Topics

TMC2130

Joe
 

I purchased a different set of GT2 gears for a better drive ratio on my Vixen GP. I had purchased a couple of the TMC2130s before doing research that unfortunately for me were the wrong version. I was unable to solder those really small jumpers on the bottom.
With the 3.16 OnStep release, if I’m reading correctly the correct driver would improve performance with using SPI? My question being is it enough of an improvement over the LV8729s to warrant buying them? 
If the Watterott brand is the preferred variety I’ll spend the money, although the link from the wiki is in a different currency than USD, not sure what they actually cost. The page also mentions BigTreeTech which seems to be pretty reasonable cost wise. Basically I’d want TMC2130 V3.0 with SPI and all of the pins on the bottom? (For price comparisons)

Khalid Baheyeldin
 

On Thu, Feb 13, 2020 at 06:48 PM, Joe wrote:
I purchased a different set of GT2 gears for a better drive ratio on my Vixen GP. I had purchased a couple of the TMC2130s before doing research that unfortunately for me were the wrong version. I was unable to solder those really small jumpers on the bottom.
Soldering of these tiny components is very hard. Requires a steady hand, good eyesight, magnifier and experience.
Average people would like to avoid doing it, unless there is no other choice.

With the 3.16 OnStep release, if I’m reading correctly the correct driver would improve performance with using SPI? My question being is it enough of an improvement over the LV8729s to warrant buying them? 
SPI itself does not give an advantage.
What gives an advantage is all the Trinamic TMCxxxx drivers have 1/256 intepolation.
So if you select 1/16, it will internally divide them to 1/256 giving finer motion.

However, that may not be needed at all.

It depends on your steps per degree, as well as your pixel scale (focal length and camera sensor).
If steps per degree is high enough, and/or pixel scale is coarser, then it does not matter if it is a TMC or LV8279 (or S109 for higher amps).
If steps per degree is is low, and/or pixel scale is finer, then the TMC may have an edge due to interpolation.

I am using a fine pixel scale (0.47"/pixel = 1310 mm focal length and Canon T4i/650D) with 28,800 steps/degree, and the LV8729 performs beautifully. I could not detect any difference between it and the TMC2130 (other than silence).

If the Watterott brand is the preferred variety I’ll spend the money, although the link from the wiki is in a different currency than USD, not sure what they actually cost. The page also mentions BigTreeTech which seems to be pretty reasonable cost wise. Basically I’d want TMC2130 V3.0 with SPI and all of the pins on the bottom? (For price comparisons)
The eBay/AliExpress/Amazon non-Watterott drivers have been proven to work.
The problem is that there are many versions, some with SPI and some without.
Some have pins up only, some have pins up down only some have both up and down.
Some look like they have all what is required, but two extra pins show that they are SPI or not.
And depending on your board, you may need a certain.
Yes, there are too many variations, and yes, it is confusing, even to me.

Note that there is no shipping from China on many items.
I ordered a switch and did not get the shipping notice for about 10 days. I contacted the seller and they said the workers can't go to the warehouse. He will refund.

Joe
 

Thanks Khalid. I’m using a Canon 6DM2, it’s a full frame as opposed to a crop sensor. I believe the pixels on the 6D are a bit bigger than the T4i, but there are more if them, probably pretty close. I’ll try to hit that 28,800 mark and see how it goes. I’ve got some concerns about the inefficiency of my mount. Are you using the PEC with your photography? Do you think that should be a priority to get some decent results?



On Thursday, February 13, 2020, 7:09 PM, Khalid Baheyeldin <kbahey@...> wrote:

On Thu, Feb 13, 2020 at 06:48 PM, Joe wrote:
I purchased a different set of GT2 gears for a better drive ratio on my Vixen GP. I had purchased a couple of the TMC2130s before doing research that unfortunately for me were the wrong version. I was unable to solder those really small jumpers on the bottom.
Soldering of these tiny components is very hard. Requires a steady hand, good eyesight, magnifier and experience.
Average people would like to avoid doing it, unless there is no other choice.

With the 3.16 OnStep release, if I’m reading correctly the correct driver would improve performance with using SPI? My question being is it enough of an improvement over the LV8729s to warrant buying them? 
SPI itself does not give an advantage.
What gives an advantage is all the Trinamic TMCxxxx drivers have 1/256 intepolation.
So if you select 1/16, it will internally divide them to 1/256 giving finer motion.

However, that may not be needed at all.

It depends on your steps per degree, as well as your pixel scale (focal length and camera sensor).
If steps per degree is high enough, and/or pixel scale is coarser, then it does not matter if it is a TMC or LV8279 (or S109 for higher amps).
If steps per degree is is low, and/or pixel scale is finer, then the TMC may have an edge due to interpolation.

I am using a fine pixel scale (0.47"/pixel = 1310 mm focal length and Canon T4i/650D) with 28,800 steps/degree, and the LV8729 performs beautifully. I could not detect any difference between it and the TMC2130 (other than silence).

