Re: [Emc-users] ADC - realtime for load cell

2019-09-20 Thread Dave Cole

So you want to balance the load on the platform with a single load cell?
That's going to be nearly impossible.

The reason is that when you drive the load up, the load cell will see an 
increase in force due to acceleration.


So you will need some position feedback as well.

Dave




On 9/18/2019 7:36 PM, Curtis Dutton wrote:

my experiment is to use a force measuring device to proportionally control
torque output on a servo attached to a ball screw.

as a thought experiment imagine a platform hoisted by said ball screw. as
you added weight to the platform the motor torque would increase to offset
the weight and keep it steady. or you could set the required force equal to
X kgs and with less than X kgs on the platform it would travel up
proportionally. the motor would increase torque to maintain the force
value. if more force than X the platform would travel down proportionally.
the motor would decrease torque to maintain the set force value.

in effect not a position commanded but a force commanded linear actuator.

is there some othere device that would be better for this type of force
measurement? in the 0 to about 500kg range.

On Wed, Sep 18, 2019, 11:51 AM Chris Albertson 
wrote:


in any case, even with lower 16-bit precision, the most important part of
the design is the analog portion that sits between the load cell and the
A/D converter.



On Wed, Sep 18, 2019 at 6:38 AM Curtis Dutton  wrote:


Chris thanks for the info.

I'm not exactly certain at this point what precision I need. I'll be
determining that experimentally later. I will keep in mind what you have
said, I'll deal with it after I get my hardware running.


On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson <

albertson.ch...@gmail.com

wrote:


If you want better than 16-bits and 1KHz then you also are going to

need

a

very good quality analog signal conditioning and instrumentation

amplifier.

   The analog circuit between the A/D converts and the load cell really

does

matter.   You have to reject common-mode or the low bits will just

be

filled with 60Hz noise.  And you can not simply use a "brick wall"

filter

or you will not get the desired bandwidth through.  (If you are needing

to

sample at 1K then you must care about 500Hz bandwidth.

This matters a LOT more than which A/D chip you use.

22 bits is better than one part per million. Do you really need that?

It

will be hard to isolate all of the electrical and mechanical noise
especially

On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton 

wrote:

I have a pet project I'm starting that will involve reading a load

cell

in

as near as real time as possible and using that input for servo

control.

I plan on using mesa hardware to interface to the equipment. Does

anyone

know of a nice way to interface hotsmot2 to a load cell?

I would like a fairly high resolution. 16 bit minimum, the higher the
better up to 22 bit I suppose. As well as the ability to sample at

the

servo thread rate of 1 Khz.


Thanks,
Curtis

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users



--

Chris Albertson
Redondo Beach, California

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users



--

Chris Albertson
Redondo Beach, California

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users



___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-20 Thread andy pugh
On Thu, 19 Sep 2019 at 18:13, Curtis Dutton  wrote:

> This encoder is a serial 17-bit with 4 wires. 0V, 5V, T+ and T-


That might be another puzzle. Mesa has a UART module, but that seems to
only want data in 8-bit bytes.

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is designed
for the especial use of mechanical geniuses, daredevils and lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-20 Thread Les Newell
That encoder is probably using a custom protocol and can only be used 
with systems specifically designed to work with it. You'll probably 
struggle to get it to work with LinuxCNC.
Is the motor brushed or brushless? Brushed motors generally have two 
main power wires, brushless normally have 3. If it's brushless, again it 
could be hard work to get it running unless you have the driver designed 
to work with it. If it's a brushed motor, pretty much any brushed motor 
driver that can put out enough voltage and current will work with it.


Les

On 19/09/2019 18:10, Curtis Dutton wrote:

This encoder is a serial 17-bit with 4 wires. 0V, 5V, T+ and T-

On Thu, Sep 19, 2019 at 11:26 AM Todd Zuercher 
wrote:


Or




___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-19 Thread Curtis Dutton
This encoder is a serial 17-bit with 4 wires. 0V, 5V, T+ and T-

On Thu, Sep 19, 2019 at 11:26 AM Todd Zuercher 
wrote:

> Or to think of it another way, closing the loop with a force meter, would
> be analogous to the air regulator on an air counterbalance.
>
> Todd Zuercher
> P. Graham Dunn Inc.
> 630 Henry Street
> Dalton, Ohio 44618
> Phone:  (330)828-2105ext. 2031
>
> -Original Message-
> From: Todd Zuercher 
> Sent: Thursday, September 19, 2019 8:32 AM
> To: Enhanced Machine Controller (EMC) 
> Subject: Re: [Emc-users] ADC - realtime for load cell
>
> I don't think closing the loop with force feedback would work, If the
> motor is applying x amount of force, adding more weight isn't going to
> change the amount of force read by the force feedback device.  It will just
> push it down.  The feedback will only be reading the force applied by the
> motor to resist the load, not the load itself, so changing the load will
> only cause acceleration.
>
> Todd Zuercher
> P. Graham Dunn Inc.
> 630 Henry Street
> Dalton, Ohio 44618
> Phone:  (330)828-2105ext. 2031
>
> -Original Message-
> From: Chris Albertson 
> Sent: Wednesday, September 18, 2019 8:31 PM
> To: Enhanced Machine Controller (EMC) 
> Subject: Re: [Emc-users] ADC - realtime for load cell
>
> One way to measure force is with a spring.  Force is measured by looking at
> spring displacement. It is kind of like the load cell idea but you use
> a spring scale.
>
> Another way that is not quite to accurate is to measure the current in the
> motor.  Current is proportional to torque.
>
> On Wed, Sep 18, 2019 at 4:39 PM Curtis Dutton  wrote:
>
> > my experiment is to use a force measuring device to proportionally
> > control torque output on a servo attached to a ball screw.
> >
> > as a thought experiment imagine a platform hoisted by said ball screw.
> > as you added weight to the platform the motor torque would increase to
> > offset the weight and keep it steady. or you could set the required
> > force equal to X kgs and with less than X kgs on the platform it would
> > travel up proportionally. the motor would increase torque to maintain
> > the force value. if more force than X the platform would travel down
> proportionally.
> > the motor would decrease torque to maintain the set force value.
> >
> > in effect not a position commanded but a force commanded linear actuator.
> >
> > is there some othere device that would be better for this type of
> > force measurement? in the 0 to about 500kg range.
> >
> > On Wed, Sep 18, 2019, 11:51 AM Chris Albertson
> > 
> > wrote:
> >
> > > in any case, even with lower 16-bit precision, the most important
> > > part of the design is the analog portion that sits between the load
> > > cell and the A/D converter.
> > >
> > >
> > >
> > > On Wed, Sep 18, 2019 at 6:38 AM Curtis Dutton 
> > wrote:
> > >
> > > > Chris thanks for the info.
> > > >
> > > > I'm not exactly certain at this point what precision I need. I'll
> > > > be determining that experimentally later. I will keep in mind what
> > > > you
> > have
> > > > said, I'll deal with it after I get my hardware running.
> > > >
> > > >
> > > > On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson <
> > > albertson.ch...@gmail.com
> > > > >
> > > > wrote:
> > > >
> > > > > If you want better than 16-bits and 1KHz then you also are going
> > > > > to
> > > need
> > > > a
> > > > > very good quality analog signal conditioning and instrumentation
> > > > amplifier.
> > > > >   The analog circuit between the A/D converts and the load cell
> > really
> > > > does
> > > > > matter.   You have to reject common-mode or the low bits will
> > just
> > > be
> > > > > filled with 60Hz noise.  And you can not simply use a "brick wall"
> > > filter
> > > > > or you will not get the desired bandwidth through.  (If you are
> > needing
> > > > to
> > > > > sample at 1K then you must care about 500Hz bandwidth.
> > > > >
> > > > > This matters a LOT more than which A/D chip you use.
> > > > >
> > > > > 22 bits is better than one part per million. Do you really need
> that?
> > > It
> > > > > will be hard to

