Date   

Re: Problem compiling SHC code...python serial module missing?

Khalid Baheyeldin
 

Not a SuSE fan ... but I have been using Linux exclusively for a decade and a half ...
Ubunut it is for me ...

In the past few years, when using 'python' it usually means some python 2.x (usually 2.6 or 2.7).
If you want Python 3 (usually 3.6, 3.7, 3.8), then it is invoked via 'python3'.

Package names are the same:
python-whatever means the whatever package for python 2.x
python3-whatever is the 3.x variant.


Re: STM32 Blue Pill and focuser with 28BYJ-48 Stepper

Khalid Baheyeldin
 

On Sat, Aug 8, 2020 at 12:55 PM, George Cushing wrote:
The red conductor is connected inside the case to both coil sets, shorting bipolar signals. The thin coil wires are appear to be joined to the PCB at two points. The Red wire's connection to both coils may be broken by cutting this trace.
There are several web pages and Youtube videos explaining how to convert a 28BYJ unipolar stepper to be bipolar.

Here is one example https://www.youtube.com/watch?v=ROFjf_7IrRA


Re: 1986 vixen super Polaris. #celestron #takahashi

George Cushing
 

I've converted several EQ-5 types the SP, depending on the type, is a little more challenging than most. I yours is a C8-SP and you don't have this part. . . 


Then the usual EQ-5 approach is a bit more difficult. If I had to do another early SP I'd machine a duplicate of the part. out of 5/16" (8mm) alloy. I eventually got something that worked but I really was't happy with it.





Re: STM32 Blue Pill and focuser with 28BYJ-48 Stepper

George Cushing
 

No, I don't think that will work: 



The red conductor is connected inside the case to both coil sets, shorting bipolar signals. The thin coil wires are appear to be joined to the PCB at two points. The Red wire's connection to both coils may be broken by cutting this trace.


Re: Problem compiling SHC code...python serial module missing?

jsalbinson@...
 

Further to my notes on opensuse tumbleweed, here are my notes on opensuse leap15.2
It seems that python3 is installed, but not pyserial v3. I used yast to install pyserial v3, but the Arduino IDE
did not compile the SHC ESP32 code, still not finding 'python' in the $PATH. A "cnf python" suggested installing python-base, which seems to refer to python2.
So I followed the suggestion: sudo zypper install python-base
and accepted the additional library to install. Still no joy on the Adruino IDE ESP32 compilation, so back to yast to install python2-serial.
This did the trick and I now have a clean compile of the SHC ESP32 code.
Now to build the system on a breadboard, and test it against the telescope test rig!
I hope these notes may be useful if you are an opensuse fan, as I am.
Regards, James Albinson


Re: Mks gen_l v2.0 pins

John Petterson
 

You are lucky. The last time I was wiring my setup when I was low on sleep I let the magic smoke out of both a MKS Gen L and a Teensy 3.2.  Expensive lesson that sleep is more important than getting it done sooner....

John


Re: Mks gen_l v2.0 pins

Eric Esch
 

Never mind... i am special sometimes... i found the picture right where it should be.. i just didn't read close enough. Wow.... i need sleep. 


Re: stand alone operation

Gary McLean
 

Hello all

I found on both kits I ordered that the Power Jack pins (3) where just a little bit to wide or

The holes where a  little bit to small to insert into the  pcb. So I increased the hole size a little.

Iam sure this somehow broke the ground connection to the  PCB ground from the Power Jack

So that allowed the ground  to float.

I did try to trim metal tabs on the Power Jack but was worried that I might break the jack.

The fix was to jumper Power Jack Gnd to the PCB Gnd passthru hole.

If others have had a hard time inserting the jack what was your fix?

Let me know if this fix is a bad idea.

Thanks

Gary

 

Sent from Mail for Windows 10

 

From: George Cushing
Sent: Friday, August 7, 2020 12:08 PM
To: main@onstep.groups.io
Subject: Re: [onstep] stand alone operation

 

No, you need to let the current flow through the LED to ground via R1. No R1, no complete circuit.


If you are using this regulator... 

HW-813 you have to cut the trace to adjustment pot and bridge the 5V contacts. The enable pin is not used and the surface with the contacts should be mounted up.

If you have this regulator....


You simply solder it in place. Voltage on Vin and Vout should be checked with reference to ground. 

 


Mks gen_l v2.0 pins

Eric Esch
 

I am having some difficulty figuring out which pins to hook up where. 

I had a mks gen_l v2 already so figured I would start with that and get a better understanding of onstep then build a faster one. I have the tmc2130 drivers already i robbed from a 3d printer here. 

The add ons I am attempting are:
Wemos d1 mini pro
Gps neo-6m
Bme280
Speaker

The wemos I tried to follow the guide and hooked up to the z axis limit serial and it would only blink. I then moved it to aux 1 in the top row and it blink and then goes solid, but i do not see the access point of onstep ever broadcasted. 

