Leica SP8 confocal - 5 colour setup takes a very long time or is crashing

[Angela Kurz-2]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Hello everyone,

We are currently experiencing problems with the Leica SP8 confocal system.
If users are attempting to set up a 5 colour run, the setup takes a very
long time (moving red progress bars when adding each colour).

Observations:

   - CPU at 60-80% when it's doing that
   - RAM usage is low
   - HDD activity is nigh on 0 (we swapped temp directory to the SSD, no
   improvement (or minimal)

I am wondering if it is the software struggling with the overlapping
filtering?
Does anyone else have these problems or has anyone found a solution?

Thank you,
Angela

--

*Dr. ANGELA KURZ*
Bioimaging Specialist | Research Officer

*Centenary Imaging | Sydney Cytometry*
*CENTENARY INSTITUTE*

*Mail* | Locked Bag 6 | Newtown | NSW | 2042 | Australia
*Courier* | Building 93 Royal Prince Alfred Hospital | Missenden Road |
Camperdown | NSW | 2050 | Australia

*E* [hidden email] <[hidden email]> | *T* +61 2 9565
61 <+61%202%209565%206173>880 |  *W *www.centenary.org.au

--
 <https://www.centenary.org.au>

[George McNamara]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Hi Angela,

Make sure you are not using sequential line mode.

I suggest (using sequential frame, single Z plane), create 5 scan tracks
of the same color from scratch, with identical emission band for each.
That is, start with just a single laser -> emission band (say
500-540nm), time it. Then duplicate, time, that. Make third, time it,
fourth, time it, fifth, time it. My expectation: all fast. then change
emission band say to 500-530nm for #5, time that. Then change others one
at a time, time each run (i.e. next up #4 to 500-520). If timing with
all five different (but close together), then can change (one at a time)
to ultimately match the user's settings, but check performance each
time. Close all Leica container data between runs - that is, new Leica
container file each time.

Make sure: no other application running (turn off antivirus, also unplug
the PC from the building network). Windows has lots of applications and
processes running - many can be killed with Task Manager (the Leica
stuff all have the same logo).

George

p.s. master beamsplitter should be the same for each scan track.

On 5/24/2020 6:49 PM, Angela Kurz wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> Hello everyone,
>
> We are currently experiencing problems with the Leica SP8 confocal system.
> If users are attempting to set up a 5 colour run, the setup takes a very
> long time (moving red progress bars when adding each colour).
>
> Observations:
>
>     - CPU at 60-80% when it's doing that
>     - RAM usage is low
>     - HDD activity is nigh on 0 (we swapped temp directory to the SSD, no
>     improvement (or minimal)
>
> I am wondering if it is the software struggling with the overlapping
> filtering?
> Does anyone else have these problems or has anyone found a solution?
>
> Thank you,
> Angela
>

[Steffen Dietzel]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Hi Angela,

nice to read from you!

Can you go into more detail how you are doing this? Probably sequential
mode, but line or frame sequential? How many different detectors for the
five colors are used?

If possible, it seems fastest to set up all channels with their
wavelength-range first and only then activate sequential scanning. So
that the mirrors do not need any movement between channels. If you need
different wavelength-range-settings (in frame sequential, if you need to
reuse a detector), try to keep the moving of mirrors to a minimum, i.e.
if you have say two wavelength settings and each of them is needed
twice, make sure you have the same setting-sequences after each other,
not intermingled.

And, finally, try to set up the channels from scratch. I have the
impression that on rare occasions, the system is introducing some
unnecessary extra loops. And once it starts with that it seems to be
easier to start over, instead of trying to repair the broken setting.

