Nikon Ti Eclipse XY stage calibration Error

[Balaji.ncbs]

Dear Nico,

We are trying to run a HCS plugin on the nikon Ti Eclipse and we are facing a few problems in calibrating the XY stage movement.

Firstly how the plugin works is that,It initializes a matrix of values for a 96/384 well plate. Now calibration is simple, if u go to your first well in the plate and hit the calibrate button, the plugin replaces the current value of the stage to the default starting value made by the plugin,

say for example,

Everytime when you start the plugin,
Well A1: X: 10840 Y: 2790 in the xy plane.

The stage on its own can have its own set of values,
Well A1: X: 11470 Y: 3850 in the xy plane.

Now to calibrate the stage,I now replace the values of the stage position to X: 10840 Y: 2790.
Thus the start up positions for the stage and the plugin matrix are the same and the movements can be synchronized.

This methodology works with prior stages but the fails when it comes to Nikon Ti XYDrive. The replacement of values does not work.What do you suggest?

Please do let me know IF my explanation doesnt make sense.

Balaji

[Nico Stuurman-4]

Hi Balaji,


> We are trying to run a HCS plugin on the nikon Ti Eclipse and we are facing
> a few problems in calibrating the XY stage movement.
>
> Firstly how the plugin works is that,It initializes a matrix of values for a
> 96/384 well plate. Now calibration is simple, if u go to your first well in
> the plate and hit the calibrate button, the plugin replaces the current
> value of the stage to the default starting value made by the plugin,

Since we can not see the source code of your plugin, we do no know what is happening (b.t.w., it would be really good to make that code open source as soon as possible!).  However, looking in the Nikon TI adapter code, I see that the function:
"setOrigin()" is not implemented.  On the other hand, the function "SetAdapterOrigin (x, y)" is implemented.  So, if you can change your code so that it uses the latter, you should be good to go (and yes, we should implement setOrigin() in the TI adapter).

B.t.w., are you still using 1.3 or is this using 1.4?

Best,

Nico



------------------------------------------------------------------------------
RSA® Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general

[Balaji.ncbs]

Dear Nico,

We are actually using 1.3 still.We will be moving to 1.4 in a week or so.We would like to test this plugin on our slides and then make it available based on jitu's suggestions.

And yes we do use setOrigin() as of now.Will change the same to SetAdapterOrigin (x, y) rightaway and check the same.

Thank you,
Balaji

On Tue, Nov 1, 2011 at 8:34 PM, Nico Stuurman-4 [via Micro-Manager] <[hidden email]> wrote:

Hi Balaji,


> We are trying to run a HCS plugin on the nikon Ti Eclipse and we are facing
> a few problems in calibrating the XY stage movement.
>
> Firstly how the plugin works is that,It initializes a matrix of values for a
> 96/384 well plate. Now calibration is simple, if u go to your first well in
> the plate and hit the calibrate button, the plugin replaces the current
> value of the stage to the default starting value made by the plugin,

Since we can not see the source code of your plugin, we do no know what is happening (b.t.w., it would be really good to make that code open source as soon as possible!).  However, looking in the Nikon TI adapter code, I see that the function:
"setOrigin()" is not implemented.  On the other hand, the function "SetAdapterOrigin (x, y)" is implemented.  So, if you can change your code so that it uses the latter, you should be good to go (and yes, we should implement setOrigin() in the TI adapter).

B.t.w., are you still using 1.3 or is this using 1.4?

Best,

Nico



------------------------------------------------------------------------------
RSA&reg; Conference 2012
Save &#36;700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general

---

If you reply to this email, your message will be added to the discussion below:

http://micro-manager.3463995.n2.nabble.com/Nikon-Ti-Eclipse-XY-stage-calibration-Error-tp6951759p6951986.html

To unsubscribe from Nikon Ti Eclipse XY stage calibration Error, click here.

[Balaji.ncbs]

Dear Nico.

The change to setAdapterOrigin(x,y) did not work.We use Micromanager 1.3.
Even seperately we tried to run the same command on Micromanager 1.4. It is able to detect the driver when I say,

mmc.getXYStageDevice()

But when I try to reset the origin using,

mmc.setAdapterOriginXY(mmc.getXYStageDevice(), 14380.0, 11240.0);

these values do not get reset in the Ti-control.

Please do let me know if you need anything else regarding the same.
Else if sending the code, is the only way to work our way around this,I can arrange for that as well by tomorrow.

Best,
Balaji

On Tue, Nov 1, 2011 at 8:40 PM, Balaji Ramalingam <[hidden email]> wrote:

Dear Nico,

We are actually using 1.3 still.We will be moving to 1.4 in a week or so.We would like to test this plugin on our slides and then make it available based on jitu's suggestions.

