Pimoroni's OctoCam - Vertical lines from cammera

General project help for Adafruit customers

Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.
User avatar
Dad4_Service
 
Posts: 10
Joined: Sun Dec 19, 2021 7:11 pm

Pimoroni's OctoCam - Vertical lines from cammera

Post by Dad4_Service »

Hello,

I bought 3 OctoCam's and have set up one so far using motionEyeOS (https://learn.pimoroni.com/article/moti ... ur-octocam). At first I had issues accessing it via the network but finally got those fixed by restarting my router. After letting it run all night I came back this morning to find vertical streaks/lines in the feed. I checked the saved images and found the streaks/lines there too. I just enabled the second OctoCam and found it's working fine and not having the streaks ... I've also restored the settings to default from a backup made before any changes were done. Also the streaks aren't there when the camera first starts but gradually get worse over time.

Oh here's details on the os:
motionEye Version 0.42.1
Motion Version 4.3.1
OS Version motionEyeOS dev20201026
Here's a link to the photo's it's taking ... The second one is giving clear images and not producing streaks. Please help.
https://drive.google.com/drive/folders/ ... sp=sharing
I've attached one of the images as well.

Edit: I have reseated the camera. Here's the latest image:
This seems to be the worst of all the images.
This seems to be the worst of all the images.
Camera1_16-05-55.jpg (1020.37 KiB) Viewed 375 times
Attachments
Camera1_16-55-18.jpg
Camera1_16-55-18.jpg (449.44 KiB) Viewed 373 times

User avatar
adafruit_support_mike
 
Posts: 67446
Joined: Thu Feb 11, 2010 2:51 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by adafruit_support_mike »

That looks like noise overlaid on generally good sensor output.

Try buffing the Pi-side contacts on the camera's flex cable with a pencil eraser to clean away any oxide or environmental gunk that might be causing a noisy connection. See if that has any effect on the received image quality.

User avatar
Dad4_Service
 
Posts: 10
Joined: Sun Dec 19, 2021 7:11 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by Dad4_Service »

adafruit_support_mike wrote:That looks like noise overlaid on generally good sensor output.

Try buffing the Pi-side contacts on the camera's flex cable with a pencil eraser to clean away any oxide or environmental gunk that might be causing a noisy connection. See if that has any effect on the received image quality.
So in my original post I mentioned that the image degrades over time, I'm not sure if that makes a difference. I wouldn't expect it to degrade if it was a contact noise issue, but seeing as I'm quite noob I have no idea. Starts out quite clear but in short order (less than 1 hour) it is very poor quality. I attempted to clean the contacts with an eraser. The results have not changed. Here is a progression of the images(limited to 3). See all in the OP link to google drive, I've added many images and .mp4's to the drive folder(currently uploading).
Camera2_20211219_191853(immediately-after-restart-and-contact-cleaning).jpg
Camera2_20211219_191853(immediately-after-restart-and-contact-cleaning).jpg (85.28 KiB) Viewed 368 times
OctoCam2_20211219_194415(camera-name-changed).jpg
OctoCam2_20211219_194415(camera-name-changed).jpg (318.29 KiB) Viewed 368 times
OctoCam2_20211219_201003.jpg
OctoCam2_20211219_201003.jpg (325.21 KiB) Viewed 368 times

User avatar
Dad4_Service
 
Posts: 10
Joined: Sun Dec 19, 2021 7:11 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by Dad4_Service »

I'm going to test the camera on a one of the other RPi Zero W's from the same purchase and see if it's the pi or the camera. (using the same SD Card and camera on dif pi)
Okay just did that before posting this. Here are the 3 images (limited to that by your system)
OctoCam2_20211220_203924-just-before-switching-pi's-had-been-running-for-50-minutes.jpg
OctoCam2_20211220_203924-just-before-switching-pi's-had-been-running-for-50-minutes.jpg (193.85 KiB) Viewed 362 times
OctoCam2_20211220_210309-first-image-after-getting-it-positioned-right-with-alt-pi.jpg
OctoCam2_20211220_210309-first-image-after-getting-it-positioned-right-with-alt-pi.jpg (96.11 KiB) Viewed 362 times
OctoCam2_20211220_214849-almost-50-min-running-and-no-streaks.jpg
OctoCam2_20211220_214849-almost-50-min-running-and-no-streaks.jpg (95.68 KiB) Viewed 362 times
So I used everything the same except the pi. I suspect the pi is broken. Any further testing needed? I'm about to use other SD Card, Camera, and Power Cord (even different power adapter) with probable failing pi.