I want to see if i can get some clarification on which pins I should be using. The bme280 will have to be over spi but its lables do not match any description I can find for the mks board.  Any help would be appreciated.  Also clarification on which connectors to use on the mks board as it calls them aux 1, exp 1, exp 2


Re: Too High WORMROT causes timeout to serial

Khalid Baheyeldin
 

This all shows that PEC is for conventional equatorial mounts, and not intended for the multi step worm-wheel-less design you have.

For example, taking 8 hours to record PEC is not practical at all. The meridian will be crossed during that time, and a meridian flip should be performed, interrupting the recording process, and so on.

And as you can see, this is why some designs insist on direct driving the worm, and having only one stage of reduction (the worm wheel + worm gear), usually at 360:1 (e.g. Losmandy G11). This is to avoid non periodic errors in GR1.

The lower end practical equatorial is the EQ3-2 with 130:1 in GR2. So in your case, 3:1 or even 5:1 is beyond the conventional designs, and therefore will stumble across such limitations.

Oh, and I see that Howard revised the Config.h wiki page to reduce the maximum for steps per worm rotation to 862,920.


Re: Too High WORMROT causes timeout to serial

snorowski@...
 

Dave,

Hi. Yes, that is one thing I was doing , playing around with placing different gear combinations in GR1 and GR2, then I realized I probably wouldn't utilize PEC anyway since I'll
be using PHD2 or metaguide.     My 1215:1  gear reduction comes from 6  stages actually:   5:1  3:1  3:1   3:1   3:1  3:1      toothwise its 90:18  90:30  90:30  90:30  90:30 90:30.
Im working with what I've got, thats why I chose those pulleys.  The only thing I needed to buy for the whole mount were 12 belts @ $4 each and some hardware. Steppers were on hand , too. Using 200 steps, 32 micro.  21600 steps/degree.

Now back to my initial observation and original post:  If STEPS_PER_WORMROT  > 590,000,  the STM32 (maybe other build platforms, too?) will not communicate through the cp2012 or wemos.

So, hypothetically, as you suggested, in order to compensate for eccentricity on the final driving pulley, a ratio of 3 would be entered into GR2, GR1 would take up the remaining 405:1 reduction.
Using Axis_Steps_per_degree = 21600,    Steps_Per_WORMROT  is still 2,592,000,  leading to serial timeout.  Also it would take 8 hours to record a full rotation of the driving gear.
Similarly ,I could try to put the sensor on the second to last driving gear.  GR2=9,  GR1=135,  Steps_per_WORMROT = 864,000.  Still get a serial timeout. And take 2.6 hours to record for PEC.
I would have to put the sensor on the 3rd from the last driving gear.   GR2=27  ,  GR1=45,    Steps_per_WORMROT = 288,000,  finally something OnStep can handle.   And PEC recording would be almost an hour,  but 1 hour on that gear would mean missing 1.6 hours of eccentricity not compensated for on the next gear and so on.
Rather than deal with all that, I decided to just rely on guiding from PHD2.  But hey, maybe I get lucky and my pulleys will be perfect.

We could go on forever about what my best options were and how I should or shouldn't have built my mount.  I just wanted people to be aware of the problem with using too high a steps_per_wormrot.
And if they have  a  build that is indicating tracking but not communicating, they should check to make sure the WORMROT steps are not too high by accident, even though the configuration page says they can be as high as 22M.    


Re: Too High WORMROT causes timeout to serial

Ufuk Sandalci
 

I built a four-stage belt (MXL) only GEM and using it for a year now. I assumed and set the first stage (6x) as GR1 and the rest three stages (288x) as GR2. I don't bother with PEC issues and recently using PHD2. I haven't encountered a problem yet.  Someone from Avalon (belt only GEM company) once mentioned that the first stage was the most important so far the PE was concerned. Maybe in the future, I will test PEC. But meanwhile, the variation of belt tension in the final stage is more of a problem. Auto-guiding has solved it but it is there.


Re: STM32 Blue Pill and focuser with 28BYJ-48 Stepper

 

On Fri, Aug 7, 2020 at 02:27 PM, Dave Schwartz wrote:
Are you saying you would like to petition Howard to make this a configuration option? Would be very easy to do.
Sure--I hate counting to '5' 


Re: Failed to init device yet again

Mike Ahner
 

Martin I haven't done much engineering design, I was on the repair side. But I seem to remember that resistors like these were merely current limiters installed to protect the (usually) very low rate drive current the MPU was designed to carry. In this case, the need is to directly tie to VDD or directly short to ground. Perhaps there is a benefit for static discharge, but I would rank that very low probability.

The other interesting point you raised is that for most logic circuits, you don't need nor seldom achieve full 5v+ or 3.3v+ or ground as a logic level signal. If you measure with an O-scope, you'll find most are a 1/10 or 2/10 vdc or so short & the specs are always much lower requirement as you pointed out. For example, 5v logic used to "HIGH" at 3.75 & above, or "LOW" at less 1v or so. I don't know what today's value requirements are, but I'm sure they are lower than we usually think.

