The OD measurements increase significantly at random time on the Pioreactors.
Software version: 25.5.22
Pioreactor version: Pioreactor 20 mL v1.1
Despite having an older software version, we did not have this issue in majority of our experiments. However, recently carrier two experiments resulted with this problem in different a total of 6 reactors. We checked the IR excel log, but the intensity did not show any difference before and after the high sudden jump. In the picture below, I am showing only 3 of them.
So I don’t think it’s related to ir_led_intensity-auto. Are you sure it’s not biological? The blue graph looks suspicious, but the other two graphs less so.
Are you using OD calibration curves? If so, do those have a odd shape/behaviour near OD=0.05?
Any sudden changes in temperature in the Pioreactors?
We were suspicious as well, though we observed this issue with different organisms, and the same organisms tested in previous expriments did not had this issue.
No, we used 20 mL vials with cross shaped stirring magnet. The media we used was also similar to water in transparency which got turbid over time with bacterial growth.
If you observer black square at the end of Plot 1 - You can see another jump in OD readings. This was triggered after pressing the REF PD inwards and replaceing some of the caps holding the REF PD. Though, the OD jumps occured prior to the latest event occured randomly in the middle of the night, but maybe it might be caused due to the reasons described below.
Can we make this system more robust by reducing the ir_led_intensity-fixed number?
So the highlighted jumps are expected. If you move the REF, it changes the normalization that occurs in the OD readings:
raw OD reading = [90deg signal] / [REF signal]
So any disturbance of the REF signal will shift the OD reading.
The other jumps appear too unnatural to be biological. I wonder - are the REF photodiodes under some tension, and shifting suddenly to a new position at some point throughout the experiment?
(Internally, we’ve used that blue sticky tac to hold a PD and LED cap in place - a better solution to PDs is coming in the future)
However, this would show up a big jump in the dataset ir_led_intensities. You can confirm you see a jump in that dataset at the 24h mark (the highlighted jumps)
Other ideas / information:
The IR LED intensity is static throughout the experiment. Even if it changed, then it would affect [90deg signal] and [REF signal] about the same, hence the ratio [90deg signal] / [REF signal] wouldn’t change (that’s the idea!). So I don’t think ir_led_intensity is the problem.
If the problem isn’t [REF signal], what might cause [90deg signal] to increase? Do any jobs start or change at the same times? Like does stirring’s RPM increase or decrease, and other LEDs come on, etc.?
Are these near a window, and lots of ambient light suddenly hits the Pioreactor?
Hmm… this might describe the jumps, though it is wierd that where we observerd this was ~10 hour after in Experiment 1, but not in Experiment 2 that has been running for 800 hours (experiments were run on separate clusters). Interestingly thoug, after we replaced vials on the 800 hour experiment, the jump occured couple hours in to the experiment - This was not a jump caused by vial change, because the intial readings were fine. The jump happened when long later when we were not there.
Yes, this was the first thing I checked but there was no big jump registered at the moment of these events.
I would also said “other jobs” as potential cause, and yes, in certain cases I observed that there was a yellow warning for RPM=0 and slight increase in temperature (likely due to RPM=0) close to the events. Though, this combination has not been consistent in all the cases.
They are located away from the window and shielded at the corner, so ambient light should not be problem.