If the Watterott brand is the preferred variety I’ll spend the money, although the link from the wiki is in a different currency than USD, not sure what they actually cost. The page also mentions BigTreeTech which seems to be pretty reasonable cost wise. Basically I’d want TMC2130 V3.0 with SPI and all of the pins on the bottom? (For price comparisons)
The eBay/AliExpress/Amazon non-Watterott drivers have been proven to work.
The problem is that there are many versions, some with SPI and some without.
Some have pins up only, some have pins up down only some have both up and down.
Some look like they have all what is required, but two extra pins show that they are SPI or not.
And depending on your board, you may need a certain.
Yes, there are too many variations, and yes, it is confusing, even to me.

Note that there is no shipping from China on many items.
I ordered a switch and did not get the shipping notice for about 10 days. I contacted the seller and they said the workers can't go to the warehouse. He will refund.

Khalid Baheyeldin
 

On Thu, Feb 13, 2020 at 07:56 PM, Joe wrote:
I’m using a Canon 6DM2, it’s a full frame as opposed to a crop sensor. I believe the pixels on the 6D are a bit bigger than the T4i, but there are more if them, probably pretty close.
If I use Canon 6D on this calculator, with my focal length (1310 mm) it gives me 1.03"/pixel.
Not sure of the Mark II is different.

I’ll try to hit that 28,800 mark and see how it goes.
The difference between the recommended steps per degree in the spreadsheet (12,800) and mine (28,800) is close to double.
So maybe with your pixel scale being about double it will even out.
And if you are using a scope at 800mm or 1000mm, then it is even more likely that an LV8729 would do the job more than adequately.

I’ve got some concerns about the inefficiency of my mount. Are you using the PEC with your photography? Do you think that should be a priority to get some decent results?
I used PEC when I was not autoguiding. It was all an experiment to postpone autoguiding.
It helped, and so did the Full Compensation Tracking.
But when I switched to autoguiding, I found out that PEC makes no difference.
If the periodic error in your mount is gradual, then autoguiding may be sufficient to take care of it.
If it is too high and happens over a short period of time, then maybe in this case PEC will help.

So, my recommendation is not to bother with it, if you are going to try autoguiding anyway.
If you inspect the graph and see that there is a period, then at that point consider trying PEC.
You can do it without a sensor initially (just record it every session, or park the scope), and if it is helpful, then you can consider adding a sensor.

Howard Dutton
 

On Thu, Feb 13, 2020 at 04:56 PM, Joe wrote:
I’ve got some concerns about the inefficiency of my mount.
I'd like to hear what the problem is before guessing at a solution.

Joe
 

Howard, 
In order to adjust the drive to the worm gear, on this mount you need to remove the steppers to access the set screws to adjust the tension. I’m trying to find some balance between smooth movement and tightness. Too loose and I have noticeable play when pushing the mount by hand, too tight and I am stressing the steppers. It’s a process to move even 1/4 turn on the set screws, it’s very time consuming. It seems I should be doing this process with the mount fully loaded as it changes with the weight dynamics. I believe I need a much lighter scope for this mount, but that’s another problem. Also while I don’t see much if any wear on my worm gears I’m aware that poor worm/gear engagement is a concern when thinking about precision, thinking ahead if PEC could help out...I did disassemble the mount, remove the old grease and used the recommended “superlube”.
If my goal is to achieve 28800 steps per degree, I’m using 16 & 40 teeth gears with a 144 worm I get: 128M/S 51200SPD, 64=25600, 32=12800. Is 128M/S ok for the LV8729 .9 steppers? The RA does move very smooth and quiet with these settings. Haven’t tried outside under the sky yet. 
The slew rate is another story. The first time I set this up was using Khalids calculator and his auto generated Config.h file.
I see that the MaxRate is now DesiredBaseSlewRate. If the STM32 is 20 you enter that number in the Config.h file and also add a lower number in Microsteps_Goto or just leave it blank? I’ve tried multiple combinations and am getting painfully slow skew rates or bad noises and skipping from the steppers. It’s likely I’m misinterpreting what values are actually needed.






On Friday, February 14, 2020, 2:16 PM, Howard Dutton <hjd1964@...> wrote:

On Thu, Feb 13, 2020 at 04:56 PM, Joe wrote:
I’ve got some concerns about the inefficiency of my mount.
I'd like to hear what the problem is before guessing at a solution.

Khalid Baheyeldin
 

On Fri, Feb 14, 2020 at 04:05 PM, Joe wrote:
If my goal is to achieve 28800 steps per degree, I’m using 16 & 40 teeth gears with a 144 worm I get: 128M/S 51200SPD, 64=25600, 32=12800. Is 128M/S ok for the LV8729 .9 steppers?
You probably do not need 28,800, and 12,800 is sufficient, given your pixel scale.
See my previous reply in this same thread for details.

The slew rate is another story. The first time I set this up was using Khalids calculator and his auto generated Config.h file.
The online configuration generator applies to 2.22 only.

I see that the MaxRate is now DesiredBaseSlewRate. If the STM32 is 20 you enter that number in the Config.h file
If you are using 3.16, then it is SLEW_RATE_BASE_DESIRED, and with 400 step motors, and 1/32 microsteps, you should not have a problem with 3.0 degrees/second.

and also add a lower number in Microsteps_Goto or just leave it blank?
Try first with AXISx_DRIVER_MICROSTEPS_GOTO set to OFF, and see if 3 deg/sec is adequate. It probably is.
If not, then try 4 or 8 for that value.
Again, all this is 3.16.

