Update: It seems that the camera flex module connectivity is documented in the OmniVision OV2640 Camera Module Hardware Application Note which is available online. I missed it at first because the AI Thinker PCB schematic uses a 24-pin FPC connector symbol whose numbering is reversed compared to the OmniVision App Note. If you reverse the pin numbering for pins 1 to 24, , the connectivity make sense. Now I just need to find a similar App note for the OV5640 module.
Another thing to consider is that the OV5640 only allows a maximum of 3.0 volts on its digital I/O V(DD-IO) rail whereas the OV2640 allows 3.3 volts. So is possible that the OV5640 module is using the ESP32-CAM's +3.3V rail (flex connector pin #14) for V(DD-IO) and thereby overdriving the sensor chip. If on the other hand it used the +2.8V rail (flex connector pin #21) for the sensor's V(DD-IO), the sensor's logic HIGH output might be too low for the ESP32 to reliably recognize. We would need to examine the OV5640 module's flex schematic to know for sure. I have searched but have been unable to find one online. Do you ever seen one? I may have to unsolder the sensor from the flex to figure out its connections. Once I know how the module is actually wired, a proper fix might not be all that difficult. This is really an unfortunate situation because running the sensor hot will degrade its noise performance and give a bad picture, especially in low light.
files.waveshare.com/upload/b/bb/OV5640-A.pdf This is all I have. As far as I know, both OV5640 and OV2640 have the same flex connector. Isn't the root cause of the heat generation due to the use of the OV5640's internal regulator? So if we input a correct voltage from an external circuit to V(DD-IO) the OV5640's internal regulator is not used. This is how I'm understanding. Do you think it would be possible to modify the pin to use 2.8V?
Yes, the solution would be to replace the 1.2V regulator on the AI-Thinker board with a 1.5V regulator, Torex XC6206P152MR. That is the SOT-23 package which I think is correct. That should disable the OC5640's internal regulator.
If I get some time I will try to drive DVDD from a bench supply to see if the camera module dissipation is reduced. You could do the same if you have a small soldering iron and some fine wire. Just tack a wire onto the 1.2V bypass cap. You could also use an external linear regulator module like an LM317. It is not a permanent solution but it 0:02 would work well enough to measure the temperature rise.
@@ThatProject I tried using an external LM317 to feed 1.5V to DVDD and it made only a small improvement to the temperature rise, maybe 5 degrees after 3 minutes. The external regulator should reduce the chip dissipation from 0.55W to 0.32W which is a substantial difference. However I was measuring the temperature of the back of the sensor with a platinum RTD and no heatsink, so it is possible that with no heatsink there is so much thermal resistance to ambient that even the 0.32W is enough to drive the temperature that high, albeit more slowly than with 0.55W. I would need to monitor the current of the DOVDD rail to make sure the internal regulator is indeed not supplying current to the core, but that is not so easy to do. Either way, it looks like the OV640 cameras need a substantial heatsink regardless of their powering scheme.
One other thing. I noticed a strange dot/dash pattern appear in part of the image as the OV5640 got hot. When I sprayed freeze spray on it, the problem went away immediately. So the problem is likely due to internal timing errors in the on-chip image processor when it is run beyond its maximum temperature rating. So for good picture quality, the heat dissipation of this camera sensor has to be managed carefully. I think this would be the case even if the chip was run at lower DOVDD voltage.
I'm still looking for a good camera module to use in an MCU environment, but unfortunately there aren't many options. Have you checked out the AMB82-Mini board? It has great performance, especially when it comes to capturing images continuously. ua-cam.com/video/TGqOUVhQQY8/v-deo.htmlsi=UlxwIbJVnBFB-Op9
@@ThatProject I love the AMB82 Mini. Far worth the little extra on price. Outperforms any other those others and better at recognition. Also easy to setup. No monkeying and moving files around like some others. Very professional all the way around to me.
I'm amazed more and more with the AMB82Mini. The compile time of all their examples is a lot faster even. That makes me think the software is written better.
Hi, i tried ESP32 cam with OV5640 autofocus version and it has same overheating problem. Is there any solution? Something similar to ESP32cam etc.? Thank you for your work!
@@VitaZVitkova There is no good solution for this yet on the ESP32CAM AI-Thinker board. Additional circuit configuration appears to be necessary to accurately input the core voltage.
![LVGL 9, [Ep. 01], From Scratch to Benchmark, For newbie/beginner/starter #LVGL #UI #ESP32 #Benchmark](/img/n.gif)
Update: It seems that the camera flex module connectivity is documented in the OmniVision OV2640 Camera Module Hardware Application Note which is available online. I missed it at first because the AI Thinker PCB schematic uses a 24-pin FPC connector symbol whose numbering is reversed compared to the OmniVision App Note. If you reverse the pin numbering for pins 1 to 24, , the connectivity make sense. Now I just need to find a similar App note for the OV5640 module.
