Dynomotion

Group: DynoMotion Message: 9065 From: daveymahomh600e Date: 2/1/2014
Subject: Ribbon Cable Woes
Hi Tom,

I am just about to yank my last hair out. I posted last week about the issue I was having with my Konnect board and the suggested solution was to add an extra connector with a capacitor to kill suspected noise. I did do that, but the problem still exists.

Further to that, it has become impossible to get the servo motors setup and I suspect this may also be a noise problem. When I run the Autophasefindfast program, it comes up with wildly different numbers for all the values every time it is run. Because my system uses a SnapAmp, there is that 50 pin ribbon cable I have to deal with. It is also impossible to keep it under 12 inches in length and I don't know where to go from here. I have spent the past 4 days trying to tune the servos and gotten nowhere.

Any suggestions would be much appreciated.

Thanks,
David.
Group: DynoMotion Message: 9066 From: Wcarrothers Yahoo Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
First let me suggest trying to sound a little more positive in your message.  As I have not gotten more  expert help from anyone more so then from Tom.  The documentation is good but requires thought.  But try to imagine how much thought goes into a production machine like a techno befor me it runs well and is ready for the market.  You will get there

That said. I also am having my first go around with a k and snaps with brushless servos.  My other system was a k with a k analog board and moog brushes servos. So in that case the movement is done by a 3rd party if you will.   There were signals in that case which I did add caps to and works well and has for years

That said. Perhaps this will help.  My new setup using the snaps has servos with differential a and b signals. Those are connected no problem but my index signal( 1 pulse per turn) was a problem.  It turned out to not be a differential signal.  It also did not provide a good signal to the lvttl inputs.  I switched to into that signal using the opto inputs and it now works great

The next step is to be sure the settings for that axis are uploaded to the k in the config and flash screen.  This is a step that I always forgot. And also the perspective for the upload and download (with respect to the pcs possition is all wrong for my mine. But that isn't his fault. Just something to pay attention to. ) I would prefer the buttons on say send to kflop and get from kflop. But that is just me

Anyway if these settings are not right things will not work.  The find c program will but a regular  axis control program will not work if this step is missed.   Perhaps will look like it is trying but will not be smooth


Now for the wiring.  Yes tenth sender pin outs for everything.  He really has no other practical choice but to do it that way as to give a screw terminal for all the io he gives would result in a huge board which would also be adding antennas perhaps to pickup more noise (ok perhaps a stretch there). But in anycase most systems would use a lower percentage is he connections so connect to the ones you need and not to he others

My solution was to use pololu connection cords. Becides being cheeper then Winford.   Also between the break out board I made and snap. Signal wires are twisted although they are differential so not sure i needed to anyway.  

I will try to add a pic here or send direct 

But after the breakout I run 30-40" cable runs to my servos twisted  signal wires and shielded.  

All seems to work.  I also have some 100uf caps on the 5vdd on my breakout to help insure cdd is clean here

image.jpeg

B



On Feb 1, 2014, at 10:47 AM, <david.m.stevenson@...> wrote:

 

Hi Tom,

I am just about to yank my last hair out. I posted last week about the issue I was having with my Konnect board and the suggested solution was to add an extra connector with a capacitor to kill suspected noise. I did do that, but the problem still exists.

Further to that, it has become impossible to get the servo motors setup and I suspect this may also be a noise problem. When I run the Autophasefindfast program, it comes up with wildly different numbers for all the values every time it is run. Because my system uses a SnapAmp, there is that 50 pin ribbon cable I have to deal with. It is also impossible to keep it under 12 inches in length and I don't know where to go from here. I have spent the past 4 days trying to tune the servos and gotten nowhere.

Any suggestions would be much appreciated.

Thanks,
David.
Group: DynoMotion Message: 9067 From: Wcarrothers Yahoo Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Think the yahoo board killed my message because of the attached pic so here it is with out

First let me suggest trying to sound a little more positive in your message.  As I have not gotten more  expert help from anyone more so then from Tom.  The documentation is good but requires thought.  But try to imagine how much thought goes into a production machine like a techno befor me it runs well and is ready for the market.  You will get there

That said. I also am having my first go around with a k and snaps with brushless servos.  My other system was a k with a k analog board and moog brushes servos. So in that case the movement is done by a 3rd party if you will.   There were signals in that case which I did add caps to and works well and has for years

