Hamamatsu ORCA-ER Corrupted images with 1x1 binning, OK with 2x2 binning

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

Hamamatsu ORCA-ER Corrupted images with 1x1 binning, OK with 2x2 binning

abasu
Hello,

I have a Hamamatsu ORCA-ER CCD camera, and I am having trouble getting it to
work with Micromanager.  

The problem is that the images are corrupted when I have it set at 1x1
Binning.  See the two examples below.  It almost looks like a
synchronization issue
<http://micro-manager.3463995.n2.nabble.com/file/t397005/capture1.png>
<http://micro-manager.3463995.n2.nabble.com/file/t397005/Capture2.png>

Here's the strange part: If I set the binning to 2x2, 4x4, or 8x8, the
corruption disappears.  See below.
<http://micro-manager.3463995.n2.nabble.com/file/t397005/capture3-2x2binning.png>  


Any idea what the problem may be? Details of my system are below.

Camera: Hamamatsu ORCA-ER AG, model number C4742-80-12AG
Micromanager: 1.4.22 (most recent as of Nov 2017)
Hamamatsu DCAM-API: 17.10.5372 (most recent as of Nov 2017)
Firewire card: Texas Instruments 1394 OHCI Compliant Host Controller (shown
in device manager)
OS: Windows 7 64 bit (fresh install Nov 2017).


Thanks a lot for your help!

Best wishes,
Amar



--
Sent from: http://micro-manager.3463995.n2.nabble.com/

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general
Reply | Threaded
Open this post in threaded view
|

Re: Hamamatsu ORCA-ER Corrupted images with 1x1 binning, OK with 2x2 binning

abasu
I also notice that if I select  very small regions of interest (maybe a
quarter of the image size), the problem also goes away.   However, I'd like
to take images at full resolution.  I appreciate your help!

Amar



--
Sent from: http://micro-manager.3463995.n2.nabble.com/

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general
Reply | Threaded
Open this post in threaded view
|

Re: Hamamatsu ORCA-ER Corrupted images with 1x1 binning, OK with 2x2 binning

Ti_Sapph
My guess is the PCIe slot that your framegrabber card is installed in does not run at the necessary speed.  Likely it's sharing bandwidth on the bus with another PCIe slot that has a card installed.  Moving the framegrabber card to another slot, especially one rated for a higher bandwidth (ie a x8 or x16 even if the card will fit in a x4) is usually a good bet.  

We had a similar issue with a Flash4.0.  Ended up needing to swap cards around for a while to find a setting that worked, but it was possible to get 2 of these to run at full resolution+speed on a standard desktop PC.  

If you can't get a slot that works and you can sacrifice a bit of speed, there could be a setting for something like 'Read Mode' in the Device Manager.  Changing this from a '1' to a '2' (or maybe the other way around) changes the speed of the pixel clock and therefore the necessary bandwidth on the PCIe bus.  You can take full-resolution images, albeit at a reduced rate.  This was a work-around on our Flash4.0 and let us run at 40 fps rather than 100 fps at full resolution.

Thanks,
Rusty

On Tue, Nov 21, 2017 at 5:26 PM, abasu <[hidden email]> wrote:
I also notice that if I select  very small regions of interest (maybe a
quarter of the image size), the problem also goes away.   However, I'd like
to take images at full resolution.  I appreciate your help!

Amar



--
Sent from: http://micro-manager.3463995.n2.nabble.com/

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general
Reply | Threaded
Open this post in threaded view
|

Re: Hamamatsu ORCA-ER Corrupted images with 1x1 binning, OK with 2x2 binning

abasu
Dear Rusty,

Thanks very much for your response.  I wasn't expecting someone to be able
to figure out my obscure problem, and I'm very grateful for your help.

Today I tried acquisition using another computer that had a full PCI card
rather than a PCI-E card, and the images came out fine.  Thank you!

However, one thing that still puzzles me is that another PC which had a full
PCI card had the same problem.

I'm curious which firewire card did you end up using?  Hamamatsu's DCAM-API
documentation suggests using a firewire card that uses the LSI FW643 chip.
I don't have such a card, but wondering if I should get one.

Thanks again for your reply!
Amar




--
Sent from: http://micro-manager.3463995.n2.nabble.com/

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general
Reply | Threaded
Open this post in threaded view
|

Re: Hamamatsu ORCA-ER Corrupted images with 1x1 binning, OK with 2x2 binning

Ti_Sapph
I do not have any Firewire cameras.  The issue I experienced was with CamLink framegrabber cards.  I have heard that a 'Texas Instrument chipset' is necessary for optimal operation for a Firewire framegrabber but can't suggest a specific part to use.

The way peripheral bus bandwidth is divided on computers is often obscure.  It depends on the CPU and motherboard design, with the necessary details being difficult to discern without going deep into data sheets (I tend to use v similar instrument PC builds repeatedly since I found a combo that works).  On top of that, which slots are occupied by other peripherals makes a difference in the performance of a given slot and often in non-obvious ways.  Basically, it takes trial and error to find a slot on a given PC that works.  This can mean pulling all peripherals off of the motherboard and getting the framegrabber slot set first before replacing other less bandwidth-greedy parts until you have a stable configuration.  All of this is hard to work out a priori.  Trying things until it works and then not changing it too much is sadly the usual way to go.

Thanks,
Rusty

On Tue, Nov 28, 2017 at 6:43 PM, abasu <[hidden email]> wrote:
Dear Rusty,

Thanks very much for your response.  I wasn't expecting someone to be able
to figure out my obscure problem, and I'm very grateful for your help.

Today I tried acquisition using another computer that had a full PCI card
rather than a PCI-E card, and the images came out fine.  Thank you!

However, one thing that still puzzles me is that another PC which had a full
PCI card had the same problem.

I'm curious which firewire card did you end up using?  Hamamatsu's DCAM-API
documentation suggests using a firewire card that uses the LSI FW643 chip.
I don't have such a card, but wondering if I should get one.

Thanks again for your reply!
Amar




--
Sent from: http://micro-manager.3463995.n2.nabble.com/

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general
Reply | Threaded
Open this post in threaded view
|

Re: Hamamatsu ORCA-ER Corrupted images with 1x1 binning, OK with 2x2 binning

abasu
Hi Randy,

I tried a few different firewire cards I had in the lab, and it turned out
the one that worked had an LSI chipset.   The model number was LS2-FAE10.

Thanks again for your help.  It is a strange problem.. but for me I found
something that works, at least for one camera.

I am having trouble doing dual camera acquisition, and made a separate post
on that.  Any suggestions would be welcome.

Best,
AB



--
Sent from: http://micro-manager.3463995.n2.nabble.com/

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
micro-manager-general mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/micro-manager-general