Re: [Emc-users] ADC - realtime for load cell

2019-09-19 Thread Todd Zuercher
Or to think of it another way, closing the loop with a force meter, would be 
analogous to the air regulator on an air counterbalance.

Todd Zuercher
P. Graham Dunn Inc.
630 Henry Street 
Dalton, Ohio 44618
Phone:  (330)828-2105ext. 2031

-Original Message-
From: Todd Zuercher  
Sent: Thursday, September 19, 2019 8:32 AM
To: Enhanced Machine Controller (EMC) 
Subject: Re: [Emc-users] ADC - realtime for load cell

I don't think closing the loop with force feedback would work, If the motor is 
applying x amount of force, adding more weight isn't going to change the amount 
of force read by the force feedback device.  It will just push it down.  The 
feedback will only be reading the force applied by the motor to resist the 
load, not the load itself, so changing the load will only cause acceleration. 

Todd Zuercher
P. Graham Dunn Inc.
630 Henry Street
Dalton, Ohio 44618
Phone:  (330)828-2105ext. 2031

-Original Message-
From: Chris Albertson 
Sent: Wednesday, September 18, 2019 8:31 PM
To: Enhanced Machine Controller (EMC) 
Subject: Re: [Emc-users] ADC - realtime for load cell

One way to measure force is with a spring.  Force is measured by looking at
spring displacement. It is kind of like the load cell idea but you use
a spring scale.

Another way that is not quite to accurate is to measure the current in the 
motor.  Current is proportional to torque.

On Wed, Sep 18, 2019 at 4:39 PM Curtis Dutton  wrote:

> my experiment is to use a force measuring device to proportionally 
> control torque output on a servo attached to a ball screw.
>
> as a thought experiment imagine a platform hoisted by said ball screw. 
> as you added weight to the platform the motor torque would increase to 
> offset the weight and keep it steady. or you could set the required 
> force equal to X kgs and with less than X kgs on the platform it would 
> travel up proportionally. the motor would increase torque to maintain 
> the force value. if more force than X the platform would travel down 
> proportionally.
> the motor would decrease torque to maintain the set force value.
>
> in effect not a position commanded but a force commanded linear actuator.
>
> is there some othere device that would be better for this type of 
> force measurement? in the 0 to about 500kg range.
>
> On Wed, Sep 18, 2019, 11:51 AM Chris Albertson 
> 
> wrote:
>
> > in any case, even with lower 16-bit precision, the most important 
> > part of the design is the analog portion that sits between the load 
> > cell and the A/D converter.
> >
> >
> >
> > On Wed, Sep 18, 2019 at 6:38 AM Curtis Dutton 
> wrote:
> >
> > > Chris thanks for the info.
> > >
> > > I'm not exactly certain at this point what precision I need. I'll 
> > > be determining that experimentally later. I will keep in mind what 
> > > you
> have
> > > said, I'll deal with it after I get my hardware running.
> > >
> > >
> > > On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson <
> > albertson.ch...@gmail.com
> > > >
> > > wrote:
> > >
> > > > If you want better than 16-bits and 1KHz then you also are going 
> > > > to
> > need
> > > a
> > > > very good quality analog signal conditioning and instrumentation
> > > amplifier.
> > > >   The analog circuit between the A/D converts and the load cell
> really
> > > does
> > > > matter.   You have to reject common-mode or the low bits will
> just
> > be
> > > > filled with 60Hz noise.  And you can not simply use a "brick wall"
> > filter
> > > > or you will not get the desired bandwidth through.  (If you are
> needing
> > > to
> > > > sample at 1K then you must care about 500Hz bandwidth.
> > > >
> > > > This matters a LOT more than which A/D chip you use.
> > > >
> > > > 22 bits is better than one part per million. Do you really need that?
> > It
> > > > will be hard to isolate all of the electrical and mechanical 
> > > > noise especially
> > > >
> > > > On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton 
> > > > 
> > > wrote:
> > > >
> > > > > I have a pet project I'm starting that will involve reading a 
> > > > > load
> > cell
> > > > in
> > > > > as near as real time as possible and using that input for 
> > > > > servo
> > > control.
> > > > >
> > > > > I plan on using mesa hardware to interface to the equipment. 
> > >

Re: [Emc-users] ADC - realtime for load cell

2019-09-19 Thread Gene Heskett
On Thursday 19 September 2019 09:39:41 Curtis Dutton wrote:

> I appreciate the help and information.
>
> One last question before I go out and try this. I found a servo motor
> with a serial encoder. What interface board should I use to read that
> if I were using an 8i20 to power the servo?
>
That depends on the format the encoder outputs.  And there are several. 