Another thing to consider is that the OV5640 only allows a maximum of 3.0 volts on its digital I/O V(DD-IO) rail whereas the OV2640 allows 3.3 volts. So is possible that the OV5640 module is using the ESP32-CAM's +3.3V rail (flex connector pin #14) for V(DD-IO) and thereby overdriving the sensor chip. If on the other hand it used the +2.8V rail (flex connector pin #21) for the sensor's V(DD-IO), the sensor's logic HIGH output might be too low for the ESP32 to reliably recognize. We would need to examine the OV5640 module's flex schematic to know for sure. I have searched but have been unable to find one online. Do you ever seen one? I may have to unsolder the sensor from the flex to figure out its connections. Once I know how the module is actually wired, a proper fix might not be all that difficult.
This is really an unfortunate situation because running the sensor hot will degrade its noise performance and give a bad picture, especially in low light.
files.waveshare.com/upload/b/bb/OV5640-A.pdf This is all I have.
As far as I know, both OV5640 and OV2640 have the same flex connector. Isn't the root cause of the heat generation due to the use of the OV5640's internal regulator? So if we input a correct voltage from an external circuit to V(DD-IO) the OV5640's internal regulator is not used. This is how I'm understanding. Do you think it would be possible to modify the pin to use 2.8V?
Yes, the solution would be to replace the 1.2V regulator on the AI-Thinker board with a 1.5V regulator, Torex XC6206P152MR. That is the SOT-23 package which I think is correct. That should disable the OC5640's internal regulator.
That's super interesting. Any plan to replace it?
If I get some time I will try to drive DVDD from a bench supply to see if the camera module dissipation is reduced. You could do the same if you have a small soldering iron and some fine wire. Just tack a wire onto the 1.2V bypass cap. You could also use an external linear regulator module like an LM317. It is not a permanent solution but it 0:02 would work well enough to measure the temperature rise.
@@davidclementson6149 I think this is a very interesting challenge. I will try it too when I get the chance. Thank you for sharing.
@@ThatProject I tried using an external LM317 to feed 1.5V to DVDD and it made only a small improvement to the temperature rise, maybe 5 degrees after 3 minutes. The external regulator should reduce the chip dissipation from 0.55W to 0.32W which is a substantial difference. However I was measuring the temperature of the back of the sensor with a platinum RTD and no heatsink, so it is possible that with no heatsink there is so much thermal resistance to ambient that even the 0.32W is enough to drive the temperature that high, albeit more slowly than with 0.55W. I would need to monitor the current of the DOVDD rail to make sure the internal regulator is indeed not supplying current to the core, but that is not so easy to do. Either way, it looks like the OV640 cameras need a substantial heatsink regardless of their powering scheme.
One other thing. I noticed a strange dot/dash pattern appear in part of the image as the OV5640 got hot. When I sprayed freeze spray on it, the problem went away immediately. So the problem is likely due to internal timing errors in the on-chip image processor when it is run beyond its maximum temperature rating. So for good picture quality, the heat dissipation of this camera sensor has to be managed carefully. I think this would be the case even if the chip was run at lower DOVDD voltage.
Could there be another manufacturer making the OV5640/ESP32 with a better design (different regulator, heat sink, etc)?
Right. It may be slightly different depending on the manufacturer, so we may need to check the schematic.
I think you fell in love with lady ESP32😁
Do you have the same problem when using the OV5640 in the Seeeduino XIAO ESP32S3 Sense Board?
@@esqueda.business I don't know exactly yet because I haven't tried it yet. I'll take a look when I get a chance.
thanks for the information👍
Very informative!
Glad you liked it
do you know if the XIAO ESP32S3 Sense can use the OV5640? And if so, is there an overheating issue?
Good point. I think I need to check its schematic first. However, it is possible that it still has the issues I mentioned.
So, what is a good alternative to either of these camera modules? At least for continuous capture.
I'm still looking for a good camera module to use in an MCU environment, but unfortunately there aren't many options. Have you checked out the AMB82-Mini board? It has great performance, especially when it comes to capturing images continuously. ua-cam.com/video/TGqOUVhQQY8/v-deo.htmlsi=UlxwIbJVnBFB-Op9
@@ThatProject Actually just ordered one. I will let you know how it goes. The reviews are great.
@@Michael-lo3ht Ok. Please let me know the test result.
@@ThatProject I love the AMB82 Mini. Far worth the little extra on price. Outperforms any other those others and better at recognition. Also easy to setup. No monkeying and moving files around like some others. Very professional all the way around to me.
I'm amazed more and more with the AMB82Mini. The compile time of all their examples is a lot faster even. That makes me think the software is written better.
Hi! Did you see this issue with the Auto-Focus version of the camera?
Good point. I'll have to check on that again.
Thank you please let me know!@@ThatProject
Hi, i tried ESP32 cam with OV5640 autofocus version and it has same overheating problem. Is there any solution? Something similar to ESP32cam etc.? Thank you for your work!
@@VitaZVitkova There is no good solution for this yet on the ESP32CAM AI-Thinker board. Additional circuit configuration appears to be necessary to accurately input the core voltage.
I see, where could I find the right core voltage info to design my own support circuit?@@ThatProject
In your view does there exists as esp32 cam which can be used for streaming , which is dead stable and does not overheat !
@@AmanSinghal-ny3ik This overheating issue occurs when using OV5640 on ESP32CAM board. Normally, this type of heat does not occur when using OV2640.
❤