Edit: It would seem both octocam's are working w/out issue. I'll post again if the issue returns. Basically just switched with camera was attached to which pi! So much fun

User avatar
Dad4_Service
 
Posts: 10
Joined: Sun Dec 19, 2021 7:11 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by Dad4_Service »

So the issue has returned. With the same camera that it was happening on before. So it probably wasn't the pi. Is wiping the contacts with rubbing alcohol acceptable?

User avatar
adafruit_support_mike
 
Posts: 67446
Joined: Thu Feb 11, 2010 2:51 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by adafruit_support_mike »

Yeah, alcohol should be fine. It may not do much.. most electrical contaminants are either oxides or oils, neither of which are seriously bothered by alcohol. You might try some nail polish remover or acetone though.. that's a decent solvent for oils. Rubbing with an eraser is mild physical abrasion that works for thin layers of oxide, and rubber will absorb most kinds of oil.

The fact that the noise increases over time suggests something accumulating, and the most obvious suspect there is heat. Try pointing a small fan at the camera, and then at the RasPi. Does either condition have any effect on the noise?

User avatar
Dad4_Service
 
Posts: 10
Joined: Sun Dec 19, 2021 7:11 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by Dad4_Service »

Thanks for the suggestion. I did clean with acetone. Didn't help.

Attempted the heat idea. I don't have a small fan but my lungs did quite well :P. I gently blew on the camera but that didn't help. I then blew on the cpu of the pie and that helped A LOT! I could see the streaks clear when I blew on the CPU. Does that mean a defective CPU? I mean the other two Pi's aren't overheating just this one. They are all running motionEyeOS.

Also note: I did switch the SD card and camera again so now it's octocam3 in the file names but the cpu is the same as previous posts.

Here are 2 images ... first is with streaks and second is after blowing on the cpu for a little while.
OctoCam3_20211223_000319.jpg
OctoCam3_20211223_000319.jpg (370.85 KiB) Viewed 343 times
OctoCam3_20211223_000524.jpg
OctoCam3_20211223_000524.jpg (102.18 KiB) Viewed 343 times
It did not take 2 min to cool it enough it was just that the image wasn't adjusting for brightness so I had to move it around a bit to figure out how to force a brightness correction. It actually cleared up about 20-30 seconds (maybe even 10-15) after starting to blow.

Any further thoughts?

User avatar
adafruit_support_mike
 
Posts: 67446
Joined: Thu Feb 11, 2010 2:51 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by adafruit_support_mike »

It sounds like video channel is getting hot enough to generate noise, at least.

Try saving an image to the SD card as well as displaying it to the screen. Do you see the same lines in the stored image?

User avatar
Dad4_Service
 
Posts: 10
Joined: Sun Dec 19, 2021 7:11 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by Dad4_Service »

I'm a little confused by that last statement/question. Have you seen the images I've been posting? Those images are stored to the SD card and viewed on the screen. They are not screenshots they are what the OS is creating and offering as downloads but the same images are being displayed when I'm viewing it live as well so yes the streaks are all the time(after heating up). And as noted before the heat being produced by the CPU is the likely culprit as the cause, for when I blow on the CPU to cool it the streaks disappear both in the live and the stored image. I promise I'm not trying to be negative or non-constructive I just don't understand what you're trying to ask me to do because it sounds like you're trying to do what has always happened with this operating system. It's called motionEyeOS and is recommended for this device by Pimoroni. The OS offers a web user interface and displays a live stream of the frames the camera is catching. It offers the option of storing all frames some frames or videos.

At this point we've already determined what's causing the streaks. It is the heat coming off of the CPU which is very hot to the touch and when cooled removes the streaks or causes the streaks to no longer exist. Of the three octo cams that I purchased only this one Raspberry Pi has the overheating issue from the CPU.