The most popular is an open ended protocol whose range is limited only by 
the limits imposed by the number of valid bits the OS supports.  This is 
commonly call ABX or similar, where there is an electrical phase angle 
of 90 degrees between the A and B signals, with an additional, 3rd 
signal of 1 pulse per rev to mark the 0 degree point in the rotation. 
Such a scheme gives the system enough data the it can know both the 
actual rotational position of the device, but which direction it is 
turning, at any time one of the A or B signals changes state. These 
encoders often come in LVDS, meaning low voltage differential, with at 
least 8 leads. These work exceptionally well in high noise environments 
as noise is usually longitudinal, its the difference  between the A+ 
lead and the A- lead that is detected, ditto for B and X.

This is very precise, but assumes the homeing switch that indirectly 
zero's all the counters, is the absolute accurate reference. I find the 
typical miniature microswitch with a rollor tipped lever, can be 
repeatable if consistently operated by the machinery to sub .001", and 
often much smaller errors, like .0002".  Spring-dome pushbuttons, also 
selling for 50 cents or so are even more accurate in terms of 
repeatability. Contaminate exclusion may be up to the user.

There are encoders that output a train of serial data in absolute format, 
but because it takes a finite amount of time to send it, also have a 
latency between recording the position, and updating the controling 
program. So the machine has to slowed considerably as it approaches the 
stopping point. I'm not a great believer in that, but it has been made 
to work.

There are other ways to do it, often proprietary. Meaning if the company 
goes under, you have an unsupported device.  That bothers me.

Your trivia factoid for the day. ;-)

> On Thu, Sep 19, 2019 at 9:06 AM Les Newell 
>
> wrote:
> > If you are trying to counterbalance a pure weight this system would
> > inevitably end up accelerating into it's travel limits. However if
> > the load has some sort of spring component it should be stable. For
> > instance if you are trying to counterbalance a load so the force
> > feedback is providing most of the force but something else is
> > providing a small stabilizing force it should work.
> >
> > It is difficult to provide better advice without knowing more about
> > the application.
> >
> > Les
> >
> > On 19/09/2019 13:31, Todd Zuercher wrote:
> > > I don't think closing the loop with force feedback would work, If
> > > the
> >
> > motor is applying x amount of force, adding more weight isn't going
> > to change the amount of force read by the force feedback device.  It
> > will just push it down.  The feedback will only be reading the force
> > applied by the motor to resist the load, not the load itself, so
> > changing the load will only cause acceleration.
> >
> >
> >
> > ___
> > Emc-users mailing list
> > Emc-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/emc-users
>
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users


Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
If we desire respect for the law, we must first make the law respectable.
 - Louis D. Brandeis
Genes Web page 


___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-19 Thread dave engvall
It is not too unusual to use the torsion in a rotating shaft as a load 
indicator. Sensors at both ends aligned at zero force. The app I saw was 
using this  on the turbine shaft for a 125 Mw generator. Just a comment. ;-)


On 9/18/19 5:30 PM, Chris Albertson wrote:

One way to measure force is with a spring.  Force is measured by looking at
spring displacement. It is kind of like the load cell idea but you use
a spring scale.

Another way that is not quite to accurate is to measure the current in the
motor.  Current is proportional to torque.

On Wed, Sep 18, 2019 at 4:39 PM Curtis Dutton  wrote:


my experiment is to use a force measuring device to proportionally control
torque output on a servo attached to a ball screw.

as a thought experiment imagine a platform hoisted by said ball screw. as
you added weight to the platform the motor torque would increase to offset
the weight and keep it steady. or you could set the required force equal to
X kgs and with less than X kgs on the platform it would travel up
proportionally. the motor would increase torque to maintain the force
value. if more force than X the platform would travel down proportionally.
the motor would decrease torque to maintain the set force value.

in effect not a position commanded but a force commanded linear actuator.

is there some othere device that would be better for this type of force
measurement? in the 0 to about 500kg range.

On Wed, Sep 18, 2019, 11:51 AM Chris Albertson 
wrote:


in any case, even with lower 16-bit precision, the most important part of
the design is the analog portion that sits between the load cell and the
A/D converter.



On Wed, Sep 18, 2019 at 6:38 AM Curtis Dutton 

wrote:

Chris thanks for the info.

I'm not exactly certain at this point what precision I need. I'll be
determining that experimentally later. I will keep in mind what you

have

said, I'll deal with it after I get my hardware running.


On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson <

albertson.ch...@gmail.com

wrote:


If you want better than 16-bits and 1KHz then you also are going to

need

a

very good quality analog signal conditioning and instrumentation

amplifier.

   The analog circuit between the A/D converts and the load cell

really

does

matter.   You have to reject common-mode or the low bits will

just

be

filled with 60Hz noise.  And you can not simply use a "brick wall"

filter

or you will not get the desired bandwidth through.  (If you are

needing

to

sample at 1K then you must care about 500Hz bandwidth.

This matters a LOT more than which A/D chip you use.

22 bits is better than one part per million. Do you really need that?

It

will be hard to isolate all of the electrical and mechanical noise
especially

On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton 

wrote:

I have a pet project I'm starting that will involve reading a load

cell

in

as near as real time as possible and using that input for servo

control.

I plan on using mesa hardware to interface to the equipment. Does

anyone

know of a nice way to interface hotsmot2 to a load cell?

I would like a fairly high resolution. 16 bit minimum, the higher

the

better up to 22 bit I suppose. As well as the ability to sample at

the

servo thread rate of 1 Khz.


Thanks,
Curtis

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users



--

Chris Albertson
Redondo Beach, California

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users



--

Chris Albertson
Redondo Beach, California

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users






___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-19 Thread Curtis Dutton
I appreciate the help and information.

One last question before I go out and try this. I found a servo motor with
a serial encoder. What interface board should I use to read that if I were
using an 8i20 to power the servo?

On Thu, Sep 19, 2019 at 9:06 AM Les Newell 
wrote:

> If you are trying to counterbalance a pure weight this system would
> inevitably end up accelerating into it's travel limits. However if the
> load has some sort of spring component it should be stable. For instance
> if you are trying to counterbalance a load so the force feedback is
> providing most of the force but something else is providing a small
> stabilizing force it should work.
>
> It is difficult to provide better advice without knowing more about the
> application.
>
> Les
>
> On 19/09/2019 13:31, Todd Zuercher wrote:
> > I don't think closing the loop with force feedback would work, If the
> motor is applying x amount of force, adding more weight isn't going to
> change the amount of force read by the force feedback device.  It will just
> push it down.  The feedback will only be reading the force applied by the
> motor to resist the load, not the load itself, so changing the load will
> only cause acceleration.
>
>
>
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-19 Thread Les Newell
If you are trying to counterbalance a pure weight this system would 
inevitably end up accelerating into it's travel limits. However if the 
load has some sort of spring component it should be stable. For instance 
if you are trying to counterbalance a load so the force feedback is 
providing most of the force but something else is providing a small 
stabilizing force it should work.