And yes we do use setOrigin() as of now.Will change the same to SetAdapterOrigin (x, y) rightaway and check the same.

Thank you,
Balaji

On Tue, Nov 1, 2011 at 8:34 PM, Nico Stuurman-4 [via Micro-Manager] <[hidden email]> wrote:

Hi Balaji,


> We are trying to run a HCS plugin on the nikon Ti Eclipse and we are facing
> a few problems in calibrating the XY stage movement.
>
> Firstly how the plugin works is that,It initializes a matrix of values for a
> 96/384 well plate. Now calibration is simple, if u go to your first well in
> the plate and hit the calibrate button, the plugin replaces the current
> value of the stage to the default starting value made by the plugin,

Since we can not see the source code of your plugin, we do no know what is happening (b.t.w., it would be really good to make that code open source as soon as possible!).  However, looking in the Nikon TI adapter code, I see that the function:
"setOrigin()" is not implemented.  On the other hand, the function "SetAdapterOrigin (x, y)" is implemented.  So, if you can change your code so that it uses the latter, you should be good to go (and yes, we should implement setOrigin() in the TI adapter).

B.t.w., are you still using 1.3 or is this using 1.4?

Best,

Nico



------------------------------------------------------------------------------
RSA&reg; Conference 2012
Save &#36;700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general

---

If you reply to this email, your message will be added to the discussion below:

http://micro-manager.3463995.n2.nabble.com/Nikon-Ti-Eclipse-XY-stage-calibration-Error-tp6951759p6951986.html

To unsubscribe from Nikon Ti Eclipse XY stage calibration Error, click here.

[Nico Stuurman-4]

On Nov 2, 2011, at 7:24 AM, Balaji.ncbs wrote:

> The change to setAdapterOrigin(x,y) did not work.We use Micromanager 1.3.
> Even seperately we tried to run the same command on Micromanager 1.4. It is able to detect the driver when I say,
>
> mmc.getXYStageDevice()
>
> But when I try to reset the origin using,
>
> mmc.setAdapterOriginXY(mmc.getXYStageDevice(), 14380.0, 11240.0);
>
> these values do not get reset in the Ti-control.

What do you mean with "Ti-control"?  The separate application from Nikon?  If so, that is exactly what you should expect.  The adapter keeps its own coordinate system that can be different from the one in TI-control.  

However, after the setAdapterOriginXY call as above, if you now send the stage to position 0.0, 0.0, it should go to what TI-control labels as position  14380.0, 11240.0.

So, I simply do not know what you mean with "it did not work"....

> Please do let me know if you need anything else regarding the same.
> Else if sending the code, is the only way to work our way around this,I can arrange for that as well by tomorrow.

Things are always easier if everyone can look at the code.

Best,

Nico



------------------------------------------------------------------------------
RSA&#174; Conference 2012
Save $700 by Nov 18
Register now&#33;
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general

[Holden Seamus John]

Hi

Im trying to do fast (1ms) acquistion on a couple of EMCCD cameras I
have been demoing. I observe “"Error: Circular buffer overflowed",
followed by abort of acquisition, when trying to acquire >2000 frames at
1ms time resolution ("AbsoluteInterval") 64x128 pixels.

Reproduced under multiple conditions:
64bit MM, Photometrics Evolve 128. With conserve ram off and on, saving
images to ram or to disl. Tried increasing ImageJ and sequence buffer
memory availability.
32 bit MM, Andor Ixon3 860. Same conditions as above.
Windows 7, 8GB RAM.
Nightly build 1.4.7 20111003.

The associated CoreLog has been submitted.

Any advice appreciated.

Thanks
Seamus


------------------------------------------------------------------------------
RSA&#174; Conference 2012
Save $700 by Nov 18
Register now&#33;
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general

[Nico Stuurman-4]

On Nov 2, 2011, at 9:12 AM, Seamus Holden wrote:

> Im trying to do fast (1ms) acquistion on a couple of EMCCD cameras I  have been demoing. I observe “"Error: Circular buffer overflowed",  followed by abort of acquisition, when trying to acquire >2000 frames at  1ms time resolution ("AbsoluteInterval") 64x128 pixels.
>
> Reproduced under multiple conditions: 64bit MM, Photometrics Evolve 128. With conserve ram off and on, saving
> images to ram or to disl. Tried increasing ImageJ and sequence buffer memory availability.
> 32 bit MM, Andor Ixon3 860. Same conditions as above.

Interesting.  Each image is only 64x128x2= 16384 bytes in size.  2000 of these should easily fit into a circular buffer (as long as it is > ~35 MB in size).

I can completely reproduce this problem by setting the size of the demo camera to 64x128, 16bit, even with a circular buffer size of 200 MB.  I suspect the issue is related to metadata storage. In any case, we should be able to figure this out.

