Frame shift in home-built two-photon laser scanning microscopy software

[Dan Fu-2]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Dear all,

We have a home-built two photon laser scanning microscopy. For that, we
wrote our own software in Labwindows/CVI. Bascially it controlls the DAQ
card's two output channels (for controlling the scan mirrors from
Cambridge Technology), two input channels (for collecting data) and the
motion controller (NEWPORT ESP7000, for moving the sample). The way we
did it is like this:
1. Make sawtooth waveform for a single frame and send that to the two
output channels.
2. Start output task and input task with a trigger.
3. Acquire data for that frame.
4. Stop output task and input task
5. Move to another position and repeat 1-4.
However, problem arises when we compare images at slightly different z
positions or want to average a few frames. They don't match up quite
well. There is a shift in the lateral or vertical position of about
5-10pixels. I was wondering if anybody had any experience with this. If
yes, is this an inherent problem with the scan mirror? Are there ways to
improve it?
Thank you for your time.

Best Regards,
Dan

---

Dan Fu
Dept. of Chemistry, Duke University
FFSC Room 2305, Box 90346
124 Science Drive
Durham, NC 27708
Tel: [phone]; Fax: [phone]

[Craig Brideau]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

We had a problem with our own home-brew rig that way.  The galvo scan
pattern gets out of synch with the data acquisition so the images
become shifted.  I wrote a flag into the hardware code for the
scanners that would be tripped at the start of a new frame.  If that
flag was not tripped when the software thought a new frame should
start, it waited until the flag was tripped before acquiring.  This
ensured that the galvos were in the right position for the start of
frame when the data acquisition software thought it was at the start
of the frame.  Otherwise it's possible to get ahead or behind the
galvos.

Hope this helps!

Craig

On 9/26/07, Dan Fu <[hidden email]> wrote:

> Search the CONFOCAL archive at
> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>
> Dear all,
>
> We have a home-built two photon laser scanning microscopy. For that, we
> wrote our own software in Labwindows/CVI. Bascially it controlls the DAQ
> card's two output channels (for controlling the scan mirrors from
> Cambridge Technology), two input channels (for collecting data) and the
> motion controller (NEWPORT ESP7000, for moving the sample). The way we
> did it is like this:
> 1. Make sawtooth waveform for a single frame and send that to the two
> output channels.
> 2. Start output task and input task with a trigger.
> 3. Acquire data for that frame.
> 4. Stop output task and input task
> 5. Move to another position and repeat 1-4.
> However, problem arises when we compare images at slightly different z
> positions or want to average a few frames. They don't match up quite
> well. There is a shift in the lateral or vertical position of about
> 5-10pixels. I was wondering if anybody had any experience with this. If
> yes, is this an inherent problem with the scan mirror? Are there ways to
> improve it?
> Thank you for your time.
>
> Best Regards,
> Dan
>
> ---
>
> Dan Fu
> Dept. of Chemistry, Duke University
> FFSC Room 2305, Box 90346
> 124 Science Drive
> Durham, NC 27708
> Tel: [phone]; Fax: [phone]
>

[Dan Fu-2]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Thanks for your input, Craig. I think I am doing that right now. Each new
frame I start a new acqusition by using a new trigger. The ouput task and
input task use the same clock (although at different frequencies) so the
synchronization should be ok.
Does the gavlo mirror needs sometime to settle down at a certain speed? For
example, if I send 1 Volt to the galvo (let's say it was at -1 V), would it
be precisely at 1 V position by some loop control or it is at some random
position close to 1V?
Any input would be appreciated.

Best,
Dan

[Craig Brideau]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Galvos can typically be driven to a kHz or two, but your mileage may
vary depending on the mirror weight and several other mechanical
factors.  If you are trying to scan out more than 1000 lines per
second you may be bumping up against inertia.  Have you tried slowing
down your scan to see what happens?

Craig

On 9/27/07, Dan Fu <[hidden email]> wrote:

> Search the CONFOCAL archive at
> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>
> Thanks for your input, Craig. I think I am doing that right now. Each new
> frame I start a new acqusition by using a new trigger. The ouput task and
> input task use the same clock (although at different frequencies) so the
> synchronization should be ok.
> Does the gavlo mirror needs sometime to settle down at a certain speed? For
> example, if I send 1 Volt to the galvo (let's say it was at -1 V), would it
> be precisely at 1 V position by some loop control or it is at some random
> position close to 1V?
> Any input would be appreciated.
>
> Best,
> Dan
>