It is difficult to provide better advice without knowing more about the 
application.


Les

On 19/09/2019 13:31, Todd Zuercher wrote:

I don't think closing the loop with force feedback would work, If the motor is 
applying x amount of force, adding more weight isn't going to change the amount 
of force read by the force feedback device.  It will just push it down.  The 
feedback will only be reading the force applied by the motor to resist the 
load, not the load itself, so changing the load will only cause acceleration.




___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-19 Thread Todd Zuercher
I don't think closing the loop with force feedback would work, If the motor is 
applying x amount of force, adding more weight isn't going to change the amount 
of force read by the force feedback device.  It will just push it down.  The 
feedback will only be reading the force applied by the motor to resist the 
load, not the load itself, so changing the load will only cause acceleration. 

Todd Zuercher
P. Graham Dunn Inc.
630 Henry Street 
Dalton, Ohio 44618
Phone:  (330)828-2105ext. 2031

-Original Message-
From: Chris Albertson  
Sent: Wednesday, September 18, 2019 8:31 PM
To: Enhanced Machine Controller (EMC) 
Subject: Re: [Emc-users] ADC - realtime for load cell

One way to measure force is with a spring.  Force is measured by looking at
spring displacement. It is kind of like the load cell idea but you use
a spring scale.

Another way that is not quite to accurate is to measure the current in the 
motor.  Current is proportional to torque.

On Wed, Sep 18, 2019 at 4:39 PM Curtis Dutton  wrote:

> my experiment is to use a force measuring device to proportionally 
> control torque output on a servo attached to a ball screw.
>
> as a thought experiment imagine a platform hoisted by said ball screw. 
> as you added weight to the platform the motor torque would increase to 
> offset the weight and keep it steady. or you could set the required 
> force equal to X kgs and with less than X kgs on the platform it would 
> travel up proportionally. the motor would increase torque to maintain 
> the force value. if more force than X the platform would travel down 
> proportionally.
> the motor would decrease torque to maintain the set force value.
>
> in effect not a position commanded but a force commanded linear actuator.
>
> is there some othere device that would be better for this type of 
> force measurement? in the 0 to about 500kg range.
>
> On Wed, Sep 18, 2019, 11:51 AM Chris Albertson 
> 
> wrote:
>
> > in any case, even with lower 16-bit precision, the most important 
> > part of the design is the analog portion that sits between the load 
> > cell and the A/D converter.
> >
> >
> >
> > On Wed, Sep 18, 2019 at 6:38 AM Curtis Dutton 
> wrote:
> >
> > > Chris thanks for the info.
> > >
> > > I'm not exactly certain at this point what precision I need. I'll 
> > > be determining that experimentally later. I will keep in mind what 
> > > you
> have
> > > said, I'll deal with it after I get my hardware running.
> > >
> > >
> > > On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson <
> > albertson.ch...@gmail.com
> > > >
> > > wrote:
> > >
> > > > If you want better than 16-bits and 1KHz then you also are going 
> > > > to
> > need
> > > a
> > > > very good quality analog signal conditioning and instrumentation
> > > amplifier.
> > > >   The analog circuit between the A/D converts and the load cell
> really
> > > does
> > > > matter.   You have to reject common-mode or the low bits will
> just
> > be
> > > > filled with 60Hz noise.  And you can not simply use a "brick wall"
> > filter
> > > > or you will not get the desired bandwidth through.  (If you are
> needing
> > > to
> > > > sample at 1K then you must care about 500Hz bandwidth.
> > > >
> > > > This matters a LOT more than which A/D chip you use.
> > > >
> > > > 22 bits is better than one part per million. Do you really need that?
> > It
> > > > will be hard to isolate all of the electrical and mechanical 
> > > > noise especially
> > > >
> > > > On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton 
> > > > 
> > > wrote:
> > > >
> > > > > I have a pet project I'm starting that will involve reading a 
> > > > > load
> > cell
> > > > in
> > > > > as near as real time as possible and using that input for 
> > > > > servo
> > > control.
> > > > >
> > > > > I plan on using mesa hardware to interface to the equipment. 
> > > > > Does
> > > anyone
> > > > > know of a nice way to interface hotsmot2 to a load cell?
> > > > >
> > > > > I would like a fairly high resolution. 16 bit minimum, the 
> > > > > higher
> the
> > > > > better up to 22 bit I suppose. As well as the ability to 
> > > > > sample at
> > the
> > > > > servo thre

Re: [Emc-users] ADC - realtime for load cell

2019-09-19 Thread Thaddeus Waldner
Thanks for correcting me on that.

Thaddeus Waldner
Newdale School
Elkton, SD 57026


From: andy pugh 
Sent: Thursday, September 19, 2019 4:03 AM
To: Enhanced Machine Controller (EMC)
Subject: Re: [Emc-users] ADC - realtime for load cell

On Thu, 19 Sep 2019 at 02:14, Thaddeus Waldner  wrote:

> Standard strain gages have a usable resolution of about 6000, as far as I
> know.
>

You can do much better than that. I used to use home-made clip-gauges when
I worked in the structural integrity field. Those were just a bent bit of
clock spring with strain gauges bonded on, but could be used to measure
sub-micron displacement on a 20mm gauge length.This was using proto-board
and an Op-amp.

Dedicated strain gauge amps are available with 24 bit resolution and 50kHz
sampling, for a price.
https://www.ni.com/en-us/shop/select/c-series-strain-bridge-input-module?modelId=122190



--
atp
"A motorcycle is a bicycle with a pandemonium attachment and is designed
for the especial use of mechanical geniuses, daredevils and lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-19 Thread andy pugh
On Thu, 19 Sep 2019 at 02:14, Thaddeus Waldner  wrote:

> Standard strain gages have a usable resolution of about 6000, as far as I
> know.
>

 You can do much better than that. I used to use home-made clip-gauges when
I worked in the structural integrity field. Those were just a bent bit of
clock spring with strain gauges bonded on, but could be used to measure
sub-micron displacement on a 20mm gauge length.This was using proto-board
and an Op-amp.