That said. Perhaps this will help.  My new setup using the snaps has servos with differential a and b signals. Those are connected no problem but my index signal( 1 pulse per turn) was a problem.  It turned out to not be a differential signal.  It also did not provide a good signal to the lvttl inputs.  I switched to into that signal using the opto inputs and it now works great

The next step is to be sure the settings for that axis are uploaded to the k in the config and flash screen.  This is a step that I always forgot. And also the perspective for the upload and download (with respect to the pcs possition is all wrong for my mine. But that isn't his fault. Just something to pay attention to. ) I would prefer the buttons on say send to kflop and get from kflop. But that is just me

Anyway if these settings are not right things will not work.  The find c program will but a regular  axis control program will not work if this step is missed.   Perhaps will look like it is trying but will not be smooth


Now for the wiring.  Yes tenth sender pin outs for everything.  He really has no other practical choice but to do it that way as to give a screw terminal for all the io he gives would result in a huge board which would also be adding antennas perhaps to pickup more noise (ok perhaps a stretch there). But in anycase most systems would use a lower percentage is he connections so connect to the ones you need and not to he others

My solution was to use pololu connection cords. Becides being cheeper then Winford.   Also between the break out board I made and snap. Signal wires are twisted although they are differential so not sure i needed to anyway.  

I will try to add a pic here or send direct 

But after the breakout I run 30-40" cable runs to my servos twisted  signal wires and shielded.  

All seems to work.  I also have some 100uf caps on the 5vdd on my breakout to help insure cdd is clean here

On Feb 1, 2014, at 10:47 AM, <david.m.stevenson@...> wrote:

 

Hi Tom,

I am just about to yank my last hair out. I posted last week about the issue I was having with my Konnect board and the suggested solution was to add an extra connector with a capacitor to kill suspected noise. I did do that, but the problem still exists.

Further to that, it has become impossible to get the servo motors setup and I suspect this may also be a noise problem. When I run the Autophasefindfast program, it comes up with wildly different numbers for all the values every time it is run. Because my system uses a SnapAmp, there is that 50 pin ribbon cable I have to deal with. It is also impossible to keep it under 12 inches in length and I don't know where to go from here. I have spent the past 4 days trying to tune the servos and gotten nowhere.

Any suggestions would be much appreciated.

Thanks,
David.
Group: DynoMotion Message: 9068 From: Wcarrothers Yahoo Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Meant to say my cable runs to the servos are up to 40 ft

On Feb 1, 2014, at 11:42 AM, Wcarrothers Yahoo <wcarrothers@...> wrote:

 

Think the yahoo board killed my message because of the attached pic so here it is with out

First let me suggest trying to sound a little more positive in your message.  As I have not gotten more  expert help from anyone more so then from Tom.  The documentation is good but requires thought.  But try to imagine how much thought goes into a production machine like a techno befor me it runs well and is ready for the market.  You will get there

That said. I also am having my first go around with a k and snaps with brushless servos.  My other system was a k with a k analog board and moog brushes servos. So in that case the movement is done by a 3rd party if you will.   There were signals in that case which I did add caps to and works well and has for years

That said. Perhaps this will help.  My new setup using the snaps has servos with differential a and b signals. Those are connected no problem but my index signal( 1 pulse per turn) was a problem.  It turned out to not be a differential signal.  It also did not provide a good signal to the lvttl inputs.  I switched to into that signal using the opto inputs and it now works great

The next step is to be sure the settings for that axis are uploaded to the k in the config and flash screen.  This is a step that I always forgot. And also the perspective for the upload and download (with respect to the pcs possition is all wrong for my mine. But that isn't his fault. Just something to pay attention to. ) I would prefer the buttons on say send to kflop and get from kflop. But that is just me

Anyway if these settings are not right things will not work.  The find c program will but a regular  axis control program will not work if this step is missed.   Perhaps will look like it is trying but will not be smooth


Now for the wiring.  Yes tenth sender pin outs for everything.  He really has no other practical choice but to do it that way as to give a screw terminal for all the io he gives would result in a huge board which would also be adding antennas perhaps to pickup more noise (ok perhaps a stretch there). But in anycase most systems would use a lower percentage is he connections so connect to the ones you need and not to he others

My solution was to use pololu connection cords. Becides being cheeper then Winford.   Also between the break out board I made and snap. Signal wires are twisted although they are differential so not sure i needed to anyway.  

I will try to add a pic here or send direct 

But after the breakout I run 30-40" cable runs to my servos twisted  signal wires and shielded.  