Also, at this point, I'm wondering what we can do about this Raspberry Pi zero W that seems to be faulty?

User avatar
adafruit_support_mike
 
Posts: 67446
Joined: Thu Feb 11, 2010 2:51 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by adafruit_support_mike »

No worries.. you've confirmed what I needed to know: the noise lives in the signal path between the camera and the CPU, not in the path between the CPU and the display.
Dad4_Service wrote:Also, at this point, I'm wondering what we can do about this Raspberry Pi zero W that seems to be faulty?
'Faulty' is a broad enough term that I hesitate to use it here. The RasPi seems to work in general.

The noise seems to be an artifact of heat in the CPU/RAM, which is kind of inevitable on the physics side of things. Thermal noise is a major category of electrical noise.

I'd lean toward saying you've found a boundary condition of the Zero's operating characteristics. It seems to be able to handle the camera signals as long as the CPU temperature remains below a certain limit. Above that limit, the thermal noise in the camera's data path becomes too large to correct. That isn't too surprising, given that the Zero is an exercise in removing anything that isn't absolutely necessary.

It sounds like adding a heatsink or a fan should keep things under control.

User avatar
Dad4_Service
 
Posts: 10
Joined: Sun Dec 19, 2021 7:11 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by Dad4_Service »

adafruit_support_mike wrote:
Dad4_Service wrote:Also, at this point, I'm wondering what we can do about this Raspberry Pi zero W that seems to be faulty?
'Faulty' is a broad enough term that I hesitate to use it here. The RasPi seems to work in general.

The noise seems to be an artifact of heat in the CPU/RAM, which is kind of inevitable on the physics side of things. Thermal noise is a major category of electrical noise.

I'd lean toward saying you've found a boundary condition of the Zero's operating characteristics. It seems to be able to handle the camera signals as long as the CPU temperature remains below a certain limit. Above that limit, the thermal noise in the camera's data path becomes too large to correct. That isn't too surprising, given that the Zero is an exercise in removing anything that isn't absolutely necessary.

It sounds like adding a heatsink or a fan should keep things under control.
What I'll probably do is just have the camera not behind the CPU/ram. However, neither of the other pi's have this issue. Why would this one be getting so hot? Also, if I don't add a heat sink or fan (by just keeping the camera away) will the CPU/ram fail?

My main concern is the fact that this one pi has the overheating fault (and I use that term purposefully). Just because it "generally" functions, doesn't mean it's not faulty. All three use the same hardware and software and should function the same, right? Maybe, I'm wrong on that.

adafruit_support_mike wrote:I'd lean toward saying you've found a boundary condition of the Zero's operating characteristics.
I would agree with that if it wasn't for the fact that all three pi's aren't having the same issue. So it's not an operating characteristic of the zero generally but this one specifically.

User avatar
adafruit_support_mike
 
Posts: 67446
Joined: Thu Feb 11, 2010 2:51 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by adafruit_support_mike »

Dad4_Service wrote:My main concern is the fact that this one pi has the overheating fault (and I use that term purposefully). Just because it "generally" functions, doesn't mean it's not faulty. All three use the same hardware and software and should function the same, right? Maybe, I'm wrong on that.
You are, but it’s a matter of degree.

Every population has some amount of variation, and it’s wrong to assume every individual will be identical to all the others under any possible set of conditions. That’s why we have specs and tolerances: tolerances let us define conditions where we agree to ignore small differences between individuals while still being able to say, “that difference is unacceptably large”. Specs keep us from cherry-picking conditions to show that every member of a population is inferior to most of the others in some way or another.

That leads to the nuances of language: we can call any unwanted behavior ‘a fault’, but to declare something ‘faulty’ we have to show that it’s out of tolerance for some agreed-upon spec.

In this case we have the Zero, which is a notably minimal and fragile design, so the base amount of variation is fairly wide. To make things more fun, the RasPi Foundation changes hardware without changing the model name or version number, to the point where one board can boot from an SD card that another board of the same name/version can’t. The boards are running a third-party OS, and are operating at a temperature well away from the generally assumed standard of 25C. There are no published tolerances and specs in that range, so it’s impossible to say the board’s performance is out of spec.