Dedicated strain gauge amps are available with 24 bit resolution and 50kHz
sampling, for a price.
https://www.ni.com/en-us/shop/select/c-series-strain-bridge-input-module?modelId=122190



-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is designed
for the especial use of mechanical geniuses, daredevils and lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread Chris Albertson
The OP said he needs a 1KHz sample rate. Those chips are about 50x to
slow.   Doing both high resolution and high bandwidth will be a hard
problem.

On Wed, Sep 18, 2019 at 6:01 PM Nathan Hartman 
wrote:

> On Wed, Sep 18, 2019 at 9:38 AM Curtis Dutton  wrote:
>
> > On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson <
> albertson.ch...@gmail.com
> > >
> > wrote:
> >
> > > If you want better than 16-bits and 1KHz then you also are going to
> need
> > a
> > > very good quality analog signal conditioning and instrumentation
> > amplifier.
> > >   The analog circuit between the A/D converts and the load cell really
> > does
> > > matter.   You have to reject common-mode or the low bits will just
> be
> > > filled with 60Hz noise.  And you can not simply use a "brick wall"
> filter
> > > or you will not get the desired bandwidth through.  (If you are needing
> > to
> > > sample at 1K then you must care about 500Hz bandwidth.
> > >
> > > This matters a LOT more than which A/D chip you use
>
>
> What you need is a Analog Front End (AFE) that is designed specifically for
> a strain gauge. Usually that's a single IC (that probably requires some
> passive components around it, which are specified in the datasheet) and if
> you're not designing your own circuit board there are little circuit boards
> ready made that you can get from Amazon that take the wires for a strain
> gauge on one end and wires for power and SPI on the other end. There's a IC
> called HX711 that's popular with Arduino enthusiasts.
>
> If you are building your own circuit board, the various AFE manufacturers
> (Texas Instruments, Analog Devices, etc) tell you exactly how in their
> datasheets, application notes, and reference designs.
>
> The difference between a AFE and A/D converter is that a AFE has the A/D
> and also all the amplification and filtering etc built in to the chip,
> whereas you'd have to build a pretty complex circuit board to do all of
> that with discretes and op amps, and your results won't be as good as the
> AFE chip.
>
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>


-- 

Chris Albertson
Redondo Beach, California

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread Thaddeus Waldner
Standard strain gages have a usable resolution of about 6000, as far as I know. 

> On Sep 18, 2019, at 7:58 PM, Nathan Hartman  wrote:
> 
> On Wed, Sep 18, 2019 at 9:38 AM Curtis Dutton  wrote:
> 
>> On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson >> 
>> wrote:
>> 
>>> If you want better than 16-bits and 1KHz then you also are going to need
>> a
>>> very good quality analog signal conditioning and instrumentation
>> amplifier.
>>>  The analog circuit between the A/D converts and the load cell really
>> does
>>> matter.   You have to reject common-mode or the low bits will just be
>>> filled with 60Hz noise.  And you can not simply use a "brick wall" filter
>>> or you will not get the desired bandwidth through.  (If you are needing
>> to
>>> sample at 1K then you must care about 500Hz bandwidth.
>>> 
>>> This matters a LOT more than which A/D chip you use
> 
> 
> What you need is a Analog Front End (AFE) that is designed specifically for
> a strain gauge. Usually that's a single IC (that probably requires some
> passive components around it, which are specified in the datasheet) and if
> you're not designing your own circuit board there are little circuit boards
> ready made that you can get from Amazon that take the wires for a strain
> gauge on one end and wires for power and SPI on the other end. There's a IC
> called HX711 that's popular with Arduino enthusiasts.
> 
> If you are building your own circuit board, the various AFE manufacturers
> (Texas Instruments, Analog Devices, etc) tell you exactly how in their
> datasheets, application notes, and reference designs.
> 
> The difference between a AFE and A/D converter is that a AFE has the A/D
> and also all the amplification and filtering etc built in to the chip,
> whereas you'd have to build a pretty complex circuit board to do all of
> that with discretes and op amps, and your results won't be as good as the
> AFE chip.
> 
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users



___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread Nathan Hartman
On Wed, Sep 18, 2019 at 9:38 AM Curtis Dutton  wrote:

> On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson  >
> wrote:
>
> > If you want better than 16-bits and 1KHz then you also are going to need
> a
> > very good quality analog signal conditioning and instrumentation
> amplifier.
> >   The analog circuit between the A/D converts and the load cell really
> does
> > matter.   You have to reject common-mode or the low bits will just be
> > filled with 60Hz noise.  And you can not simply use a "brick wall" filter
> > or you will not get the desired bandwidth through.  (If you are needing
> to
> > sample at 1K then you must care about 500Hz bandwidth.
> >
> > This matters a LOT more than which A/D chip you use


What you need is a Analog Front End (AFE) that is designed specifically for
a strain gauge. Usually that's a single IC (that probably requires some
passive components around it, which are specified in the datasheet) and if
you're not designing your own circuit board there are little circuit boards
ready made that you can get from Amazon that take the wires for a strain
gauge on one end and wires for power and SPI on the other end. There's a IC
called HX711 that's popular with Arduino enthusiasts.

If you are building your own circuit board, the various AFE manufacturers
(Texas Instruments, Analog Devices, etc) tell you exactly how in their
datasheets, application notes, and reference designs.

The difference between a AFE and A/D converter is that a AFE has the A/D
and also all the amplification and filtering etc built in to the chip,
whereas you'd have to build a pretty complex circuit board to do all of
that with discretes and op amps, and your results won't be as good as the
AFE chip.

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread Chris Albertson
One way to measure force is with a spring.  Force is measured by looking at
spring displacement. It is kind of like the load cell idea but you use
a spring scale.

Another way that is not quite to accurate is to measure the current in the
motor.  Current is proportional to torque.