All seems to work.  I also have some 100uf caps on the 5vdd on my breakout to help insure cdd is clean here

On Feb 1, 2014, at 10:47 AM, <david.m.stevenson@...> wrote:

 

Hi Tom,

I am just about to yank my last hair out. I posted last week about the issue I was having with my Konnect board and the suggested solution was to add an extra connector with a capacitor to kill suspected noise. I did do that, but the problem still exists.

Further to that, it has become impossible to get the servo motors setup and I suspect this may also be a noise problem. When I run the Autophasefindfast program, it comes up with wildly different numbers for all the values every time it is run. Because my system uses a SnapAmp, there is that 50 pin ribbon cable I have to deal with. It is also impossible to keep it under 12 inches in length and I don't know where to go from here. I have spent the past 4 days trying to tune the servos and gotten nowhere.

Any suggestions would be much appreciated.

Thanks,
David.
Group: DynoMotion Message: 9069 From: David Stevenson Date: 2/1/2014
Subject: Re: Ribbon Cable Woes [1 Attachment]
Hello B,

Thank you for your detailed reply.

Without question, Tom provides the best support I have seen anywhere. Sorry for the negative pitch, but my frustration has maxed out. Not being an electronics guy means I have to spend countless hours (days) researching the things that are common knowledge for electrical people. I now also have a very angry customer adding to the problem.

This is the third such machine I have built and do understand the process to tune the drives. I don't understand the process of how to eliminate noise, if that is actually the problem here. I guess what I am looking for is confirmation that this probably is noise and I will try to eliminate it. The servos are very good quality Baldor units with very expensive shielded cables.

Thanks again,
David.


On 2/1/2014 11:38 AM, Wcarrothers Yahoo wrote:
First let me suggest trying to sound a little more positive in your message.  As I have not gotten more  expert help from anyone more so then from Tom.  The documentation is good but requires thought.  But try to imagine how much thought goes into a production machine like a techno befor me it runs well and is ready for the market.  You will get there

That said. I also am having my first go around with a k and snaps with brushless servos.  My other system was a k with a k analog board and moog brushes servos. So in that case the movement is done by a 3rd party if you will.   There were signals in that case which I did add caps to and works well and has for years

That said. Perhaps this will help.  My new setup using the snaps has servos with differential a and b signals. Those are connected no problem but my index signal( 1 pulse per turn) was a problem.  It turned out to not be a differential signal.  It also did not provide a good signal to the lvttl inputs.  I switched to into that signal using the opto inputs and it now works great

The next step is to be sure the settings for that axis are uploaded to the k in the config and flash screen.  This is a step that I always forgot. And also the perspective for the upload and download (with respect to the pcs possition is all wrong for my mine. But that isn't his fault. Just something to pay attention to. ) I would prefer the buttons on say send to kflop and get from kflop. But that is just me

Anyway if these settings are not right things will not work.  The find c program will but a regular  axis control program will not work if this step is missed.   Perhaps will look like it is trying but will not be smooth


Now for the wiring.  Yes tenth sender pin outs for everything.  He really has no other practical choice but to do it that way as to give a screw terminal for all the io he gives would result in a huge board which would also be adding antennas perhaps to pickup more noise (ok perhaps a stretch there). But in anycase most systems would use a lower percentage is he connections so connect to the ones you need and not to he others

My solution was to use pololu connection cords. Becides being cheeper then Winford.   Also between the break out board I made and snap. Signal wires are twisted although they are differential so not sure i needed to anyway.  

I will try to add a pic here or send direct 

But after the breakout I run 30-40" cable runs to my servos twisted  signal wires and shielded.  

All seems to work.  I also have some 100uf caps on the 5vdd on my breakout to help insure cdd is clean here

image.jpeg

B



On Feb 1, 2014, at 10:47 AM, <david.m.stevenson@...> wrote:

 

Hi Tom,

I am just about to yank my last hair out. I posted last week about the issue I was having with my Konnect board and the suggested solution was to add an extra connector with a capacitor to kill suspected noise. I did do that, but the problem still exists.

Further to that, it has become impossible to get the servo motors setup and I suspect this may also be a noise problem. When I run the Autophasefindfast program, it comes up with wildly different numbers for all the values every time it is run. Because my system uses a SnapAmp, there is that 50 pin ribbon cable I have to deal with. It is also impossible to keep it under 12 inches in length and I don't know where to go from here. I have spent the past 4 days trying to tune the servos and gotten nowhere.

Any suggestions would be much appreciated.