Best,

Nico

> Windows 7, 8GB RAM.
> Nightly build 1.4.7 20111003.
>
> The associated CoreLog has been submitted.
>
> Any advice appreciated.
>
> Thanks
> Seamus
>
>
> ------------------------------------------------------------------------------
> RSA&#174; Conference 2012
> Save $700 by Nov 18
> Register now&#33;
> http://p.sf.net/sfu/rsa-sfdev2dev1
> _______________________________________________
> micro-manager-general mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/micro-manager-general

------------------------------------------------------------------------------
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general

[Nico Stuurman-4]

On Nov 2, 2011, at 10:18 PM, Nico Stuurman wrote:

> On Nov 2, 2011, at 9:12 AM, Seamus Holden wrote:
>
>> Im trying to do fast (1ms) acquistion on a couple of EMCCD cameras I  have been demoing. I observe “"Error: Circular buffer overflowed",  followed by abort of acquisition, when trying to acquire >2000 frames at  1ms time resolution ("AbsoluteInterval") 64x128 pixels.
>>
>> Reproduced under multiple conditions: 64bit MM, Photometrics Evolve 128. With conserve ram off and on, saving
>> images to ram or to disl. Tried increasing ImageJ and sequence buffer memory availability.
>> 32 bit MM, Andor Ixon3 860. Same conditions as above.
>
> Interesting.  Each image is only 64x128x2= 16384 bytes in size.  2000 of these should easily fit into a circular buffer (as long as it is > ~35 MB in size).
>
> I can completely reproduce this problem by setting the size of the demo camera to 64x128, 16bit, even with a circular buffer size of 200 MB.  

Turns out that the circular buffer used in Micro-Manager had a "sanity" check and never allocated space for more than 1000 images.  I increased that limit to 10000 and that fixed the problem with the demo camera.  I suspect that will also fix the issue for you (as long as you do not acquire more than about 12000 images at that speed, you should also make the circular buffer big enough to hold the images that you want to acquire).  If needed, we can increase this maximum further (I do not like taking it away altogether as it may protect the application from crazy situations).  

This change should be in tomorrow's nightly build.

Best,

Nico



------------------------------------------------------------------------------
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general

[John Oreopoulos]

I've actually been acquiring 25000 frames at 30fps for a few applications recently. Occasionally I have run into the circular buffer error as well and was going to report it after the Neuroscience meeting after having some more time to explore the conditions under which it occurs. Since it's being discussed now, I thought it best to mention my experience with it as well. So will the update allow the user to set the size of the buffer? Thanks for addressing the issue.

John Oreopoulos

On 2011-11-05, at 11:58 PM, Nico Stuurman <[hidden email]> wrote:

>
> On Nov 2, 2011, at 10:18 PM, Nico Stuurman wrote:
>
>> On Nov 2, 2011, at 9:12 AM, Seamus Holden wrote:
>>
>>> Im trying to do fast (1ms) acquistion on a couple of EMCCD cameras I  have been demoing. I observe “"Error: Circular buffer overflowed",  followed by abort of acquisition, when trying to acquire >2000 frames at  1ms time resolution ("AbsoluteInterval") 64x128 pixels.
>>>
>>> Reproduced under multiple conditions: 64bit MM, Photometrics Evolve 128. With conserve ram off and on, saving
>>> images to ram or to disl. Tried increasing ImageJ and sequence buffer memory availability.
>>> 32 bit MM, Andor Ixon3 860. Same conditions as above.
>>
>> Interesting.  Each image is only 64x128x2= 16384 bytes in size.  2000 of these should easily fit into a circular buffer (as long as it is > ~35 MB in size).
>>
>> I can completely reproduce this problem by setting the size of the demo camera to 64x128, 16bit, even with a circular buffer size of 200 MB.  
>
> Turns out that the circular buffer used in Micro-Manager had a "sanity" check and never allocated space for more than 1000 images.  I increased that limit to 10000 and that fixed the problem with the demo camera.  I suspect that will also fix the issue for you (as long as you do not acquire more than about 12000 images at that speed, you should also make the circular buffer big enough to hold the images that you want to acquire).  If needed, we can increase this maximum further (I do not like taking it away altogether as it may protect the application from crazy situations).  
>
> This change should be in tomorrow's nightly build.
>
> Best,
>
> Nico
>
>
>
> ------------------------------------------------------------------------------
> RSA(R) Conference 2012
> Save $700 by Nov 18
> Register now
> http://p.sf.net/sfu/rsa-sfdev2dev1
> _______________________________________________
> micro-manager-general mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/micro-manager-general

------------------------------------------------------------------------------
RSA(R) Conference 2012
Save $700 by Nov 18
Register now
http://p.sf.net/sfu/rsa-sfdev2dev1
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general