Good luck

Steffen

Am 25.05.2020 um 00:49 schrieb Angela Kurz:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> Hello everyone,
>
> We are currently experiencing problems with the Leica SP8 confocal system.
> If users are attempting to set up a 5 colour run, the setup takes a very
> long time (moving red progress bars when adding each colour).
>
> Observations:
>
>     - CPU at 60-80% when it's doing that
>     - RAM usage is low
>     - HDD activity is nigh on 0 (we swapped temp directory to the SSD, no
>     improvement (or minimal)
>
> I am wondering if it is the software struggling with the overlapping
> filtering?
> Does anyone else have these problems or has anyone found a solution?
>
> Thank you,
> Angela
>

--
------------------------------------------------------------
Steffen Dietzel, PD Dr. rer. nat
Ludwig-Maximilians-Universität München
Biomedical Center (BMC)
Head of the Core Facility Bioimaging

Großhaderner Straße 9
D-82152 Planegg-Martinsried
Germany

http://www.bioimaging.bmc.med.uni-muenchen.de

[Tina Carvalho-2]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Hi,

Yeah, what Steffen said. Set up each sequence, one at a time, test it well
before moving on to the next sequence. I often avoid Dye Assistant when I'm
doing this. Do frame sequence. I've only ever tried using a different
detector for each, but I bet you could double up.

And since they are stuck at home, the helpful Leica applications people are
waiting for your call!

Aloha,
Tina

On Tue, May 26, 2020 at 6:23 AM Steffen Dietzel <[hidden email]> wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> Hi Angela,
>
> nice to read from you!
>
> Can you go into more detail how you are doing this? Probably sequential
> mode, but line or frame sequential? How many different detectors for the
> five colors are used?
>
> If possible, it seems fastest to set up all channels with their
> wavelength-range first and only then activate sequential scanning. So
> that the mirrors do not need any movement between channels. If you need
> different wavelength-range-settings (in frame sequential, if you need to
> reuse a detector), try to keep the moving of mirrors to a minimum, i.e.
> if you have say two wavelength settings and each of them is needed
> twice, make sure you have the same setting-sequences after each other,
> not intermingled.
>
> And, finally, try to set up the channels from scratch. I have the
> impression that on rare occasions, the system is introducing some
> unnecessary extra loops. And once it starts with that it seems to be
> easier to start over, instead of trying to repair the broken setting.
>
> Good luck
>
> Steffen
>
> Am 25.05.2020 um 00:49 schrieb Angela Kurz:
> > *****
> > To join, leave or search the confocal microscopy listserv, go to:
> > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > Post images on http://www.imgur.com and include the link in your
> posting.
> > *****
> >
> > Hello everyone,
> >
> > We are currently experiencing problems with the Leica SP8 confocal
> system.
> > If users are attempting to set up a 5 colour run, the setup takes a very
> > long time (moving red progress bars when adding each colour).
> >
> > Observations:
> >
> >     - CPU at 60-80% when it's doing that
> >     - RAM usage is low
> >     - HDD activity is nigh on 0 (we swapped temp directory to the SSD, no
> >     improvement (or minimal)
> >
> > I am wondering if it is the software struggling with the overlapping
> > filtering?
> > Does anyone else have these problems or has anyone found a solution?
> >
> > Thank you,
> > Angela
> >
> --
> ------------------------------------------------------------
> Steffen Dietzel, PD Dr. rer. nat
> Ludwig-Maximilians-Universität München
> Biomedical Center (BMC)
> Head of the Core Facility Bioimaging
>
> Großhaderner Straße 9
> D-82152 Planegg-Martinsried
> Germany
>
> http://www.bioimaging.bmc.med.uni-muenchen.de
>

--
Tina Weatherby Carvalho

University of Hawaii
Pacific Biosciences Research Center
Biological Electron Microscope Facility
2538 McCarthy Mall
Honolulu, HI 96822
808-956-6251

[Angela Kurz-2]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Hi,

Thank you everyone for your feedback.

The user, who is having problems, uses Dye Assistent, frame sequential and
uses all our 5 detectors for the 5 colours. In addition he is setting up a
huge tilescan (20GB).

We will try to set it up manually.
I was also in contact with Leica and they will set up our workstation from
scratch.

Let's see how it goes.
It wasn't like that before.