Thanks,
David.
Group: DynoMotion Message: 9070 From: Wcarrothers Yahoo Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
The key for mine working without what looked like noise was getting the index signal right or moving it to one of the opto inputs once I did that the find program stated working seeing the pulse was clear.  Also. If you use the opto in you have to change the c program mark so it is opposit as the signal is pulled high when off And goes low when on or the opposit of that.  !Mark variables.

Also would be curious of your source for 5v being clean and servo source supply   

B

On Feb 1, 2014, at 12:00 PM, David Stevenson <david.m.stevenson@...> wrote:

Hello B,

Thank you for your detailed reply.

Without question, Tom provides the best support I have seen anywhere. Sorry for the negative pitch, but my frustration has maxed out. Not being an electronics guy means I have to spend countless hours (days) researching the things that are common knowledge for electrical people. I now also have a very angry customer adding to the problem.

This is the third such machine I have built and do understand the process to tune the drives. I don't understand the process of how to eliminate noise, if that is actually the problem here. I guess what I am looking for is confirmation that this probably is noise and I will try to eliminate it. The servos are very good quality Baldor units with very expensive shielded cables.

Thanks again,
David.


On 2/1/2014 11:38 AM, Wcarrothers Yahoo wrote:
First let me suggest trying to sound a little more positive in your message.  As I have not gotten more  expert help from anyone more so then from Tom.  The documentation is good but requires thought.  But try to imagine how much thought goes into a production machine like a techno befor me it runs well and is ready for the market.  You will get there

That said. I also am having my first go around with a k and snaps with brushless servos.  My other system was a k with a k analog board and moog brushes servos. So in that case the movement is done by a 3rd party if you will.   There were signals in that case which I did add caps to and works well and has for years

That said. Perhaps this will help.  My new setup using the snaps has servos with differential a and b signals. Those are connected no problem but my index signal( 1 pulse per turn) was a problem.  It turned out to not be a differential signal.  It also did not provide a good signal to the lvttl inputs.  I switched to into that signal using the opto inputs and it now works great

The next step is to be sure the settings for that axis are uploaded to the k in the config and flash screen.  This is a step that I always forgot. And also the perspective for the upload and download (with respect to the pcs possition is all wrong for my mine. But that isn't his fault. Just something to pay attention to. ) I would prefer the buttons on say send to kflop and get from kflop. But that is just me

Anyway if these settings are not right things will not work.  The find c program will but a regular  axis control program will not work if this step is missed.   Perhaps will look like it is trying but will not be smooth


Now for the wiring.  Yes tenth sender pin outs for everything.  He really has no other practical choice but to do it that way as to give a screw terminal for all the io he gives would result in a huge board which would also be adding antennas perhaps to pickup more noise (ok perhaps a stretch there). But in anycase most systems would use a lower percentage is he connections so connect to the ones you need and not to he others

My solution was to use pololu connection cords. Becides being cheeper then Winford.   Also between the break out board I made and snap. Signal wires are twisted although they are differential so not sure i needed to anyway.  

I will try to add a pic here or send direct 

But after the breakout I run 30-40" cable runs to my servos twisted  signal wires and shielded.  

All seems to work.  I also have some 100uf caps on the 5vdd on my breakout to help insure cdd is clean here

<mime-attachment.jpg>

B



On Feb 1, 2014, at 10:47 AM, <david.m.stevenson@...> wrote:

 

Hi Tom,

I am just about to yank my last hair out. I posted last week about the issue I was having with my Konnect board and the suggested solution was to add an extra connector with a capacitor to kill suspected noise. I did do that, but the problem still exists.

Further to that, it has become impossible to get the servo motors setup and I suspect this may also be a noise problem. When I run the Autophasefindfast program, it comes up with wildly different numbers for all the values every time it is run. Because my system uses a SnapAmp, there is that 50 pin ribbon cable I have to deal with. It is also impossible to keep it under 12 inches in length and I don't know where to go from here. I have spent the past 4 days trying to tune the servos and gotten nowhere.

Any suggestions would be much appreciated.

Thanks,
David.
Group: DynoMotion Message: 9071 From: Wcarrothers Yahoo Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
I know if is a bad pic. But if you can believe it by k and kanalog live in here with the moog amps in the back the reactor coils for ac power under it and ac filters to the right.  It i didn't start out this messy and I wished I could have used a larger box but features it could handle kept getting added on top of things as I build and ended up with this

image.jpeg

Note my huge caps on the right for 5 and 12vdc.  These can keep the system alive for a long time if the supply sees an interruption on the ac in but probably don't help any with noise as there esr is not very low and are probably more of a time bomb as if I did get any shorts they would unload and probably nuke everything but they are there and to much work to remove how. 

