Been using the OT-2 for couple of years now, and I have two very annoying issues.
Analyzing protocols where 384-well plates are used take 10+ minutes, whereas all other protocols with 96 plates are instantly ready to start. When looking at the developer tools in the opentrons app there is a constant 404 error with what seems to be the robot’s internal server
“Failed to load resource: the server responded with a status of 404 (Not Found) [robot’s IP:31950]”. I get these hundreds of these repeatedly until it “magically” finishes analysis and the Start run button appears.
When running protocols with 384-well plates only (protocols with 96-well plates run without issues), the pipettor introduces random delays before executing actions. These delays occur regardless of the specific action being performed (tips pickup, movement, aspiration, dispense). For certain steps, the delay is consistent and reproducible. For some transfers there is always a delay immediately after the pipettor picks up tips. The robot pauses at the tip rack location for ~1 minute or longer before proceeding with the transfer.
In other instances, the arm stalls over the reservoir during bulk transfers when moving to aspirate liquid from the next well. This is also independent of connection type.
Unfortunately after providing all possible logs to OT support they could not figure out the issue. This is all true if the robot is connected via USB or WiFi, and I keep the software updated constantly.
Anyone ever experienced these issues before? And if so, did you find a solution?
I remember there being a correlation between the time required to analyze a protocol and the protocol’s complexity and the complexity of the associated labware. I thought that issue had been resolved by the many updates to the protocol analysis they introduced. We noticed protocols with 384 plates would take much longer than 96, and we thought it was related to how the OT-2 processes the labware data.
What is the layout of your labware? How many 384-well plates do you have?
I have 2 384 plates. I didn’t check the firewall, but since protocls with 96 plates takes couple seconds I don’t thinks that’s the issue.
The fact I get these 404 and 403 errors in the console for 10+ minutes and suddenly it finishes make me think the internal server is faulty. Btw, when starting a run from last runs (under the device section) it starts instantly.
I didn’t try to reset but I did reflash the SD card. I’m not inclined to reset all my setting honestly, rather wait than setup all the calibration and hoping nothing breaks
I don’t have an ot2 to test it on at the moment, but I’d be curious to see if you still experience this if you load a 96 well plate and pipette as if it were a 384 well plate with offsets to see if you would have the same issue (or just 4x the pipette steps and don’t load a plate)
I always chalked it up to the amount of actual steps the robot takes which dictates how long a protocol takes.
Interesting you see the error messages, maybe you could provide the logs here? I doubt I’d solve it but it’s interesting to see
I do know the ot app saves protocols that are (to it) identical. If you load multiple protocols with 96 well plates that are the same it should load instantly, but maybe you’re changing the ones with 384? Just a random thought
If you reflashed the SD card, that is the hardest type of reset you can perform.
Remembering things a bit better, and I think the 404 errors are just something the OT-2 does and might be a red herring for this specific issue. Happy to review the logs, but Opentrons support team might be more up to date with what to look for.
The OT-2 would use multiple layers of dead reckoning to determine where a well is on a plate. I would assume that having to go through all the calculations for each well on the 384 plate might slow it down. I would think other people would be experiencing this issue if that were the case.
I would agree with your theory if I had none of these issues when I ran 8 96-well plates in a single protocol.
Btw, I just recalled that there were a few times when analyzing was quick and it was always after updating the firmware or the app.