being, like for us, a fried PCB, the other ones having more mechanical issues.

So, this section of the webpage will be dedicated to how to on fixing the various problems that I have seen so far.

Fried/not working PCB

As I said, I had 3 people contacting me because of fried PCB. The first contact was from Korea with www.sllab.co.kr who are using the FLI to move the secondary on a RC type scope!

When you have this issue, the FLI will not register on USB when plugged in...

Anyhow, I ended up creating a new set of PCB which have the stepper chip build in and are not using sheilds anymore. This means that there is no more need for a physical modification of the FLI anymore.

I do have 2 PCB left here that I can send to people that might need them.

Else, you can get the project at https://oshwlab.com/cyrille.de.brebisson/focus_copy_copy and you should be able to order the PCB from JlcPcb...

Do not forget to also order some ATTiny85 dijispark boards from aliExpress (take 2 or 3, it's easy to mess up the soldering and burn the CPU) as well as some 100µf capacitors.

Once you receive all the parts. First, unsolder the ATTiny from it's board and put it on the receiver board. make sure that you can put the FW on it before going further. If this gets messed up, start again with a new set of boards.

The wiering that I used for the stepper engine is not the same as the one used by FLI. See picture on the left for the wire order I used. To "swap" wires on the connector, using the tip of an exacto knife, life the plastic tab on the connector and pull the lead out. From memory, you need to swap the green and black wires. Test that the motor turns when the PC says so before doing anything else!

Then and only then unsolder the homing switch and connector from the original PCB and put them on the receiver boards.

Get the files here.

I stronly suggest that you get in touch with me if you want to replace your board it will most likley help!

The schematics of the PCB that fits perfectly

They are much less things here than in the previous version. But it fits.

Note that the driver is directly on the board, not on a carrier

PCB itself

Here is the PCB, adn you can see the real thing on the picture just above.

And here are the 5 brand new PCB (not yet mounted with the Attiny or capactiors)

At the top, you can see my original model and on the top right the Atlas original PCB

Mechanical issues

You might want to read this part of the article even if your issue is only with the PCB as it discusss the openning of the device and the mechanical intricacies of puttings things back together... Which is a PITA.


Step 1: opening the FLI step 1

The FLI is held closed by 3 large screws (4mm hex) and 7(8?) smaller ones (2mm from memory). Unscrew them and separate the 2 parts.

Note that on some of these, they have put some type of glue in one of the screw as an anty opening deterent. you will need to break it with a small nail or other pointy thingy...

WARNING, on some FLI, there is a small spring somewhere between the 2 parts. I have not manged to figure out where it comes from. So, if you do, tell me!

Once the 2 parts are separated, you will need to disconnect the stepper motor lead which will allow you to complete the process.

On the top of each of the 4 posts are 2 washers and 1 needle bearing. Make sure that they stay where they need to be and that you do not mix them as they are incredibely height sensitive.

Step 2: opening the FLI step 2

If you need to continue pulling things appart (I strongly advise you not to if you don't need to).

For example because your FLI is disgusting and you need to clean it up (see the engine pulley on the right) or because the belt broke. Here are the steps...

First, get the moving part of the FLI as "high" as you can. The best way to do it is to put a 2mm hex wrench in the top of one of the poles and to turn.

Then, using some type of tape, marker or something like this, try to "register" the poles positions to the moving part so that, when it comes times to reassemble things they will not have moved...

Then, you will need to unhook the belt from the tentionner where the PCB is located.

Then, using needle nose pliers, try to unhook the belt from the 4 cogs at the bottom of the 4 poles by moving the belt down as you pull (gently) the moving part up... This is the easy part.

You should now be able to disasemble everything.

If you need to fully remove the belt, you will need to remove the engine.

It is held by 2 screws on the outside of the box, but they can be hard to remove. They are using, I think US sized hex wrench, not a metric one. But they can be incredibely tight and can strip.

The best way to handle it can be to use a dremmel and just cut a slit in the head and then use a straight screwdriver. (picture of the 2 screws on the right, before I cut the slit). I discourage you from removing the engine the way I did in the top picture as it is a pain to put it back together!


The underside of the carrige

A problematic belt

Step 3: cleanup

Use denatured alchool for the cleanup. And then white lithium grease to regrease all moving parts...

Step 4: Reassembly of the carrige

This part is the HARDEST of them all!

Assuming that you have registered the 4 poles, and that you have headed by advice to fully retrace the carrige before removing things, try to route the belt properly and insert the carrige.

Then, using needle nose  pliers, try to place the belt in the proper location. Sometimes it is easier to first get the belt "too high", and bring it down in position at the end.