Some day when I retire ill clean up the wiring. Ha. 

B

On Feb 1, 2014, at 12:00 PM, David Stevenson <david.m.stevenson@...> wrote:

Hello B,

Thank you for your detailed reply.

Without question, Tom provides the best support I have seen anywhere. Sorry for the negative pitch, but my frustration has maxed out. Not being an electronics guy means I have to spend countless hours (days) researching the things that are common knowledge for electrical people. I now also have a very angry customer adding to the problem.

This is the third such machine I have built and do understand the process to tune the drives. I don't understand the process of how to eliminate noise, if that is actually the problem here. I guess what I am looking for is confirmation that this probably is noise and I will try to eliminate it. The servos are very good quality Baldor units with very expensive shielded cables.

Thanks again,
David.


On 2/1/2014 11:38 AM, Wcarrothers Yahoo wrote:
First let me suggest trying to sound a little more positive in your message.  As I have not gotten more  expert help from anyone more so then from Tom.  The documentation is good but requires thought.  But try to imagine how much thought goes into a production machine like a techno befor me it runs well and is ready for the market.  You will get there

That said. I also am having my first go around with a k and snaps with brushless servos.  My other system was a k with a k analog board and moog brushes servos. So in that case the movement is done by a 3rd party if you will.   There were signals in that case which I did add caps to and works well and has for years

That said. Perhaps this will help.  My new setup using the snaps has servos with differential a and b signals. Those are connected no problem but my index signal( 1 pulse per turn) was a problem.  It turned out to not be a differential signal.  It also did not provide a good signal to the lvttl inputs.  I switched to into that signal using the opto inputs and it now works great

The next step is to be sure the settings for that axis are uploaded to the k in the config and flash screen.  This is a step that I always forgot. And also the perspective for the upload and download (with respect to the pcs possition is all wrong for my mine. But that isn't his fault. Just something to pay attention to. ) I would prefer the buttons on say send to kflop and get from kflop. But that is just me

Anyway if these settings are not right things will not work.  The find c program will but a regular  axis control program will not work if this step is missed.   Perhaps will look like it is trying but will not be smooth


Now for the wiring.  Yes tenth sender pin outs for everything.  He really has no other practical choice but to do it that way as to give a screw terminal for all the io he gives would result in a huge board which would also be adding antennas perhaps to pickup more noise (ok perhaps a stretch there). But in anycase most systems would use a lower percentage is he connections so connect to the ones you need and not to he others

My solution was to use pololu connection cords. Becides being cheeper then Winford.   Also between the break out board I made and snap. Signal wires are twisted although they are differential so not sure i needed to anyway.  

I will try to add a pic here or send direct 

But after the breakout I run 30-40" cable runs to my servos twisted  signal wires and shielded.  

All seems to work.  I also have some 100uf caps on the 5vdd on my breakout to help insure cdd is clean here

<mime-attachment.jpg>

B



On Feb 1, 2014, at 10:47 AM, <david.m.stevenson@...> wrote:

 

Hi Tom,

I am just about to yank my last hair out. I posted last week about the issue I was having with my Konnect board and the suggested solution was to add an extra connector with a capacitor to kill suspected noise. I did do that, but the problem still exists.

Further to that, it has become impossible to get the servo motors setup and I suspect this may also be a noise problem. When I run the Autophasefindfast program, it comes up with wildly different numbers for all the values every time it is run. Because my system uses a SnapAmp, there is that 50 pin ribbon cable I have to deal with. It is also impossible to keep it under 12 inches in length and I don't know where to go from here. I have spent the past 4 days trying to tune the servos and gotten nowhere.

Any suggestions would be much appreciated.

Thanks,
David.
Group: DynoMotion Message: 9072 From: David Stevenson Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Thanks B.

I am going to move the index and see if that helps. Thank you for the pointers and pictures.

David

On 2/1/2014 12:07 PM, Wcarrothers Yahoo wrote:
 
The key for mine working without what looked like noise was getting the index signal right or moving it to one of the opto inputs once I did that the find program stated working seeing the pulse was clear.  Also. If you use the opto in you have to change the c program mark so it is opposit as the signal is pulled high when off And goes low when on or the opposit of that.  !Mark variables.

