Fixed that one too.

But having 3 places to update will cause them to go out of sync eventually (as they did here).

Howard, can you make the latest spreadsheet be without any extension, as it used to be, then the older ones have a version?

So, ... assets/downloads/spreadsheet.xls is always the latest version. That way the three pages link to the latest one.
Then as you update it for the next version, you rename assets/downloads/spreadsheet.xls to assets/downloads/spreadsheet118.xls?

The correct link is on the Construction page. It is to version 1.17.

I fixed the other one so they are the same now.

Um...where is the v1.16 spreadsheet?  The Wiki, Uploading the Firmware, still points to the v1.12 download.

About your spreadsheet values.

First you are using an old spreadsheet. Download the new one (version 1.16).

Second, you have too much (virtual) gear reduction there.
The StepsPerDegreeAxisX are usually below 61,200 with a tested maximum of 102,400.
You are at 160,000.

The new spreadsheet version will request that you enter the desired slew rate, and it will calculate the slowest/fastest uSec/step rate.

Should look like this I think... the encoders are 250 line so 1000 count right?  I put the base slew rate at 2.0 deg/s and max is 4.0 deg/s on an STM32 (without using pulse mode which probably isn't advisable in this case.)

On Fri, Jan 10, 2020 at 07:50 PM, --Oz-- wrote:

Q1: What driver do i choose in the OnStep configuration? (there is no mode adjustment using the misan/dcservo)
(DRV8825, A4988, LV8729, RAPS128, ST820, S109, TMC2208)

None of the above.

In your case, you will use one of the GENERIC or SERVO external drivers.

Look through the Models.h file, lines 27 to 34, and choose the one that best matches the dcservo stuff.

Q2: I am hoping to get 5 degrees per sec slew rate, but could live with 3 (just may put up with 2).

Having 3 deg/sec is very adequate, specially when you will be doing mainly conventional astrophotography.
That is, you only slew between objects a few times per session.
Even 2.5 deg/sec is adequate.

Here is some numbers from the excel (I am using the stm32 processor)
maxR   Slew   rpm     Khz
1          6.25    4687   31
2          3.13    2343   15.6
3          2.08    1562   10.4
5         1.25     9.37    6.3

The motors are fine at these rates (but the pro-mini may not be, I am guessing the stm32 is fine at these speeds)
Here is the notes in the excel file, am I already doomed?  :)

General MaxRate recommendations:

Mega2560 >= 32,  Teensy3.2 >= 16

Teensy3.5 >= 12,  Teensy3.6 >=

The figures for MaxRate is uSec/step (not kHz), and based on the desired maximum slew rate that you enter, the spreadsheet will calculate a slowest and fastest rate for you. For the STM32, it can do 20 uSec/step. If the slowest/fastest are within that range, then you are good

Good question about the Pro Mini. Don't know the answer. I guess you will know when you try things out.

As i am first trying servo motors and the misan/dcsero on the arduino pro mini with Gemini servo motors with 250 line encoders with 25:1 gearbox and 180t main gear.

Q1: What driver do i choose in the OnStep configuration? (there is no mode adjustment using the misan/dcservo)
(DRV8825, A4988, LV8729, RAPS128, ST820, S109, TMC2208)

Q2: I am hoping to get 5 degrees per sec slew rate, but could live with 3 (just may put up with 2).
Here is some numbers from the excel (I am using the stm32 processor)
maxR   Slew   rpm     Khz
1          6.25    4687   31
2          3.13    2343   15.6
3          2.08    1562   10.4
5         1.25     9.37    6.3

The motors are fine at these rates (but the pro-mini may not be, I am guessing the stm32 is fine at these speeds)
Here is the notes in the excel file, am I already doomed?  :)

General MaxRate recommendations:

Mega2560 >= 32,  Teensy3.2 >= 16

Teensy3.5 >= 12,  Teensy3.6 >= 8

I got the stm32 kit C, and I have say a few words:
Quick shipment to California, excellent packing, assembled nice and quick. I am impressed with everything, great job guys. 

Now to the more difficult programming for me. 

As I mentioned, I am going to try servo motors first (as i have them already mounted, so here is my next question:

I will be using the https://github.com/misan/dcservo 
FW with arduino and was going to use A4950 motor driver

It looks to me to be compatible with the L298 servo driver, what do you guys think?
Chopper mode is just current limiting mode and should not be a factor in servo motor mode.

I think the normal TMC5160 is $15 I think, so saving $6 for two by making two adapters is a question of time. Let us know how it works out.

The new driver (forgot the number) is $20 each.

On Sat, Dec 28, 2019 at 05:33 AM, Tong wrote:

Here is the low-cost TMC5160 for $12. The seller promised me to mail the test product to me on Monday

Note that this is not in the StepStick form factor.
Therefore, it cannot plug directly in any of the current OnStep boards.

You may still get it to work on a separate custom Perfboard, but you have to run wires from your OnStep board to that.