I’ve tried multiple combinations and am getting painfully slow skew rates or bad noises and skipping from the steppers.
That means the motor is stalling. Try the above (OFF then 4 or 8) and see if things work. Again, with 3.16.
 

Howard Dutton
 

On Fri, Feb 14, 2020 at 01:05 PM, Joe wrote:
In order to adjust the drive to the worm gear, on this mount you need to remove the steppers to access the set screws to adjust the tension. I’m trying to find some balance between smooth movement and tightness. Too loose and I have noticeable play when pushing the mount by hand, too tight and I am stressing the steppers. It’s a process to move even 1/4 turn on the set screws, it’s very time consuming. It seems I should be doing this process with the mount fully loaded as it changes with the weight dynamics. I believe I need a much lighter scope for this mount, but that’s another problem.
EQ5 class mounts are only so good.  Fact of life.  Perhaps just me but I don't get too OCD about tight worm/wheel engagement.  More important to be sure the worm doesn't press against and bind with the wheel anywhere in the 360° travel.  Then when imaging be sure you have the weights East biased.  Even Dec can be done similarly.

Also while I don’t see much if any wear on my worm gears I’m aware that poor worm/gear engagement is a concern when thinking about precision, thinking ahead if PEC could help out...
Khalid answered this.

I did disassemble the mount, remove the old grease and used the recommended “superlube”.
Good move.

If my goal is to achieve 28800 steps per degree, I’m using 16 & 40 teeth gears with a 144 worm I get: 128M/S 51200SPD, 64=25600, 32=12800. Is 128M/S ok for the LV8729 .9 steppers?
Higher microstep doesn't hurt but doesn't help much at all (vanishingly small improvements.)

The RA does move very smooth and quiet with these settings. Haven’t tried outside under the sky yet. 
The slew rate is another story. The first time I set this up was using Khalids calculator and his auto generated Config.h file.
I see that the MaxRate is now DesiredBaseSlewRate. If the STM32 is 20 you enter that number in the Config.h file and also add a lower number in Microsteps_Goto or just leave it blank? I’ve tried multiple combinations and am getting painfully slow skew rates or bad noises and skipping from the steppers. It’s likely I’m misinterpreting what values are actually needed.
Not enough info to really answer this.

I suggest using release-3.16 and the new config.h file format.  It's a lot better and much easier to understand.

Joe
 

Thank you both for your replies. Yes I am on 3.16 and it’s a big improvement!
Using 8 for driver_microstep_goto works perfectly. Slew speed is great and plenty of tork, couldn’t stop it by hand at all. It’s like a different mount now. Makes some stepper noise, but it’s the good kind... Almost like a musical scale when it winds up.
...Of course now it’s going to be cloudy for two months...
Thanks again!




On Friday, February 14, 2020, 5:26 PM, Howard Dutton <hjd1964@...> wrote:

On Fri, Feb 14, 2020 at 01:05 PM, Joe wrote:
In order to adjust the drive to the worm gear, on this mount you need to remove the steppers to access the set screws to adjust the tension. I’m trying to find some balance between smooth movement and tightness. Too loose and I have noticeable play when pushing the mount by hand, too tight and I am stressing the steppers. It’s a process to move even 1/4 turn on the set screws, it’s very time consuming. It seems I should be doing this process with the mount fully loaded as it changes with the weight dynamics. I believe I need a much lighter scope for this mount, but that’s another problem.
EQ5 class mounts are only so good.  Fact of life.  Perhaps just me but I don't get too OCD about tight worm/wheel engagement.  More important to be sure the worm doesn't press against and bind with the wheel anywhere in the 360° travel.  Then when imaging be sure you have the weights East biased.  Even Dec can be done similarly.

Also while I don’t see much if any wear on my worm gears I’m aware that poor worm/gear engagement is a concern when thinking about precision, thinking ahead if PEC could help out...
Khalid answered this.

I did disassemble the mount, remove the old grease and used the recommended “superlube”.
Good move.

If my goal is to achieve 28800 steps per degree, I’m using 16 & 40 teeth gears with a 144 worm I get: 128M/S 51200SPD, 64=25600, 32=12800. Is 128M/S ok for the LV8729 .9 steppers?
Higher microstep doesn't hurt but doesn't help much at all (vanishingly small improvements.)

The RA does move very smooth and quiet with these settings. Haven’t tried outside under the sky yet. 
The slew rate is another story. The first time I set this up was using Khalids calculator and his auto generated Config.h file.
I see that the MaxRate is now DesiredBaseSlewRate. If the STM32 is 20 you enter that number in the Config.h file and also add a lower number in Microsteps_Goto or just leave it blank? I’ve tried multiple combinations and am getting painfully slow skew rates or bad noises and skipping from the steppers. It’s likely I’m misinterpreting what values are actually needed.
Not enough info to really answer this.

I suggest using release-3.16 and the new config.h file format.  It's a lot better and much easier to understand.