Thanks again everyone!
Angela



On Wed, 27 May 2020 at 08:35, Tina Carvalho <[hidden email]> wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> Hi,
>
> Yeah, what Steffen said. Set up each sequence, one at a time, test it well
> before moving on to the next sequence. I often avoid Dye Assistant when I'm
> doing this. Do frame sequence. I've only ever tried using a different
> detector for each, but I bet you could double up.
>
> And since they are stuck at home, the helpful Leica applications people are
> waiting for your call!
>
> Aloha,
> Tina
>
> On Tue, May 26, 2020 at 6:23 AM Steffen Dietzel <[hidden email]>
> wrote:
>
> > *****
> > To join, leave or search the confocal microscopy listserv, go to:
> > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > Post images on http://www.imgur.com and include the link in your
> posting.
> > *****
> >
> > Hi Angela,
> >
> > nice to read from you!
> >
> > Can you go into more detail how you are doing this? Probably sequential
> > mode, but line or frame sequential? How many different detectors for the
> > five colors are used?
> >
> > If possible, it seems fastest to set up all channels with their
> > wavelength-range first and only then activate sequential scanning. So
> > that the mirrors do not need any movement between channels. If you need
> > different wavelength-range-settings (in frame sequential, if you need to
> > reuse a detector), try to keep the moving of mirrors to a minimum, i.e.
> > if you have say two wavelength settings and each of them is needed
> > twice, make sure you have the same setting-sequences after each other,
> > not intermingled.
> >
> > And, finally, try to set up the channels from scratch. I have the
> > impression that on rare occasions, the system is introducing some
> > unnecessary extra loops. And once it starts with that it seems to be
> > easier to start over, instead of trying to repair the broken setting.
> >
> > Good luck
> >
> > Steffen
> >
> > Am 25.05.2020 um 00:49 schrieb Angela Kurz:
> > > *****
> > > To join, leave or search the confocal microscopy listserv, go to:
> > > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > > Post images on http://www.imgur.com and include the link in your
> > posting.
> > > *****
> > >
> > > Hello everyone,
> > >
> > > We are currently experiencing problems with the Leica SP8 confocal
> > system.
> > > If users are attempting to set up a 5 colour run, the setup takes a
> very
> > > long time (moving red progress bars when adding each colour).
> > >
> > > Observations:
> > >
> > >     - CPU at 60-80% when it's doing that
> > >     - RAM usage is low
> > >     - HDD activity is nigh on 0 (we swapped temp directory to the SSD,
> no
> > >     improvement (or minimal)
> > >
> > > I am wondering if it is the software struggling with the overlapping
> > > filtering?
> > > Does anyone else have these problems or has anyone found a solution?
> > >
> > > Thank you,
> > > Angela
> > >
> > --
> > ------------------------------------------------------------
> > Steffen Dietzel, PD Dr. rer. nat
> > Ludwig-Maximilians-Universität München
> > Biomedical Center (BMC)
> > Head of the Core Facility Bioimaging
> >
> > Großhaderner Straße 9
> > D-82152 Planegg-Martinsried
> > Germany
> >
> > http://www.bioimaging.bmc.med.uni-muenchen.de
> >
>
>
> --
> Tina Weatherby Carvalho
>
> University of Hawaii
> Pacific Biosciences Research Center
> Biological Electron Microscope Facility
> 2538 McCarthy Mall
> Honolulu, HI 96822
> 808-956-6251
>

--

*Dr. ANGELA KURZ*
Bioimaging Specialist | Research Officer

*Centenary Imaging | Sydney Cytometry*
*CENTENARY INSTITUTE*

*Mail* | Locked Bag 6 | Newtown | NSW | 2042 | Australia
*Courier* | Building 93 Royal Prince Alfred Hospital | Missenden Road |
Camperdown | NSW | 2050 | Australia

*E* [hidden email] <[hidden email]> | *T* +61 2 9565
61 <+61%202%209565%206173>880 |  *W *www.centenary.org.au