Also would be curious of your source for 5v being clean and servo source supply   

B

On Feb 1, 2014, at 12:00 PM, David Stevenson <david.m.stevenson@...> wrote:

Hello B,

Thank you for your detailed reply.

Without question, Tom provides the best support I have seen anywhere. Sorry for the negative pitch, but my frustration has maxed out. Not being an electronics guy means I have to spend countless hours (days) researching the things that are common knowledge for electrical people. I now also have a very angry customer adding to the problem.

This is the third such machine I have built and do understand the process to tune the drives. I don't understand the process of how to eliminate noise, if that is actually the problem here. I guess what I am looking for is confirmation that this probably is noise and I will try to eliminate it. The servos are very good quality Baldor units with very expensive shielded cables.

Thanks again,
David.


On 2/1/2014 11:38 AM, Wcarrothers Yahoo wrote:
First let me suggest trying to sound a little more positive in your message.  As I have not gotten more  expert help from anyone more so then from Tom.  The documentation is good but requires thought.  But try to imagine how much thought goes into a production machine like a techno befor me it runs well and is ready for the market.  You will get there

That said. I also am having my first go around with a k and snaps with brushless servos.  My other system was a k with a k analog board and moog brushes servos. So in that case the movement is done by a 3rd party if you will.   There were signals in that case which I did add caps to and works well and has for years

That said. Perhaps this will help.  My new setup using the snaps has servos with differential a and b signals. Those are connected no problem but my index signal( 1 pulse per turn) was a problem.  It turned out to not be a differential signal.  It also did not provide a good signal to the lvttl inputs.  I switched to into that signal using the opto inputs and it now works great

The next step is to be sure the settings for that axis are uploaded to the k in the config and flash screen.  This is a step that I always forgot. And also the perspective for the upload and download (with respect to the pcs possition is all wrong for my mine. But that isn't his fault. Just something to pay attention to. ) I would prefer the buttons on say send to kflop and get from kflop. But that is just me

Anyway if these settings are not right things will not work.  The find c program will but a regular  axis control program will not work if this step is missed.   Perhaps will look like it is trying but will not be smooth


Now for the wiring.  Yes tenth sender pin outs for everything.  He really has no other practical choice but to do it that way as to give a screw terminal for all the io he gives would result in a huge board which would also be adding antennas perhaps to pickup more noise (ok perhaps a stretch there). But in anycase most systems would use a lower percentage is he connections so connect to the ones you need and not to he others

My solution was to use pololu connection cords. Becides being cheeper then Winford.   Also between the break out board I made and snap. Signal wires are twisted although they are differential so not sure i needed to anyway.  

I will try to add a pic here or send direct 

But after the breakout I run 30-40" cable runs to my servos twisted  signal wires and shielded.  

All seems to work.  I also have some 100uf caps on the 5vdd on my breakout to help insure cdd is clean here

<mime-attachment.jpg>

B



On Feb 1, 2014, at 10:47 AM, <david.m.stevenson@...> wrote:

 

Hi Tom,

I am just about to yank my last hair out. I posted last week about the issue I was having with my Konnect board and the suggested solution was to add an extra connector with a capacitor to kill suspected noise. I did do that, but the problem still exists.

Further to that, it has become impossible to get the servo motors setup and I suspect this may also be a noise problem. When I run the Autophasefindfast program, it comes up with wildly different numbers for all the values every time it is run. Because my system uses a SnapAmp, there is that 50 pin ribbon cable I have to deal with. It is also impossible to keep it under 12 inches in length and I don't know where to go from here. I have spent the past 4 days trying to tune the servos and gotten nowhere.

Any suggestions would be much appreciated.

Thanks,
David.
Group: DynoMotion Message: 9073 From: Tom Kerekes Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Hi David,

If the AutoPhaseFindFast program is coming up with wildly varying number then that clearly shows one problem area.  The process is pretty simple. 

#1 Energizing the coils should cause the motor to rotate fairly smoothly,
#2 The index pulse needs to be perfectly reliable to only go high at one point on a rev
#3 The encoders must count correctly

Usually the problem is noise on the index pulse.  Could you explain what type of signal it is?  Differential or Single ended?  Shielded?  Termination Resistor?

You mentioned that you have motor wiring with shields?  It is usually best to connect those shield to an Earth ground.  But the Encoder and Index signal's shield should be connected on the KFLOP end only to KFLOP DC GND.