I found a box of stepper motors:

85x85x66mm 1.8 degree 2.4v 4.5a
56x56x75mm ?? (with driver)
42x42x40mm 1.8 degree 2.8v 2.0A (with encoder / driver)
42x42x23mm 1.8 degree 2.0v 2.0A (with encoder / driver)


The bottom three have some electronics in them, some of the components are:
IRFB4332PbF 42A FET x2
80mΩ current sense resistors x2,
8-Bit Parallel DAC,
ATtiny13 mpu (atmel 0514D),
something that looks like a memory chip (s29al008070tf102) , 
large 270uH inductor plus a couple 47uF caps for LC filter.

If I can't figure out the DB9 connector signals, I could just bypass the electronics and use external driver if needed (but lower 1.8 degree steps).



On Ali I seen this this low cost one (only 32 micro steps)
https://www.aliexpress.com/item/32957327440.html?src=google&src=google&albch=shopping&acnt=494-037-6276&isdl=y&slnk=&plac=&mtctp=&albbt=Google_7_shopping&aff_platform=google&aff_short_key=UneMJZVf&&albagn=888888&albcp=1582410664&albag=59754279756&trgt=743612850714&crea=en32957327440&netw=u&device=c&gclid=Cj0KCQiAxfzvBRCZARIsAGA7YMzfdTVNbAtH6e9YEBCeZfyZyKI7wgOlp247sMTEj76MG5G_P1jpXb4aAp6HEALw_wcB&gclsrc=aw.ds

On Fri, Dec 20, 2019 at 01:03 PM, Khalid Baheyeldin wrote:

Again and anecdotal observation: I found no practical difference between TMC2130 and LV8729 when testing star trails.

I didn't really find a measurable improvement in tracking with my G11 when switching from a DRV8825 to the TMC2130 either.  And that's at 360:1 where you might actually see it.  Not something like say 3240:1 where the gear reduction makes the motor motion control far less critical.  Perhaps I wasn't looking hard enough though.

No model was better than the other tracking wise. The TMC2130 was quieter for sure, but the LV8729 won because of far less heat dissipation.

Since my four mounts use TMC2100's or TMC2130's (both have similar heat dissipation characteristics) I have to say that this was never an issue for me.  My EM10 mount with the MPJA NEMA11 motors, for example, doesn't even have heat-sinks on the drivers.  The much bigger G11 w/ 0.85A motors worked fine back when its controller had stock heat-sinks on the drivers and they never ran "hot."

On Fri, Dec 20, 2019 at 01:03 PM, Khalid Baheyeldin wrote:

What I meant is specifying 1/256 in the configuration file for OnStep, not interpolation: no one (I know) uses 1/256 for any driver model.

For "torque and angle accuracy reasons" you wrote so I figured interpolation applied, at-least somewhat.  And no, they don't do so used often because TMC's can't mode switch [on the fly] and also usually OnStep max steps per degree would be exceeded.  If running my G11 360:1 400 step motors with LV8729's I'd definitely use 128x mode, there is no reason not to.

On Wed, Dec 18, 2019 at 02:22 PM, Howard Dutton wrote:

On Wed, Dec 18, 2019 at 09:48 AM, Khalid Baheyeldin wrote:

Almost no one runs at 128 or 256, for torque reasons and angle accuracy reasons.

Almost everyone running the TMC5160, 2130, 2209, 2208, and 2100 takes advantage of 256x mode since the driver automatically interpolates lower modes to 256x for smooth vibration free motion.

What I meant is specifying 1/256 in the configuration file for OnStep, not interpolation: no one (I know) uses 1/256 for any driver model.

Again and anecdotal observation: I found no practical difference between TMC2130 and LV8729 when testing star trails. No model was better than the other tracking wise. The TMC2130 was quieter for sure, but the LV8729 won because of far less heat dissipation.

As the TMC5160 gets more popular, it could be the driver of choice, since it has all the features:
- high current (if needed)
- 256 interpolation
- low heat dissipation
- configurable from OnStep

On Fri, Dec 20, 2019 at 12:49 AM, --Oz-- wrote:

On a side note, where to find the lowest cost TMC5160 drivers? (anyone, link plz, tia!).

They are relatively new, so the price has not dropped much yet.

Here is the cheapest eBay listing that I can find, for now.
You can also try AliExpress.

On Wed, Dec 18, 2019 at 03:38 PM, George Cushing wrote:

Yes, A4988s can be had for $0.50

The A4988 is an OK performer (no vibration), but being limited to 1/16 max, it is less flexible. If you need 1/32 (per the spreadsheet), you are out of luck. As you say, there are better options out there, for not too much money.

and DVR8825s for around a buck

The DRV8825 used to have an advantage in that it can drive more current for more torque.
But it is notoriously shaky, and inaccurate too.
No one should use it for new builds.

I ordered a couple days ago the stm32 blue pill kit-C, but havent seen a email to pay for it. I am not in a hurry, just wondering.

On a side note, where to find the lowest cost TMC5160 drivers? (anyone, link plz, tia!).