On Wed, Sep 18, 2019 at 4:39 PM Curtis Dutton  wrote:

> my experiment is to use a force measuring device to proportionally control
> torque output on a servo attached to a ball screw.
>
> as a thought experiment imagine a platform hoisted by said ball screw. as
> you added weight to the platform the motor torque would increase to offset
> the weight and keep it steady. or you could set the required force equal to
> X kgs and with less than X kgs on the platform it would travel up
> proportionally. the motor would increase torque to maintain the force
> value. if more force than X the platform would travel down proportionally.
> the motor would decrease torque to maintain the set force value.
>
> in effect not a position commanded but a force commanded linear actuator.
>
> is there some othere device that would be better for this type of force
> measurement? in the 0 to about 500kg range.
>
> On Wed, Sep 18, 2019, 11:51 AM Chris Albertson 
> wrote:
>
> > in any case, even with lower 16-bit precision, the most important part of
> > the design is the analog portion that sits between the load cell and the
> > A/D converter.
> >
> >
> >
> > On Wed, Sep 18, 2019 at 6:38 AM Curtis Dutton 
> wrote:
> >
> > > Chris thanks for the info.
> > >
> > > I'm not exactly certain at this point what precision I need. I'll be
> > > determining that experimentally later. I will keep in mind what you
> have
> > > said, I'll deal with it after I get my hardware running.
> > >
> > >
> > > On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson <
> > albertson.ch...@gmail.com
> > > >
> > > wrote:
> > >
> > > > If you want better than 16-bits and 1KHz then you also are going to
> > need
> > > a
> > > > very good quality analog signal conditioning and instrumentation
> > > amplifier.
> > > >   The analog circuit between the A/D converts and the load cell
> really
> > > does
> > > > matter.   You have to reject common-mode or the low bits will
> just
> > be
> > > > filled with 60Hz noise.  And you can not simply use a "brick wall"
> > filter
> > > > or you will not get the desired bandwidth through.  (If you are
> needing
> > > to
> > > > sample at 1K then you must care about 500Hz bandwidth.
> > > >
> > > > This matters a LOT more than which A/D chip you use.
> > > >
> > > > 22 bits is better than one part per million. Do you really need that?
> > It
> > > > will be hard to isolate all of the electrical and mechanical noise
> > > > especially
> > > >
> > > > On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton 
> > > wrote:
> > > >
> > > > > I have a pet project I'm starting that will involve reading a load
> > cell
> > > > in
> > > > > as near as real time as possible and using that input for servo
> > > control.
> > > > >
> > > > > I plan on using mesa hardware to interface to the equipment. Does
> > > anyone
> > > > > know of a nice way to interface hotsmot2 to a load cell?
> > > > >
> > > > > I would like a fairly high resolution. 16 bit minimum, the higher
> the
> > > > > better up to 22 bit I suppose. As well as the ability to sample at
> > the
> > > > > servo thread rate of 1 Khz.
> > > > >
> > > > >
> > > > > Thanks,
> > > > >Curtis
> > > > >
> > > > > ___
> > > > > Emc-users mailing list
> > > > > Emc-users@lists.sourceforge.net
> > > > > https://lists.sourceforge.net/lists/listinfo/emc-users
> > > > >
> > > >
> > > >
> > > > --
> > > >
> > > > Chris Albertson
> > > > Redondo Beach, California
> > > >
> > > > ___
> > > > Emc-users mailing list
> > > > Emc-users@lists.sourceforge.net
> > > > https://lists.sourceforge.net/lists/listinfo/emc-users
> > > >
> > >
> > > ___
> > > Emc-users mailing list
> > > Emc-users@lists.sourceforge.net
> > > https://lists.sourceforge.net/lists/listinfo/emc-users
> > >
> >
> >
> > --
> >
> > Chris Albertson
> > Redondo Beach, California
> >
> > ___
> > Emc-users mailing list
> > Emc-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/emc-users
> >
>
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>


-- 

Chris Albertson
Redondo Beach, California

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread Curtis Dutton
my experiment is to use a force measuring device to proportionally control
torque output on a servo attached to a ball screw.

as a thought experiment imagine a platform hoisted by said ball screw. as
you added weight to the platform the motor torque would increase to offset
the weight and keep it steady. or you could set the required force equal to
X kgs and with less than X kgs on the platform it would travel up
proportionally. the motor would increase torque to maintain the force
value. if more force than X the platform would travel down proportionally.
the motor would decrease torque to maintain the set force value.

in effect not a position commanded but a force commanded linear actuator.

is there some othere device that would be better for this type of force
measurement? in the 0 to about 500kg range.

On Wed, Sep 18, 2019, 11:51 AM Chris Albertson 
wrote:

> in any case, even with lower 16-bit precision, the most important part of
> the design is the analog portion that sits between the load cell and the
> A/D converter.
>
>
>
> On Wed, Sep 18, 2019 at 6:38 AM Curtis Dutton  wrote:
>
> > Chris thanks for the info.
> >
> > I'm not exactly certain at this point what precision I need. I'll be
> > determining that experimentally later. I will keep in mind what you have
> > said, I'll deal with it after I get my hardware running.
> >
> >
> > On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson <
> albertson.ch...@gmail.com
> > >
> > wrote:
> >
> > > If you want better than 16-bits and 1KHz then you also are going to
> need
> > a
> > > very good quality analog signal conditioning and instrumentation
> > amplifier.
> > >   The analog circuit between the A/D converts and the load cell really
> > does
> > > matter.   You have to reject common-mode or the low bits will just
> be
> > > filled with 60Hz noise.  And you can not simply use a "brick wall"
> filter
> > > or you will not get the desired bandwidth through.  (If you are needing
> > to
> > > sample at 1K then you must care about 500Hz bandwidth.
> > >
> > > This matters a LOT more than which A/D chip you use.
> > >
> > > 22 bits is better than one part per million. Do you really need that?
> It
> > > will be hard to isolate all of the electrical and mechanical noise
> > > especially
> > >
> > > On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton 
> > wrote:
> > >
> > > > I have a pet project I'm starting that will involve reading a load
> cell
> > > in
> > > > as near as real time as possible and using that input for servo
> > control.
> > > >
> > > > I plan on using mesa hardware to interface to the equipment. Does
> > anyone
> > > > know of a nice way to interface hotsmot2 to a load cell?
> > > >
> > > > I would like a fairly high resolution. 16 bit minimum, the higher the
> > > > better up to 22 bit I suppose. As well as the ability to sample at
> the
> > > > servo thread rate of 1 Khz.
> > > >
> > > >
> > > > Thanks,
> > > >Curtis
> > > >
> > > > ___
> > > > Emc-users mailing list
> > > > Emc-users@lists.sourceforge.net
> > > > https://lists.sourceforge.net/lists/listinfo/emc-users
> > > >
> > >
> > >
> > > --
> > >
> > > Chris Albertson
> > > Redondo Beach, California
> > >
> > > ___
> > > Emc-users mailing list
> > > Emc-users@lists.sourceforge.net
> > > https://lists.sourceforge.net/lists/listinfo/emc-users
> > >
> >
> > ___
> > Emc-users mailing list
> > Emc-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/emc-users
> >
>
>
> --
>
> Chris Albertson
> Redondo Beach, California
>
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread Chris Albertson
in any case, even with lower 16-bit precision, the most important part of
the design is the analog portion that sits between the load cell and the
A/D converter.



