FYSETC S6 and TMC5160 strangeness.


John Petterson
 

Finally, after a dozen redos involving repeated redesigns of the daughter card and shipments of these from China, several generations of 3D printed cases, I have a controller running my AZ8 mount well.  The 5160s are a nice upgrade from the 2130s, quieter and capable of faster slews.  Fast slews, accurate tracking with high accuracy encoders to correct for missed steps, I am really happy with all of it.  But now it has developed a new symptom.

Testing in my kitchen last night I started it up and did a couple of slews.  Then I set it to a target and left the room to watch TV with my wife. After some amount of time, undetermined exactly how long but less than 2 hours, the motors stopped moving.  The OnStep thinks they are still available, and if you ask for a slew to a new target it shows the scope moving on the CdC or Sky Safari screen, but the motors do not turn, the scope does not move, and immediately after the ‘slew’ completes the encoders reset the pointer to the actual location which has not changed.  Any method of moving the motors fails - Android app, web page, BT gamepad, SHC all do nothing just as the planetarium programs do. It still shows tracking on.

Motors are cool, they seem energized (I can turn them but there is some resistance and I feel clicks).  If I turn the power to OnStep off and immediately back on, set date/time and start tracking, the motors work perfectly again.

Has any one seen this before?  Since it is both axis, should I suspect the FYSETC?  Is there some other software parameter I can look at or anything else to try?  I am running it with a 24V supply if that affects anything.


Khalid Baheyeldin
 

I have not seen that. I can do a good 5 or more hours under the stars with no issues.

What I would do is disable the encoders completely within OnStep, then try the exact same test to see if there is any difference.


Drew 🔭📷🚴‍♂️
 

Sounds like you hit a limit. Look for the error on the WebServer status page. Can't be entirely sure from your symptoms but that is where I would look first.


John Petterson
 

Good thought, but I doubt that limits are involved Drew.  I will check that page when (if?) it happens again though.

After it failed once last night, I left it tracking a circumpolar star (Mirfak), when I went to bed.  The time I started it it was about 10 degrees above the horizon in the NE.  When I checked today, the telescope stopped moving, pointing about 50 degrees up in the north east, while Mirfak was close to or just past the meridian.

I am at 42 degrees North, Mirfak is 49 degrees 56 minutes, so figure 50 degrees north of the equator and my limits are set to -10 degrees on the horizon and 87 degrees elevation.

break....

I caught it this afternoon with the Alt motor continuing to work while the Azm motor was not working.  I am starting to think it might be a heat issue with the drivers.  I am going to rig a fan inside the case to see if I can prove that.  

If this is the case, once they fail it seems the only way to fix it is power reset.  Turning tracking off and on does not help.  But it seems they get confused as to direction, the Alt motor is running backwards now.  After a power reset they are both OK and moving the right directions again.  But I am also having trouble with the Bluetooth gamepad not connecting.  So maybe that is being affected by heat as well.

I am going to have to stop working on this and concentrate on packing the house if I ever want to get moved.  So maybe you will not hear for me for a couple of weeks or months - but I will be back in a bit.


Drew 🔭📷🚴‍♂️
 

On Sat, Jun 5, 2021 at 06:03 PM, John Petterson wrote:
I am starting to think it might be a heat issue with the drivers.
That is completely unacceptable. We will have to see what could be wrong when you come back.


Khalid Baheyeldin
 

If you think it is a heating TMC5160 then that would be very odd. They don't heat up at all, specially when the motors are set to way less than the huge maximum of the TMC5160.

If you have an infrared thermometer, you can point it at the driver, provided it is not too shiny, and get a temperature reading. Do that before you do the fan thing.


Dave Schwartz
 

Maybe John didn't install the heatsinks. On the 5160, the MOSFET drivers (the part that does the power switching and thus could heat up) are the four little square chips and are not integrated on the logic chip like every other module we use. When installed, the heatsink sticks to the top of these 4 chips and not the TMC5160 chip. If the heat sink is in its proper place, overheating would be very strange because the MOSFETs used have quite a low on-resistance and thus they don't have much power loss unless abused.

On 2021-06-05 7:54 p.m., Khalid Baheyeldin wrote:
If you think it is a heating TMC5160 then that would be very odd. They don't heat up at all, specially when the motors are set to way less than the huge maximum of the TMC5160.

If you have an infrared thermometer, you can point it at the driver, provided it is not too shiny, and get a temperature reading. Do that before you do the fan thing.


John Petterson
 

On Sat, Jun 5, 2021 at 07:45 PM, Dave Schwartz wrote:
Maybe John didn't install the heatsinks.

John spent 20 years working with radios and other electronics while employed at Motorola, John understands heat sinks and installs the ones that come with electronic equipment.  :-).  While I am not a EE, I did lead the team that researched, diagnosed, and designed repairs for field issues in the two way radio products.  I have some kills in that area after doing that for several years.  I won’t say I am 100% perfect, and I am not offended by being asked if I did something stupid (because I am not immune to those wrong steps), but I am usually pretty good about getting things right.

