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

Hi confocal list,

First, thanks to everyone for their wonderful suggestions. 
Both commercial entities and scientists had lots of nice advice.

As for shuttering, I'm currently doing exactly what Aryeh 
mentions: Using the direct chip-exposure hardware trigger 
from the camera (CoolSnapHQ) to control a Uniblitz shutter 
so that no software is invovled. (We use some OR and AND 
gates on a breadboard as Aryeh described to control when the 
shutter is responsive to the exposure signal of the 
CoolSnap.) The CoolSnap can also be told to use a pre-set 
delay between shutter opening and exposure, which we set at 
5msec (the Uniblitz is rated for 3msec). Since we use a blue 
LED as our excitation source, I could just turn that on and 
off directly too (we were using a Hg lamp before). I think 
the take-home message is that it is generally better to have 
the camera directly control the shutter/LED rather than 
worry about software synchronization, which will never be as 
precise.

Since Cameron asked, I'm using custom Matlab software to 
control the stage, camera, shutters, filters, etc. The PVCAM 
C interface makes it easy to directly control the camera. I 
am confident that the software is taking exposures 
accurately. In particular, I'm taking about 200msec 
exposures with 3x3 binning (blue LED, 60x/0.7NA objective, 
CoolSnapHQ) every 10 minutes and the cells stay viable but 
the signal is weak making it difficult for me to trace fine 
structures in my cultured neurons. I'm working on changing 
to a better objective and, as you know, a more sensitive CCD.


Still, I would welcome any further ideas on experiences with 
different EMCCDs (not from commercial vendors) and ideas on 
how to test EMCCDs. Right now, for demoing, I'm hoping to 
take some dark frames at various exposure times (dark noise) 
and some pictures of fluorescent beads with a set exposure 
time (overall QE test). I could also test the read noise 
with some bias images, but I'm hoping that these EMCCDs will 
have no issue there.

Thanks again for all the help... I've certainly learned a 
lot from your replies! Best,

- Neville

Aryeh Weiss wrote:
> Search the CONFOCAL archive at
> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
> 
> Nuno Moreno wrote:
> 
>> Search the CONFOCAL archive at
>> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>>
>> Hi!
>>
>> If it is for cell imaging you might consider a BT camera.
>>
>> Forcing the previous point, which I thing is very important:
>>
>> I'm having the same problem with an ORCA AG with ImagePro Plus.  What 
>> I have now is a small program called by macros that controls an 
>> external shutter via TTL, which makes a delay till it opens the 
>> shutter. I found out that the driver takes almost 500ms to get the 
>> camera taking a picture which is unacceptable! However this solution 
>> is far from being elegant and sometimes the pseudo synchronization 
>> fails! Is this a Hamamatsu problem? With a Hamamatsu BT1024 which has 
>> a QE>90% it happends the same thing with metamorph.
>>
>> Regarding to the previous commercial response, does anyone have a 
>> cheaper and more flexible solution?
>>
> 
> The way to do this is by directly controlling the shutter from the 
> chip-exposure line of the camera. This bypasses the software completely. 
> Most high end cameras  have such a control line, although it goes by 
> various names. I know that the PCO sensicam series, Roper Coolsnap 
> series, Orca AG (probably the other Orcas also),  QImaging, all have 
> such a line. I use it with a Uniblitz shutter. You still have to worry 
> about the the 5-10ms shutter opening and closing time. In a camera like 
> the PCO (again, probably in others)  the camera can be set up to start 
> integrating after a delay time set by the user. This increases your 
> sample exposure by 10-20ms, but that is really not bad compared to the 
> 1-2 seconds that some software controlled systems waste before closing 
> the shutter.
> 
> Of course, if you connect the control line directly to a system like 
> CoolLed or an AOTF, you can get very fast switching.
> 
> You may need to build a little bit of digital logic. For example, when 
> doing a live preview for focusing, we want the shutter to remain open 
> continuously. But when using the chip expose line, the shutter will 
> cycle rapidly as the camera switches between integration and image 
> transfer. Therefore, we set it up so that in live preview mode the 
> Uniblitz system was commanded to stay open independent of the hardware 
> control line. I also added an external switch that could force the input 
> to the Uniblitz control to be on (ie, I OR'ed the camera control line 
> with a manually switched voltage). This was a matter of a few NAND gates 
> on two 74LSxx series integrated circuits (your electronics shop will 
> certainly have these).
> 
> I know that this solution is used in other labs, so maybe I am missing 
> something in this thread about the difficulty that is being discussed.
> 
> --aryeh