On Wed, Sep 18, 2019 at 6:38 AM Curtis Dutton  wrote:

> Chris thanks for the info.
>
> I'm not exactly certain at this point what precision I need. I'll be
> determining that experimentally later. I will keep in mind what you have
> said, I'll deal with it after I get my hardware running.
>
>
> On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson  >
> wrote:
>
> > If you want better than 16-bits and 1KHz then you also are going to need
> a
> > very good quality analog signal conditioning and instrumentation
> amplifier.
> >   The analog circuit between the A/D converts and the load cell really
> does
> > matter.   You have to reject common-mode or the low bits will just be
> > filled with 60Hz noise.  And you can not simply use a "brick wall" filter
> > or you will not get the desired bandwidth through.  (If you are needing
> to
> > sample at 1K then you must care about 500Hz bandwidth.
> >
> > This matters a LOT more than which A/D chip you use.
> >
> > 22 bits is better than one part per million. Do you really need that?  It
> > will be hard to isolate all of the electrical and mechanical noise
> > especially
> >
> > On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton 
> wrote:
> >
> > > I have a pet project I'm starting that will involve reading a load cell
> > in
> > > as near as real time as possible and using that input for servo
> control.
> > >
> > > I plan on using mesa hardware to interface to the equipment. Does
> anyone
> > > know of a nice way to interface hotsmot2 to a load cell?
> > >
> > > I would like a fairly high resolution. 16 bit minimum, the higher the
> > > better up to 22 bit I suppose. As well as the ability to sample at the
> > > servo thread rate of 1 Khz.
> > >
> > >
> > > Thanks,
> > >Curtis
> > >
> > > ___
> > > Emc-users mailing list
> > > Emc-users@lists.sourceforge.net
> > > https://lists.sourceforge.net/lists/listinfo/emc-users
> > >
> >
> >
> > --
> >
> > Chris Albertson
> > Redondo Beach, California
> >
> > ___
> > Emc-users mailing list
> > Emc-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/emc-users
> >
>
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>


-- 

Chris Albertson
Redondo Beach, California

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread Dave Cole

On 9/18/2019 11:15 AM, Jon Elson wrote:
On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson 
 wrote:
If you want better than 16-bits and 1KHz then you also are going to 
need a
very good quality analog signal conditioning and instrumentation 
amplifier.


I've never heard of anybody getting 16-bit resolution from a load cell 
or strain gauge.
There are full-bridge strain gauges with additional adjustments to 
balance the common mode noise down to the minimum, but still, 16 bits 
is one part in 65,536, which is really tough for
any analog electronics.  Most people say that anything better than 1 
part in 10,000 is getting extreme.  Anyway, strain gauges produce very 
small signals, so a really good amplifier that cancels out all noises 
is critical.  Ambient RF that gets rectified in the amplifier inputs 
will likely be the greatest problem.


Jon





If you need 16+ bits of resolution to see enough the changes to drive 
your servos, your control design is likely incorrect.


Try and change your mechanics or sensing method so you have a more 
significant signal, otherwise noise will drive your control loop/s crazy.


Strain gauges have their limits.

Dave




___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread dave engvall
MAX11905 might do the job. 20 bit. but spendy. probably $50 in singles. 
QSPI and SPI for interfaces. data rate 1600 ksps. SAR > 16 bit gets 
really touchy... CMRR, noise, etc. Good luck.


On 9/18/19 7:01 AM, Curtis Dutton wrote:

Right now my test set up is going to consist of a 6i25, 7i85 and an 8i20. I
wonder if the 7i85 could expose a BSPI interface. Or could I tack a BSPI on
to the 6i25's spare 25 pin connector?

I also am considering just writing a parallel port driver running at some
higher rate than the servo thread that would allow me to communicate with
the spi device.

On Wed, Sep 18, 2019 at 9:36 AM Curtis Dutton  wrote:


Chris thanks for the info.

I'm not exactly certain at this point what precision I need. I'll be
determining that experimentally later. I will keep in mind what you have
said, I'll deal with it after I get my hardware running.


On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson 
wrote:


If you want better than 16-bits and 1KHz then you also are going to need a
very good quality analog signal conditioning and instrumentation
amplifier.
   The analog circuit between the A/D converts and the load cell really
does
matter.   You have to reject common-mode or the low bits will just be
filled with 60Hz noise.  And you can not simply use a "brick wall" filter
or you will not get the desired bandwidth through.  (If you are needing to
sample at 1K then you must care about 500Hz bandwidth.

This matters a LOT more than which A/D chip you use.

22 bits is better than one part per million. Do you really need that?  It
will be hard to isolate all of the electrical and mechanical noise
especially

On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton  wrote:


I have a pet project I'm starting that will involve reading a load cell

in

as near as real time as possible and using that input for servo control.

I plan on using mesa hardware to interface to the equipment. Does anyone
know of a nice way to interface hotsmot2 to a load cell?

I would like a fairly high resolution. 16 bit minimum, the higher the
better up to 22 bit I suppose. As well as the ability to sample at the
servo thread rate of 1 Khz.


Thanks,
Curtis

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users



--

Chris Albertson
Redondo Beach, California

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users



___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread Jon Elson
On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson 
 wrote:

If you want better than 16-bits and 1KHz then you also are going to need a
very good quality analog signal conditioning and instrumentation amplifier.

I've never heard of anybody getting 16-bit resolution from a 
load cell or strain gauge.
There are full-bridge strain gauges with additional 
adjustments to balance the common mode noise down to the 
minimum, but still, 16 bits is one part in 65,536, which is 
really tough for
any analog electronics.  Most people say that anything 
better than 1 part in 10,000 is getting extreme.  Anyway, 
strain gauges produce very small signals, so a really good 
amplifier that cancels out all noises is critical.  Ambient 
RF that gets rectified in the amplifier inputs will likely 
be the greatest problem.


Jon


___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread Curtis Dutton
Right now my test set up is going to consist of a 6i25, 7i85 and an 8i20. I
wonder if the 7i85 could expose a BSPI interface. Or could I tack a BSPI on
to the 6i25's spare 25 pin connector?

I also am considering just writing a parallel port driver running at some
higher rate than the servo thread that would allow me to communicate with
the spi device.