They should be a little bit of play. 

Once this is done, you should be able to put the belt back in the tentionner...

This part can take some time and will require some calm and concentration... Be patient, it is not that hard! But if you don't have small little pliers that can reach around tight corners, it will not be easy.

Another thing which you need to concider is how you will "hold" the FLI so that it rests on the housing and not the carrige. use spacers or anything to make it work.

You sometime will need to force a little bit on the belt to get it on the cogs. Make sure that they do not turn as you do so (or that they ALL turn at the same time).

Once you have complete this, you need to make sure that you have not messed up the balance of the carrige. i.e. make sure that it does not wobble when you press right/left on 2 diagonally opposed corners.

If so, you will need to turn one, and only one of the 4 screws until the wobble is gone on both diagonals (remove the belt from the tentionner and, while blocking the other 3 wheels, force turn the screw of interest).

Step 5: Test the PCB

If you have changed the PCB, you will need to test it.

Open the ASCOM config window first.

Connect the PCB, to the stepper motor, make sure to update the stepper motor wire as descibed above if/as needed.

Connect the system to power and USB and click on connect on the driver window.

If the motor starts moving, click on stop ASAP.

If the motor does not start moving, check the max position and current position. Put some values in them (like 100 000 for max pos and 50000 for current pos, validate) and then try to do a move in or out to make sure that the motor turns. If yes, then stop, you are good.

Disconnect power.

Re-attach the PCB. But first, make sure that you put some tape on the housing to avoid shorts. rescrew the PCB (I think that they are 1.3mm hex).

During all this time, make absolutely sure not to get any shorts.

Also, do DO NOT WANT to connect the stepper motor when the system is powered. This is the best way to fry the board.

Step 6: Close the FLI

Before closing make sure that you put some white lithium grease on the bearings!

This one can be tricky for 3 reasons:

1) The washer and bearings tend to move as you lower the top part and be pain the rear end.

2) The connection need to be done between the PCB and the second homming switch. If the connector is slightly out of shape, it will be problematic

3) There is not a lot of play to work...

For 1) I have no good solution. Sorry.

For 2) my advice is to first make sure that these 2 are ligned up. To do so, do NOT connect the stepper cable, but tuck it in, move the carrige to give a good wiew from the inside of the light path and try to see if/why the connector does not go in properly (bent connector, bent capacitor, something in the way) and fix this. Once this is ok, and only then, connect the motor and finish the closing.

Step 7: Test

You are now moslty done.

Reconnect USB and power, reconnect the driver and attempt a homming cycle.

If all goes well, the FLI should first go one way, then the other, and finally go to the mid point.

If something goes wrong, then it will try to overshoot the bottom or the top they should be some indication on the driver page telling you what it detects. In this case, just contact me to trouble shooting. Do not let it continue.

Rescrew the closing screws.

Then try a homing cycle again.

One thing that might happen is that the FLI might force at some place and you will hear the engine wine. This is mostly caused by overtightening of one of the 4 vertical poles. So it will have to be loosened a little bit.

First, make sure that this is the problem by loosening the closing screws and trying a homing cycle again. If it then works, then this is most likely your problem. You can try figuring out which pole is too long by see if there is an uneven gap between the 2 halfs of the FLI. or by trying to only tighten some of the closing screws...

Using a small wrench and a 2mm hex, loosen the nut on the problematic pole and move the top plate down, re-tighten and try again.

You want to minimize the play, but not put pressure on the pole. Hence the dificulty. Good luck, this can take some time!

Downloads

Here are all the files that you will need.

The driver

The rest of the SW files if you want to play with them

Driver Configuration window

This is basically the user manual for the driver setup window!

If it does not show connected, they should be a connect button and by clicking on it, if the device is connected through USB and powered up, it should connect.

Setup the step size. Normally you should not have to change it, but you can if/as needed.

The step delay in µs is the speed of the focusser. It represents the interval in microsecond between 2 stepper steps. So 120 means one step every 120 micro second or 8333 steps per second or around 5 turns per second of the motor (1600 steps per turn) or around 0.1mm per second. I think.

Max position is the maximum position (in steps) where the focusser can go. The min position being 0 of course. Normally, this is automatically detected when doing a homing/init sequence (see button), but you can manually change it (for example if your homing does not work). Do not forget to press "change max pos" to validate.

Position and mouvement allows you to move the focusser. you will see here the current position and can enter a new value and "go to", or "sync to". althrough this 2nd option is probably not somehting that you will want to do (see comment on max position).

in/out buttons allow you to move the focusser in a controled way to test things out.