It will be hard to get an angle to check the temp.  The FYSETC is now at the bottom of a box that has the daughter card mounted above it.  There are four 30mm spacers that separate the two boards and mount them inside the 3D printed box.  There are some holes in it (to connect USB cables to the various boards), but it does get a little bit warm after an hour of running.  And I cannot actually see the heat sinks without removing the screws and lifting the daughter card off.  I will ty to post a picture.


John Petterson
 

Here is the box.



And when I take the cover of it looks like this..



The 3mm screws in the corners mount it, taking those of and flipping the board over exposed the FYSETC.



And if I zoom in on the driver's we see this.



The


Ken
 

Hi John,

I am not saying we have the exact same issue here but.....this issue sounds familiar to what I am experiencing in my FYSETC S6 v2 / TMC5160 from BigTreeTech (v1.2) journey. The motors work (planetary) for a while and then stop until a power cycle of OnStep. When I say work, I mean 3 star slews to align and then goto/tracking works for ~1-2 additional objects. Then the issues start. The ALT motor stops first after random time in use passes and then occasionally the AZ one goes but usually I reset everything manually and reboot before getting too frustrated. I am using build 4.23 and my setup is moving a DOB so it is Alt/Az not EQ. I do not have my setup connected to CdC or Sky Safari to check similar behavior there.


Drew 🔭📷🚴‍♂️
 

On Sun, Jun 6, 2021 at 03:03 AM, John Petterson wrote:
Here is the box.
I see a box that needs a fan. I use small 12V fans. I hook them up to the 5V supply. That runs them slow and quiet.


John Petterson
 

Ken,  It sounds very similar.  I found a 40mm square 24v fan that I had as a spare part for a 3D printer and cut some holes in the back of the case to mount it.  With that installed the combination ran for 4 hours this afternoon with no problems, but it sounded like a jet plane taking off.  I found a quieter 24V fan and I tied running it on 12v but that caused the steppers to stop almost immediately so I went back to 24V. I have some even quieter fans, and I will try using two (one blowing in and one out) at 12V. But at least I am pretty sure I have found both the issue and a solution.  Thanks for the discussion all.

John


Drew 🔭📷🚴‍♂️
 

Still, my TMC5160's run cool. What are your settings in config.h?


Khalid Baheyeldin
 

On Sun, Jun 6, 2021 at 07:36 AM, Ken wrote:
this issue sounds familiar to what I am experiencing in my FYSETC S6 v2 / TMC5160 from BigTreeTech (v1.2) journey. The motors work (planetary) for a while and then stop until a power cycle of OnStep.
That is really odd.
I also use the FYSETC S6 v2 and the BTT TMC5160 v1.2, but with small motors.
IRUN is 675 mA, and IGOTO is 850 mA.

Lasts for many hours of slewing and tracking without incident, as it should ...


Drew 🔭📷🚴‍♂️
 

On Mon, Jun 7, 2021 at 12:30 PM, Khalid Baheyeldin wrote:
Lasts for many hours of slewing and tracking without incident, as it should ...
Likewise, mine with similar settings operate for hours without error. Also, like yours they operate coolly. This is why I think his config.h settings are a factor here.


John Petterson
 

IRUN is 900 and IGOTO is 1400.  Slews are pretty quick at 9 degrees per second (150% of the 6 degrees I requested). Setting iRun lower nets missed steps in tracking, with the 2130s I used to be able to hear those missed steps,  Now with the 6150 it just loses steps pretty much silently but I can watch the belt and pulleys bounce backwards every few seconds while tracking.  I have not had time to investigate what mechanical issue causes that.


Khalid Baheyeldin
 

On Mon, Jun 7, 2021 at 02:41 PM, John Petterson wrote:
IRUN is 900 and IGOTO is 1400.  
That well within the low side of their rating. They have a 4,400 mA maximum, which means that the practical limit is 3,500 mA.
You are way below that.

So that does not make sense ...

Setting iRun lower nets missed steps in tracking, with the 2130s I used to be able to hear those missed steps, Now with the 6150 it just loses steps pretty much silently but I can watch the belt and pulleys bounce backwards every few seconds while tracking.
That is also odd.
If you can make a Youtube video of that motion, that would be useful to others. Something to watch for.


John Petterson
 

I will record it some time here...  If I don't get it quickly, remind me in a couple of months when I am set up at the new house....


John Petterson
 

The 24v fans I had around were too loud at 24v and did not move enough air at 12v, I picked up a Noctura 12v fan and put that in today.  These are the best I have found (I use them on the 3D printer), lots of airflow and quiet.  Testing now to see how it does in the case keeping the drivers cool, but I expect it to be fine.  The case looks a little rough where I dremmeled out the air slots, but it will do for now.  One of these days I will print final versions of all my cases.

I got the final version of the ESP SHC design with my 5V power fix, I will take an hour and put one together.