You might read Bill's Thread on cnczone regarding a similar issue.  This post has a little program to blink an LED and Print every time the Index pulse is detected (run it in some unused Thread).  You might use that and rotate the motor either by hand or with AutoPhaseFindFast to observe if it flashes/prints always at the right place.

http://www.cnczone.com/forums/dynomotion_kflop_kanalog/208274-brushless_c_code_example_snap_amp.html#post1420974

Regards
TK




Group: DynoMotion Message: 9074 From: Wcarrothers Yahoo Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
What is the index on now?   With a meter my index signal would show 3mv or so then on pulse would see maybe a spike to 60-90mv.  On a scope it would look better but the point is on the ttl it would not see the signal regularly. I did try cap and resistor to drive then I went to the opto positive and grounded the opto negative and it had no more issues

B

On Feb 1, 2014, at 12:21 PM, David Stevenson <david.m.stevenson@...> wrote:

 

Thanks B.

I am going to move the index and see if that helps. Thank you for the pointers and pictures.

David

On 2/1/2014 12:07 PM, Wcarrothers Yahoo wrote:
 
The key for mine working without what looked like noise was getting the index signal right or moving it to one of the opto inputs once I did that the find program stated working seeing the pulse was clear.  Also. If you use the opto in you have to change the c program mark so it is opposit as the signal is pulled high when off And goes low when on or the opposit of that.  !Mark variables.

Also would be curious of your source for 5v being clean and servo source supply   

B

On Feb 1, 2014, at 12:00 PM, David Stevenson <david.m.stevenson@...> wrote:

Hello B,

Thank you for your detailed reply.

Without question, Tom provides the best support I have seen anywhere. Sorry for the negative pitch, but my frustration has maxed out. Not being an electronics guy means I have to spend countless hours (days) researching the things that are common knowledge for electrical people. I now also have a very angry customer adding to the problem.

This is the third such machine I have built and do understand the process to tune the drives. I don't understand the process of how to eliminate noise, if that is actually the problem here. I guess what I am looking for is confirmation that this probably is noise and I will try to eliminate it. The servos are very good quality Baldor units with very expensive shielded cables.

Thanks again,
David.


On 2/1/2014 11:38 AM, Wcarrothers Yahoo wrote:
First let me suggest trying to sound a little more positive in your message.  As I have not gotten more  expert help from anyone more so then from Tom.  The documentation is good but requires thought.  But try to imagine how much thought goes into a production machine like a techno befor me it runs well and is ready for the market.  You will get there

That said. I also am having my first go around with a k and snaps with brushless servos.  My other system was a k with a k analog board and moog brushes servos. So in that case the movement is done by a 3rd party if you will.   There were signals in that case which I did add caps to and works well and has for years

That said. Perhaps this will help.  My new setup using the snaps has servos with differential a and b signals. Those are connected no problem but my index signal( 1 pulse per turn) was a problem.  It turned out to not be a differential signal.  It also did not provide a good signal to the lvttl inputs.  I switched to into that signal using the opto inputs and it now works great

The next step is to be sure the settings for that axis are uploaded to the k in the config and flash screen.  This is a step that I always forgot. And also the perspective for the upload and download (with respect to the pcs possition is all wrong for my mine. But that isn't his fault. Just something to pay attention to. ) I would prefer the buttons on say send to kflop and get from kflop. But that is just me

Anyway if these settings are not right things will not work.  The find c program will but a regular  axis control program will not work if this step is missed.   Perhaps will look like it is trying but will not be smooth


Now for the wiring.  Yes tenth sender pin outs for everything.  He really has no other practical choice but to do it that way as to give a screw terminal for all the io he gives would result in a huge board which would also be adding antennas perhaps to pickup more noise (ok perhaps a stretch there). But in anycase most systems would use a lower percentage is he connections so connect to the ones you need and not to he others

My solution was to use pololu connection cords. Becides being cheeper then Winford.   Also between the break out board I made and snap. Signal wires are twisted although they are differential so not sure i needed to anyway.  

I will try to add a pic here or send direct 

But after the breakout I run 30-40" cable runs to my servos twisted  signal wires and shielded.  

All seems to work.  I also have some 100uf caps on the 5vdd on my breakout to help insure cdd is clean here

<mime-attachment.jpg>

B



On Feb 1, 2014, at 10:47 AM, <david.m.stevenson@...> wrote:

 

Hi Tom,

I am just about to yank my last hair out. I posted last week about the issue I was having with my Konnect board and the suggested solution was to add an extra connector with a capacitor to kill suspected noise. I did do that, but the problem still exists.