--
 <https://www.centenary.org.au>

[George McNamara]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

Hi Angela,

I strongly recommend avoiding using Leica's Dye Assistant (or anyone
else's). 5 fluorophores, 5 detectors ... if you have HyD(s) with photon
counting, these are likely superior to Leica's PMTs. I suggest starting
with HyD(s), 10 line accumulation (no frame accumulation). Very
important: Leica TCS mode on power up defaults to 400 Hz scan speed. 600
Hz is much better: (i) still enables use 0f 0.75x zoom (biggest field of
view), (ii) 1.5x faster than 400 Hz, (iii) less photobleaching
(fluorophore destruction is a >1photon process ... easiest way to test
if your user uses 400 Hz: try 40 Hz, no line accumulation vs 400 Hz, 10
line accumulation). Superiority of SP8 2nd gen HyD vs Leica PMT at any
gain (and offset setting, also typically will need to average, not
accumulate): usually pretty obvious (and if any PMT is superior, check
each HyD with same laser - fluorophore - emission bandpass, to find out
if any need replacing). If modest field of view (ex. zoom 1.28 or 2.25
... though higher zoom means more stitching), could go faster than 600 Hz.

20 Gb is going to overwhelm the PC (probably an HP Z model). Some short
term things to mitigate:

* don't do 20 Gb acquisitions! Even 10 Gb would be better.

* more ram. Possibly exit LAS X, exit Windows, restart PC after every run.

* if you don't have it, get extensive demo of Navigator ... and see if
the stitching step can be done on good offline PC (and negotiate good
price for the Navigator license).

* put in a 10 Gbe (maybe even 40 Gbe?) Ethernet card ... on both the
Leica PC and nearby destination PC (our core has 10 Gbe server, 10 GBe
switch, CAT7 cables). The Leica scanhead data is through a (probably) 1
Gbe Ethernet card (dual card, and if you have SMD, other port is for
that), and static IP address [can be found in several places in Windows,
and the Ethernet card typically has the static IP address printed) ...
could probably be replaced with a dual 10 Gbe Ethernet card, one for SP8
scanhead (good explanations on internet how to set the static IP address
for the new card -- or work with your I.T. folks ... also needed even if
you just move their card over a slot), one for data (because next item
suggests an additional PCIe card).

* I have in our SP8 an ASUS Hyper M2 card, with 1 TB M.2 NVMe SSDs
(configured for speed) ... current card is PCIe gen 4,

https://www.amazon.com/ASUS-M-2-X16-Expansion-Card/dp/B084HMHGSP

... can be moved over to a new motherboard PC when your budget enables
(tip: more CPU threads --> more PCIe lanes --> more overall data
throughput).

* Leica's HP PC's often ship with PCIe card(s) that are not used by the
SP8, so can be taken out (good to check with your local Leica folks), to
put in better hardware (better GPU for Huygens in our case).

George

p.s. occasionally our SP8 gives better data with frame accumulation
because line accumulation sometimes 'glitches' (not obvious why ...
intermittent, maybe intensity related). Our users almost entirely use
fixed specimens, so I could recommend frame accumulation ... simpler to
show the user's who need to avoid lines what line vs frame assumulation
means.

2nd gen HyDs ... pretty good ... the newer (~2 years ago) SMD HyDs
better, the new Stellaris confocal (SP9? SP11? ... I was hoping for SP
XP) POWER "HyDs" look good numbers wise (still not seen any QE curves)
on Leica's marketing online.

Any version of HyD in photon counting mode (with some modest number of
line or frame accumulation, and 600 Hz or faster scan speed) - in my
opinion - is better than Leica's PMTs.

More on speed test: Jonathan Boyd, at the time Leica applications (later
became a salesperson, then moved to industry) showed me:

a) 80 Hz * 1 accumulation

b) 800 Hz * 10 line accumulation

c) 8000 Hz * 100 line accumulation (resonant scan mode).

"c" was by far superior.

The thanks to Jonathan for doing this. Faster scanning with appropriate
photon counting accumulation is the way to go.