To get physical evidence showing the presence or absence of defective hardware would destroy the board and cost several hundred times more than the board itself, so that option is out.

The next step is to go by known patterns in similar hardware. The effects of normal variation are deterministic: they respond to the input conditions in a predictable way and remain the same over time. Hardware defects tend to get worse over time, and to cause new failures as they go. In this case, the behavior you’ve mentioned sounds repeatable and consistent, so I lean toward calling it an effect of normal variation.

The next step is to estimate the amount of compromise. Accepting the appearance of noise in the video channel at high temperature as a fault condition, the RasPi behaves the same way as the others at lower temperature. As an untested proposition, I’d expect the board to perform the same way as the others running Raspbian at room temperature. I also don’t see anything to suggest the board would have a shorter working lifespan than the others while running Raapbian at room temperature. That puts the board as likely to be within tolerances for the specs that do exist, despite being noticeably worse than the others in a specific range of out-of-spec operating conditions.

I can’t turn that into a recommendation to chuck that board in the garbage and buy another, which is a starting requirement for authorizing a replacement.. we’d also have to confirm that the board is less than 30 days of purchase, and that it hasn’t been exposed to operating conditions outside our control that could have produced the fault condition.

User avatar
Dad4_Service
 
Posts: 10
Joined: Sun Dec 19, 2021 7:11 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by Dad4_Service »

That huge, long, convoluted response seems to be ignoring the issue. This Pi is faulty. The OS it's running is exactly the same as the other two (it's also much simpler than Raspbian).
adafruit_support_mike wrote:Hardware defects tend to get worse over time, and to cause new failures as they go. In this case, the behavior you’ve mentioned sounds repeatable and consistent, so I lean toward calling it an effect of normal variation.
The only reason it's not getting worse is because I'm not leaving it powered on for extended periods. Part of my previous post asked what will happen if I leave it on and let it run in the overheating state? I did recently see that the RPi hardware is supposed to automatically reduce its speed when it's overheating to prevent damage. Is that something the OS does or the hardware? Here's a quote from your product page
https://www.adafruit.com/product/3368 wrote:Note that you do not need a fan for your Pi, as it will automatically adjust its speed to avoid overheating, this is just for people who want to run it at top speed for a long time.
I'm not attempting to run it at "top speed" I'm not so well versed on the OS that I could even cause it to run at "top speed"
Dad4_Service wrote:What I'll probably do is just have the camera not behind the CPU/ram. However, neither of the other pi's have this issue. Why would this one be getting so hot? Also, if I don't add a heat sink or fan (by just keeping the camera away) will the CPU/ram fail?
I've not done that yet, and have added a heatsink to see if that helps (it doesn't).
adafruit_support_mike wrote:The boards are running a third-party OS, and are operating at a temperature well away from the generally assumed standard of 25C.
That statement is only partially correct. "The boards are running a third-party OS," is true. However, they are NOT "operating at a temperature well away from the generally assumed standard of 25C." Two of them are operating just fine and not hot at all to the touch. Only this faulty one is. I've not left this faulty one running for an extended period of time, but do suspect doing so would result in a dead processor in less than 24-48hours. I've only had it run for about 12 hours the first time it was set up.
adafruit_support_mike wrote:As an untested proposition, I’d expect the board to perform the same way as the others running Raspbian at room temperature.
I don't understand why you would expect that, especially considering the overheating. In any case you don't seem to be willing to replace this faulty board and I'm not sure what I can do to convince you. The board was less than 30 days old after purchase when this thread began (Ordered Dec 6, 2021 and thread started Dec 20, 2021). The CPU/RAM should not be getting this hot. It has not been "exposed to operating conditions" ... "that could have produced the fault condition." It was taken out of the box, put together, and turned on. The only operating conditions it's been in were ~70F, room temp.
adafruit_support_mike wrote:That puts the board as likely to be within tolerances for the specs that do exist, despite being noticeably worse than the others in a specific range of out-of-spec operating conditions.
I'm not sure how you've determined it to be "within tolerances" or that it's being used in an "out-of-spec operating condition." How is it in an "out-of-spec operating condition?" The OS? The OS is the one specified for the product I purchased.