Having said all of that, one very valid reason that occurs to me for current limiting on such pins, would to be for battery life. So many modern devices are intended for battery power and even if not, the power savings in worldwide quantities would be substantial.

I appreciate your time to calculate, loose ends like this always bother me but seldom do we have the chance to find the answer these days. I too, believe the part is either defect, wrong rating, or poorly soldered. While the mpu itself my be genuine, the board is still manufactured by some other entity and Gerry says, "Things happen"  


Re: Failed to init device yet again

Martin Bonfiore <marsbonfire0@...>
 

All of this begs the question why the resistors in the first place and not a direct short?   Two possibilities come to mind.  One is that,when the jumper is in place it is possible for something else to drive the chip pins without fighting either ground or VDD...just needs to drive 100K.  The other is more esoteric and probably not the case; the 100k make those pins more safe relative to static discharge...these are nodes that are touched by highly charged astronomer fingers.  So...shorting the resistor(s) is probably not a problem in this application and certainly not a problem as a temporary troubleshooting technique.


Re: STM32 Blue Pill and focuser with 28BYJ-48 Stepper

Dave Schwartz
 

On 2020-08-07 5:03 p.m., Richard Shagam wrote:
On Fri, Aug 7, 2020 at 07:28 AM, Dave Schwartz wrote:

Increasing the value of AXIS4_LIMIT_MIN_RATE will make the
slow-rate movement faster.

Are you saying that increasing the value from, say 10 to 50 increases the step rate? Or do I have that backwards?
Increasing that setting will speed up the 'slow' rate... that is, it sets the minimum micron/sec rate at which the focuser will move and OnStep will adjust the step rate to meet that according to the STEPS_PER_MICRON value you set previously.

No, the duration of the hold-down until it switches to fast mode
is not adjustable.
Isn't that delay time set anywhere in the program?  Seems like it must be programmed somewhere.
It is hardcoded in the button-down time comparison "buttonPad.F.timeDown()>5000" and "buttonPad.f.timeDown()>5000" on line 523 and 524 of SmartController.cpp (line numbers from current master).

Are you saying you would like to petition Howard to make this a configuration option? Would be very easy to do.


Thanks


Re: 1986 vixen super Polaris. #celestron #takahashi

Rafael Barberá Córdoba
 

I’ve just transformed an Skywatcher EQ5 (almost the same mount) with an ESP32/CNC3, a pair of LV8729 and two 400 step 12v 1.68A motors. I’ve put a 15:60 pulleys for a 4:1 reduction. 


the advantages of the ESP32/CNC3 is not soldering requiered, so the electronics “builds” on minutes and almost without experience. 


Re: STM32 Blue Pill and focuser with 28BYJ-48 Stepper

 

On Fri, Aug 7, 2020 at 07:28 AM, Dave Schwartz wrote:
Increasing the value of AXIS4_LIMIT_MIN_RATE will make the slow-rate movement faster.
Are you saying that increasing the value from, say 10 to 50 increases the step rate? Or do I have that backwards?
No, the duration of the hold-down until it switches to fast mode is not adjustable.
Isn't that delay time set anywhere in the program?  Seems like it must be programmed somewhere.

Thanks


Re: STM32 Blue Pill and focuser with 28BYJ-48 Stepper

 
Edited

You are aware that the 28BYJ-48 is a unipolar design?
Yes, I am--but a very simple mod (pull off the blue cap and cut the red wire) makes it bipolar


Re: Too High WORMROT causes timeout to serial

Dave Schwartz
 

I would suggest that, even in your case, GR2 always be the final reduction stage.

In a worm gear set this would be the number of teeth on the ring gear (144, 180, 360, etc). In your all-belt system this would be the ratio between the number of teeth of the pulley fixed to the RA shaft and the pulley driving it. GR1 is the product of all ratios before that.

This makes a single rule for all.

It also means that if you need to implement PEC, the sensor would be mounted on the driving gear of the final stage. With this all-belt-reduction system, that would compensate for eccentricity of the driving pulley of that driving gear. There is no way to deal with the eccentricity of the pulley fixed to the RA shaft because its period of rotation is a full day and there's no way OnStep can record and save that.

Stephen: what are the tooth counts of the pulleys in the final stage? I'm also thinking of building an all-belt-reduction mount.

On 2020-08-07 4:38 p.m., snorowski@... wrote:
The gears work great, undetectable backlash,  I will be using PHD2, no index sensor, no PEC, that's why I made WORMROT= 0 in the end.

Again, my initial intent was to notify builders and coders of the possible timeout on the com port if a WORMROT value is over 590,000. Whether they intentionally are tracking a period that large or if they
like me, unknowingly  set it too high. And to not trust the configuration page that says recommended values can be as high as 22M.