On Wed, Sep 18, 2019 at 9:36 AM Curtis Dutton  wrote:

> Chris thanks for the info.
>
> I'm not exactly certain at this point what precision I need. I'll be
> determining that experimentally later. I will keep in mind what you have
> said, I'll deal with it after I get my hardware running.
>
>
> On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson 
> wrote:
>
>> If you want better than 16-bits and 1KHz then you also are going to need a
>> very good quality analog signal conditioning and instrumentation
>> amplifier.
>>   The analog circuit between the A/D converts and the load cell really
>> does
>> matter.   You have to reject common-mode or the low bits will just be
>> filled with 60Hz noise.  And you can not simply use a "brick wall" filter
>> or you will not get the desired bandwidth through.  (If you are needing to
>> sample at 1K then you must care about 500Hz bandwidth.
>>
>> This matters a LOT more than which A/D chip you use.
>>
>> 22 bits is better than one part per million. Do you really need that?  It
>> will be hard to isolate all of the electrical and mechanical noise
>> especially
>>
>> On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton  wrote:
>>
>> > I have a pet project I'm starting that will involve reading a load cell
>> in
>> > as near as real time as possible and using that input for servo control.
>> >
>> > I plan on using mesa hardware to interface to the equipment. Does anyone
>> > know of a nice way to interface hotsmot2 to a load cell?
>> >
>> > I would like a fairly high resolution. 16 bit minimum, the higher the
>> > better up to 22 bit I suppose. As well as the ability to sample at the
>> > servo thread rate of 1 Khz.
>> >
>> >
>> > Thanks,
>> >Curtis
>> >
>> > ___
>> > Emc-users mailing list
>> > Emc-users@lists.sourceforge.net
>> > https://lists.sourceforge.net/lists/listinfo/emc-users
>> >
>>
>>
>> --
>>
>> Chris Albertson
>> Redondo Beach, California
>>
>> ___
>> Emc-users mailing list
>> Emc-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/emc-users
>>
>

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-18 Thread Curtis Dutton
Chris thanks for the info.

I'm not exactly certain at this point what precision I need. I'll be
determining that experimentally later. I will keep in mind what you have
said, I'll deal with it after I get my hardware running.


On Tue, Sep 17, 2019 at 6:17 PM Chris Albertson 
wrote:

> If you want better than 16-bits and 1KHz then you also are going to need a
> very good quality analog signal conditioning and instrumentation amplifier.
>   The analog circuit between the A/D converts and the load cell really does
> matter.   You have to reject common-mode or the low bits will just be
> filled with 60Hz noise.  And you can not simply use a "brick wall" filter
> or you will not get the desired bandwidth through.  (If you are needing to
> sample at 1K then you must care about 500Hz bandwidth.
>
> This matters a LOT more than which A/D chip you use.
>
> 22 bits is better than one part per million. Do you really need that?  It
> will be hard to isolate all of the electrical and mechanical noise
> especially
>
> On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton  wrote:
>
> > I have a pet project I'm starting that will involve reading a load cell
> in
> > as near as real time as possible and using that input for servo control.
> >
> > I plan on using mesa hardware to interface to the equipment. Does anyone
> > know of a nice way to interface hotsmot2 to a load cell?
> >
> > I would like a fairly high resolution. 16 bit minimum, the higher the
> > better up to 22 bit I suppose. As well as the ability to sample at the
> > servo thread rate of 1 Khz.
> >
> >
> > Thanks,
> >Curtis
> >
> > ___
> > Emc-users mailing list
> > Emc-users@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/emc-users
> >
>
>
> --
>
> Chris Albertson
> Redondo Beach, California
>
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-17 Thread Chris Albertson
If you want better than 16-bits and 1KHz then you also are going to need a
very good quality analog signal conditioning and instrumentation amplifier.
  The analog circuit between the A/D converts and the load cell really does
matter.   You have to reject common-mode or the low bits will just be
filled with 60Hz noise.  And you can not simply use a "brick wall" filter
or you will not get the desired bandwidth through.  (If you are needing to
sample at 1K then you must care about 500Hz bandwidth.

This matters a LOT more than which A/D chip you use.

22 bits is better than one part per million. Do you really need that?  It
will be hard to isolate all of the electrical and mechanical noise
especially

On Tue, Sep 17, 2019 at 2:26 PM Curtis Dutton  wrote:

> I have a pet project I'm starting that will involve reading a load cell in
> as near as real time as possible and using that input for servo control.
>
> I plan on using mesa hardware to interface to the equipment. Does anyone
> know of a nice way to interface hotsmot2 to a load cell?
>
> I would like a fairly high resolution. 16 bit minimum, the higher the
> better up to 22 bit I suppose. As well as the ability to sample at the
> servo thread rate of 1 Khz.
>
>
> Thanks,
>Curtis
>
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>


-- 

Chris Albertson
Redondo Beach, California

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-17 Thread andy pugh
On Tue, 17 Sep 2019 at 22:41, andy pugh  wrote:

This .comp file shows how to control/decode data gathered by the Mesa BSPI
> driver:
>
> https://github.com/LinuxCNC/linuxcnc/blob/master/src/hal/drivers/mesa_7i65.comp
>

(Note that the Mesa 7i65 also has analog-digital conversion, but again only
12 bits, but that should make it a good basis for a component to decode SPI
analog inputs)

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is designed
for the especial use of mechanical geniuses, daredevils and lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users


Re: [Emc-users] ADC - realtime for load cell

2019-09-17 Thread andy pugh
On Tue, 17 Sep 2019 at 22:26, Curtis Dutton  wrote:

>
> I would like a fairly high resolution. 16 bit minimum, the higher the
> better up to 22 bit I suppose. As well as the ability to sample at the
> servo thread rate of 1 Khz.


The obvious answer is probably:
http://store.mesanet.com/index.php?route=product/product&product_id=126&search=analog
Except that does not have your required resolution.

I am using one of these in a project:
https://www.ebay.co.uk/itm/ADS1256-24-Bit-8-Channel-ADC-AD-Module-High-Precision-Data-Acquisition-Card-BSG/122917917245

Output is SPI, and there is an SPI input module for the Mesa cards. (though
the HAL driver side isn't super-convenient,
This .comp file shows how to control/decode data gathered by the Mesa BSPI
driver:
https://github.com/LinuxCNC/linuxcnc/blob/master/src/hal/drivers/mesa_7i65.comp


-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is designed
for the especial use of mechanical geniuses, daredevils and lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1916

___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users