Further to that, it has become impossible to get the servo motors setup and I suspect this may also be a noise problem. When I run the Autophasefindfast program, it comes up with wildly different numbers for all the values every time it is run. Because my system uses a SnapAmp, there is that 50 pin ribbon cable I have to deal with. It is also impossible to keep it under 12 inches in length and I don't know where to go from here. I have spent the past 4 days trying to tune the servos and gotten nowhere.

Any suggestions would be much appreciated.

Thanks,
David.
Group: DynoMotion Message: 9075 From: Wcarrothers Yahoo Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Ya the thread might help on the zone.   Although the light program seemed to work with success when connected to the ttl inputs until the coils were active. Then it would appear as more of a random signal

B
On Feb 1, 2014, at 12:22 PM, Tom Kerekes <tk@...> wrote:

 

Hi David,

If the AutoPhaseFindFast program is coming up with wildly varying number then that clearly shows one problem area.  The process is pretty simple. 

#1 Energizing the coils should cause the motor to rotate fairly smoothly,
#2 The index pulse needs to be perfectly reliable to only go high at one point on a rev
#3 The encoders must count correctly

Usually the problem is noise on the index pulse.  Could you explain what type of signal it is?  Differential or Single ended?  Shielded?  Termination Resistor?

You mentioned that you have motor wiring with shields?  It is usually best to connect those shield to an Earth ground.  But the Encoder and Index signal's shield should be connected on the KFLOP end only to KFLOP DC GND.

You might read Bill's Thread on cnczone regarding a similar issue.  This post has a little program to blink an LED and Print every time the Index pulse is detected (run it in some unused Thread).  You might use that and rotate the motor either by hand or with AutoPhaseFindFast to observe if it flashes/prints always at the right place.


Regards
TK




Group: DynoMotion Message: 9076 From: Tom Kerekes Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Hi Bill,

Well that was sort of the idea.  To show exactly where the problem was and what was causing it.

Regards
TK

Group: DynoMotion Message: 9077 From: David Stevenson Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Thank you Gents,

The encoder is shielded differential and doesn't have a resistor that I am aware of.

I do know the motor shield and encoder shield are both connected to earth ground, so I will change that.

B, the encoders including the index are connected using the SnapAmp pins between 15 and 30. I am going to move the index to pins 31-34 and see if that helps.

Thanks again,
David.


On 2/1/2014 12:41 PM, Tom Kerekes wrote:
 
Hi Bill,

Well that was sort of the idea.  To show exactly where the problem was and what was causing it.

Regards
TK

Group: DynoMotion Message: 9078 From: daveymahomh600e Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Moving the index signals over to the opto inputs and removing the common grounds between the encoder and motor shields seems to have cured the APFfast problem. Thank you!!!

With regard to the Konnect cap, how can I check to be sure the cap is doing it's job? Should it be soldered into the connector? It had quite small leads so i folded them over, plugged them into the connector and hoped for the best. I guess a continuity check would tell me if it is engaged properly?

Thanks again,
David.
Group: DynoMotion Message: 9079 From: Wcarrothers Yahoo Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Sweet. Not sure i know what the konnect cap is though

B

On Feb 1, 2014, at 3:08 PM, <david.m.stevenson@...> wrote:

 

Moving the index signals over to the opto inputs and removing the common grounds between the encoder and motor shields seems to have cured the APFfast problem. Thank you!!!

With regard to the Konnect cap, how can I check to be sure the cap is doing it's job? Should it be soldered into the connector? It had quite small leads so i folded them over, plugged them into the connector and hoped for the best. I guess a continuity check would tell me if it is engaged properly?

Thanks again,
David.
Group: DynoMotion Message: 9080 From: Tom Kerekes Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Hi David,

True that those connectors are supposed to be used with larger, square, gold coated pins.  You could plug in a mating connector and then solder the cap onto that if you like.  You can also check with an ohm meter if there is a good connection.  The bottom line is that there should be no unexpected KFLOP Resets.

Regards
TK

Group: DynoMotion Message: 9081 From: Tom Kerekes Date: 2/1/2014
Subject: Re: Ribbon Cable Woes
Hi David,

True that those connectors are supposed to be used with larger, square, gold coated pins.  You could plug in a mating connector and then solder the cap onto that if you like.  You can also check with an ohm meter if there is a good connection.  The bottom line is that there should be no unexpected KFLOP Resets.

Regards
TK