On 5/26/2020 9:01 PM, Angela Kurz wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> Hi,
>
> Thank you everyone for your feedback.
>
> The user, who is having problems, uses Dye Assistent, frame sequential and
> uses all our 5 detectors for the 5 colours. In addition he is setting up a
> huge tilescan (20GB).
>
> We will try to set it up manually.
> I was also in contact with Leica and they will set up our workstation from
> scratch.
>
> Let's see how it goes.
> It wasn't like that before.
>
> Thanks again everyone!
> Angela
>
>
>
> On Wed, 27 May 2020 at 08:35, Tina Carvalho <[hidden email]> wrote:
>
>> *****
>> To join, leave or search the confocal microscopy listserv, go to:
>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>> Post images on http://www.imgur.com and include the link in your posting.
>> *****
>>
>> Hi,
>>
>> Yeah, what Steffen said. Set up each sequence, one at a time, test it well
>> before moving on to the next sequence. I often avoid Dye Assistant when I'm
>> doing this. Do frame sequence. I've only ever tried using a different
>> detector for each, but I bet you could double up.
>>
>> And since they are stuck at home, the helpful Leica applications people are
>> waiting for your call!
>>
>> Aloha,
>> Tina
>>
>> On Tue, May 26, 2020 at 6:23 AM Steffen Dietzel <[hidden email]>
>> wrote:
>>
>>> *****
>>> To join, leave or search the confocal microscopy listserv, go to:
>>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>>> Post images on http://www.imgur.com and include the link in your
>> posting.
>>> *****
>>>
>>> Hi Angela,
>>>
>>> nice to read from you!
>>>
>>> Can you go into more detail how you are doing this? Probably sequential
>>> mode, but line or frame sequential? How many different detectors for the
>>> five colors are used?
>>>
>>> If possible, it seems fastest to set up all channels with their
>>> wavelength-range first and only then activate sequential scanning. So
>>> that the mirrors do not need any movement between channels. If you need
>>> different wavelength-range-settings (in frame sequential, if you need to
>>> reuse a detector), try to keep the moving of mirrors to a minimum, i.e.
>>> if you have say two wavelength settings and each of them is needed
>>> twice, make sure you have the same setting-sequences after each other,
>>> not intermingled.
>>>
>>> And, finally, try to set up the channels from scratch. I have the
>>> impression that on rare occasions, the system is introducing some
>>> unnecessary extra loops. And once it starts with that it seems to be
>>> easier to start over, instead of trying to repair the broken setting.
>>>
>>> Good luck
>>>
>>> Steffen
>>>
>>> Am 25.05.2020 um 00:49 schrieb Angela Kurz:
>>>> *****
>>>> To join, leave or search the confocal microscopy listserv, go to:
>>>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
>>>> Post images on http://www.imgur.com and include the link in your
>>> posting.
>>>> *****
>>>>
>>>> Hello everyone,
>>>>
>>>> We are currently experiencing problems with the Leica SP8 confocal
>>> system.
>>>> If users are attempting to set up a 5 colour run, the setup takes a
>> very
>>>> long time (moving red progress bars when adding each colour).
>>>>
>>>> Observations:
>>>>
>>>>      - CPU at 60-80% when it's doing that
>>>>      - RAM usage is low
>>>>      - HDD activity is nigh on 0 (we swapped temp directory to the SSD,
>> no
>>>>      improvement (or minimal)
>>>>
>>>> I am wondering if it is the software struggling with the overlapping
>>>> filtering?
>>>> Does anyone else have these problems or has anyone found a solution?
>>>>
>>>> Thank you,
>>>> Angela
>>>>
>>> --
>>> ------------------------------------------------------------
>>> Steffen Dietzel, PD Dr. rer. nat
>>> Ludwig-Maximilians-Universität München
>>> Biomedical Center (BMC)
>>> Head of the Core Facility Bioimaging
>>>
>>> Großhaderner Straße 9
>>> D-82152 Planegg-Martinsried
>>> Germany
>>>
>>> http://www.bioimaging.bmc.med.uni-muenchen.de
>>>
>>
>> --
>> Tina Weatherby Carvalho
>>
>> University of Hawaii
>> Pacific Biosciences Research Center
>> Biological Electron Microscope Facility
>> 2538 McCarthy Mall
>> Honolulu, HI 96822
>> 808-956-6251
>>
>