OOR prevalence in generation 4 Internal Processing Units - Alert: Red
From: James Anderson
To: Artemis Robotics Gen 4 PE Team
Hello team
Let's get the good news out of the way, we have managed to get all units past the 2000 hour mark! Outstanding considering just a few years ago we barely got our Gen 3s past 1500 hours. Miraculously IPU-012-C has just passed 2500 hours, and was able to join the other Gen 4s in the latest live fire testing on Barex-4!
Now for the alert item. Pushing that 2000 hour mark has revealed some new stress points in the design of the current IPU. The techs have begun to notice a peculiar development, we have begun to refer to as Onset optical rejection (OOR). In which the IPUs around 2000 hours of continual use of the main body sensors. The secondary sensors, installed from the supplier, began to show signs of degradation.
This isn’t to say it gets worse, but instead it appears the IPUs main chip begins to favor the main unit sensors solely. This is argued to be a software bug of the units, though is unfixable at the current state. The current recommendation is to have the QA team in PrePilot for Gen 5 look for early signs.
Unfortunately once OOR takes hold the IPUs main processor takes an increasing amount of stress to eventually burn out. We’re currently projecting a scrap rate of 70% at 2000 hours
Cost to replace all unit failures ahead of contracted window, $13.8 million
Projected cost for redesign, $234.6 million
Full cost breakdown in 00R-01F
Recommendations
- Identify root cause of OOR and how to identify defects early to repair or discard IPUs at first sign.
- QA team in PrePilot for Gen 5 to start tracking optical data.
- Investigate further into additional OOR like defects that target other components.
We recommend this be marked down to a yellow alert considering the recent changes in IPU calibration that should bring us within expected ranges.
James Anderson
Snr. Project Engineer
Ext: 41-7896
Artemis Robotics - “Do what others can’t, Do what others won’t”
_____________________________________________________
OOR prevalence in generation 4 Internal Processing Units - Alert: Red
From: Elizabeth Moore
To: Artemis Robotics Gen 4 PE Team
Excellent work with the summary as always!
I’ll be heading to the Banard system soon where I can discuss with the supplier how they are working on decreasing calibration time further. I’ll make sure they are ok with all changes.
Do we have a list of all affected units of OOR? While the cost isn’t too high, I would like to see how OOR affects combat performance. We know that eventually they burn out, but this might be to our advantage if their performance increases. How long is the burn out period, and have the techs identified how to drag that out?
Elizabeth Moore
Director of Product Development
Ext: 25-3761
Artemis Robotics - “Do what others can’t, Do what others won’t”
_____________________________________________________
OOR prevalence in generation 4 Internal Processing Units - Alert: Red
From: James Anderson
To: Artemis Robotics Gen 4 PE Team
The techs wanted to mention that not all affected units have begun the process of burnout. It appears if the IPU doesn’t diagnose the issue it doesn’t fall into a feedback loop. Meaning that potentially once the issue is identified as long as the unit remains in use, it won’t degrade.
Current list of active affected units include (Gen 3s included)
012-C, 002-D, 012-D, 015-D, 021-D, 034-D, 056-D, 078-D, 103-D, 176-D, 224-D
Here is the data pertaining to all units, this includes the burned out IPUs not listed above
[Attachment scrubbed]
Upon looking at their statistics, aside from 012-C, the affected group were run through the calibration process a lot longer than the others. The remaining live units unaffected are either under 2000 hours, or an unidentified reason.
As for burnout period we still don’t have a good grasp on it, it seems highly personal to each IPU, it can last from days to months. The techs are writing down all relevant data including fuel intake and impact monitors in case this is a handling of the IPU issue.
As for the good news when it comes to performance. 176-D, while recently flagged as contaminated, has seen an increase of performance. There is a potential that this increase is either due to the contamination or the onset of OOR, it's still too early to tell.
James Anderson
Snr. Project Engineer
Ext: 41-7896
Artemis Robotics - “Do what others can’t, Do what others won’t”
_____________________________________________________
OOR prevalence in generation 4 Internal Processing Units - Alert: Red
From: Elizabeth Moore
To: Artemis Robotics Gen 4 PE Team
I just read through 176-D’s data and it appears to be failing to process Handler requests, is this going to be an ongoing issue? Do we need to wipe its local memory and start a new, or just scrap the unit?
Elizabeth Moore
Director of Product Development
Ext: 25-3761
Artemis Robotics - “Do what others can’t, Do what others won’t”
_____________________________________________________
OOR prevalence in generation 4 Internal Processing Units - Alert: Red
From: James Anderson
To: Artemis Robotics Gen 4 PE Team
It could be an ongoing issue, but at this point with over 2000 hours of runtime it would be a waste of resources not to run it till it burns out.
As for your second question, it wouldn’t be advised, an IPUs local memory is linked with its calibration, and we’ve yet to be able to figure out how to keep it, while discarding the rest.
James Anderson
Snr. Project Engineer
Ext: 41-7896
Artemis Robotics - “Do what others can’t, Do what others won’t”
_____________________________________________________
Fwd: OOR prevalence in generation 4 Internal Processing Units - Alert: Red
From: Jack Mann
[This message has been delivered outside of company networks, all attachments have been scrubbed]
[This message and all information held within are CONFIDENTIAL and protected under applicable law. If you are not the intended recipient, you are now notified that any disclosure, copying, or distribution is STRICTLY PROHIBITED. Please contact the sender of this message, and delete all copies, digital or hard. Artemis thanks you for your cooperation.]