Are you asking that I try putting raspbian on it and see how it runs? That is possible, I just don't see the point. The "third-party OS," as you say, is running on two others w/out any overheating issue. This is clearly a faulty RPi Zero W. I'm not sure why you are denying that.

In one of my previous posts I did say, "So it probably wasn't the pi." However, since then I've come to realize it's been the same Pi the entire time. That was Dec 23, 2021 and I had taken a trip on Dec 21 that resulted in mixing up the devices, lots of stuff was happening so close to Christmas.

User avatar
Dad4_Service
 
Posts: 10
Joined: Sun Dec 19, 2021 7:11 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by Dad4_Service »

adafruit_support_mike wrote: I can’t turn that into a recommendation to chuck that board in the garbage and buy another, which is a starting requirement for authorizing a replacement.. we’d also have to confirm that the board is less than 30 days of purchase, and that it hasn’t been exposed to operating conditions outside our control that could have produced the fault condition.
I didn't mention in my previous reply the broad reaching statement you made there at the end. Well, I almost did, but I didn't point out the fatal logic in it. Which I feel compelled to do so now. Your statement that you "have to confirm that the board...hasn't been exposed to operating conditions outside our control that could have produced the fault condition." basically is impossible to confirm for any product you sell. There is no way to be 100% certain that any product you have sold wasn't exposed to operating conditions outside your control. In fact it is 100% certain everything you sell to 3rd parties (such as myself) has been exposed to operating conditions outside of your control. In that you are not in control of the environment during shipping, during delivery, during unboxing, during assembly, during power up or any process after it's left your facility.

As I have not received a response, I have not tested the faulty RPi using raspbian. Let me know if you would like me to do that. Also, would you like me to leave the RPi running continuously to see when it shuts down due to heat failure? If so please tell me what OS to use during this test to be sure it had nothing to do with "out-of-spec" conditions.

This RPi by the way has had less than 24 hours of powered state. As I mentioned in my previous message, it was on for about 12 hours the first time and since then I've kept it on for no more than 1 hour to test the overheating issues.

Edit: I'm actually a big fan of adafruit, however this runaround I'm being put through for a faulty RPi zero w is kind of putting me off to the company. Is there a good reason this issue is going on so long? Did I fail to make it clear that the 3 pi's were being used exactly the same way with exactly the same os and only this one pi was failing with overheating issues? Your last post seemed to indicate you were under the impression that all 3 were getting hot but only this one was having the streaking issue. Well that isn't the case. Only this one is getting hot. The others can be touched and they are only slightly warm to the finger as if they were slightly above room temp of 21.11111 and probably running at say the 25 you suggested. Isn't that interesting :/ ?

Also, please note that I reported this issue well w/in the 30 days of purchase and as such the delays in your responses will not affect the fact that this issue started from day one of use and that I reported it as soon as I found the issue(12 hours after turning it on for the first time). I do not know if you are delaying responses to go past the 30 days (which we are now) or not but it will not be cool if that is/was the case.

P.S. If I'm not being positive or constructive in my questions or comments, I'm sorry. I have a hard time understanding what that would look like for a report of a faulty product. At least I'm not being vulgar though right :) ?

User avatar
Dad4_Service
 
Posts: 10
Joined: Sun Dec 19, 2021 7:11 pm

Re: Pimoroni's OctoCam - Vertical lines from cammera

Post by Dad4_Service »

It's been over a month since I asked what you'd like me to do. I set the Pi in question aside and waited.

It was not powered continuously since I received it (as stated in previous posts). I put it back in the box it came in and waited for your response.

Since you did not reply I did nothing. Well today I decided to attempt testing with raspbian w/out your reply/direction. I can confirm using motioneye via Raspbian Bullseye (Full) brought the temp upto 55'C resulting in the same failure of the camera's signal.

I will now leave the RPi on and running the camera for the next few days to see what happens. I'll cause the camera to not be directly behind the cpu/ram so it doesn't get damaged by the heat.

Locked
Please be positive and constructive with your questions and comments.

Return to “General Project help”