That's precious!

Just keep replacing parts until "works for me!"

If the devices were only bigger, they'd kick 'em!

Did anybody try a hammer?

Not sure what you mean in your second question...

This should clarify things

What I'm thinking is that the various ESP32-CAM boards should be compatible with the ESP32-CAM-MB hat. This would be helpful in breaking out those pins and, at a  minimum, allow me to recover something from the bad module.

...have you really proved it is defective?

Partly yes, but mostly no.

Yes, in that, out of the box, it reports an error during initialization that AFAICT is a PSRAM problem. When the package arrived, before I opened it, I got the CameraWebServer sketch to compile. There was some confusion about which configuration settings  to  use. There's no Aideepen board in Arduino; I could find nothing on the web; I sent an email to Aideepen about it and got no response. I finally decided to just use AI_Thinker and see what happens. If needs be, I'd go through all the settings. Well, needs be but I got the same results for all of them. I did search on the error code and I found Camera probe failed with error 0x105 · Issue #450 After reading the reports of the error my reaction was that there's a lot of "works for me" but no real diagnosis of  the problem. However, from my unpaid position, this is sufficient evidence of a defect.

Mostly no because, as a software engineer, I know this this is far from a definitive diagnosis of the problem. What's needed is to dig down into the espressif code and find out why the error is generated and try to replicate it; see if there a way to test the PSRAM or whatever is causing the problem and verify a fault. There's a nagging part of me that says the hardware is fine, the software is defective.

But...I'm still learning the ESP32 chip and haven't gotten the fundamentals yet. I want the low level code to identify the chip, how much memory is installed, etc. This data is available (shows up in  the boot sequence) but I haven't learned how to do it yet. And that's a lot of work that I'm planning to get to.

But not before I finish DroneBot Boot Camp 101 and not because I can't get the stupid $15 camera working.

Is there any possibility you had a static event?

What? What are you saying here? That this is somehow my fault? That I'm to blame? Well, that's not very nice! Now you've hurt my feelings. I may cry.

(I'm 67 yrs. old (you're 83, I believe, so you as well) and a static discharge event is not a high priority event for me. I'm more concerned with the sudden, unexpected, and explosive kind.)

But seriously, you're right, it may have been a static discharge. That's a possible cause. I just want to confirm the failure result. The two boards should be identical, and one reports working PSRAM. I need to find out how to address that RAM, verify I can address it on the working board and can't on the defect board. I don't know how to do that yet.

Are any of these settings incorrect?

Again, maybe yes and maybe no. But not from my hand.

In the CameraWebServer sketch the config settings are preset using compile time constants and selected using another constant to select the model, e.g., CAMERA_MODEL_AI_THINKER. However, I note that in one of versions of the sketch I found, the one from RandomNerds,  there's a check of the PSRAM performed while the configuration is being initialized.

void setup() {
    //...
    camera_config_t config;
    config.ledc_channel = LEDC_CHANNEL_0;
    config.ledc_timer = LEDC_TIMER_0;
    config.pin_d0 = Y2_GPIO_NUM;
    //...
    if (psramFound()) {
        config.frame_size = FRAMESIZE_UXGA;
        config.jpeg_quality = 10;
        config.fb_count = 2;
    }
    //...

So, I would guess that the configuration doesn't affect the PSRAM addressability. But, again, I haven't confirmed that PSRAM is the culprit. The error is a report of "resource not found" during a probe of the psram. That's where I'm at.

My next step is to try to program the working board and see where that takes me.

Of course, if there's any failure, I will blame others.

AS to 3rd question, some will say that is true, but say have you really proved it is defective?

Is there any possibility you had a static event?

Are any of these settings incorrect?

I have over a dozen MB boards and also just the carrier board to accommodate my non MB boards. I have at least 2 kinds, one I have to press the PGM/BOOT button to upload, the other I don;t (different USB/Serial chips)

That is the second 105 error I have seen in two days.

I have a pair of Aideepen ESP32-CAM-MB modules and I'm having trouble getting them to work. I'm using the CameraWebServer sketch.

One of them reports error E105 during setup

E (513) camera: Camera probe failed with error 0x105(ESP_ERR_NOT_FOUND)
Camera init failed with error 0x105

from CameraWebSerever.ino

void setup() {
    //...

    // camera init
    esp_err_t err = esp_camera_init(&config);
    if (err != ESP_OK) {
        Serial.printf("Camera init failed with error 0x%x", err);
        return;
    }

    //...
}

I looked into this without much success. What I found so far says this is a PSRAM issue, (which I'm not clear about) but not much else. Before I dig further  into that, I wanted to try the second module. But...

(First Question)

Do these modules somehow come preprogrammed?

The Aideepen video tells you to just turn it on and connect to server and go to the web page. Yes, I know that seems unlikely but since I had trouble programming the first one, I wanted to see if the second was functioning to some degree.

Sure enough, it is. Thankfully, you can't connect to it, but at least I know it runs,

ets Jul 29 2019 12:21:46

rst:0x1 (POWERON_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:2
load:0x3fff0030,len:4
load:0x3fff0034,len:6968
load:0x40078000,len:13072
ho 0 tail 12 room 4
load:0x40080400,len:3896
entry 0x40080688
[0;32mI (31) boot: ESP-IDF v4.1-dirty 2nd stage bootloader[0m
...

[0;32mI (1159) boot: Disabling RNG early entropy source...[0m

E (911) sdmmc_common: sdmmc_init_ocr: send_op_cond (1) returned 0x107
Card Mount Failed
PSRAM OK 
ESP32-CAM-MB

Hopefully I can program that one.

(Second Question)

The ESP32-CAM module should be replaceable with one from a different manufacturer so long as it has the headers. In other words, I should be able to use the hat with a different module, right?

(Third Question)

Is this common with electronics manufacturers today? They ship defective parts. The time I would spend returning the thing is more than it's worth.

I've now spent more time with this than the cost to buy dozens of them. Before I spend more time  trying to diagnose the issue, if someone has a answer, I'd be grateful.

  Ah .. sorry if I have misunderstood the origin of the picture that @ddaved showed ... was it from Bill's blog or similar?

In which case, Ron you are right ... we need to back track a step -- hopefully my effort will come in handy later on!

----------------

And unfortunately, there is a considerable inconvenience looming ...

Previously, the FTDI board was displayed as

and following one of my 'briefer' question-style sermons, @ddaved wrote:

Thanks for the help. The 3v/5v switch switches the VCC and TxD voltages from 3v to 5v.

Looking at the board there is only 1 power output pin labelled VCC

So it appears you either get 3.3V power and 3.3V logic levels, OR 5V power and 5V logic levels.

I do not see a way of getting a combination of 5V power and 3.3V logic.

------------------

Ron, you are copying a valid suggestion .. namely use 5V power connected to the 5V input pin on the ESP32-CAM board

BUT the ESP32 (and possibly the FTDI chip) is liable to fail if you feed it 5V logic levels

----------------------

The only solution I can see is to provide a separate 5V supply.

If @ddaved is prepared to do a little surgery to his FTDI card, it might be feasible to take the 5V supply directly from the USB input part of the FTDI card.

Or some other variation on this theme of accessing USB 5V power.

---------------------

Alternately, provide an independent 5V supply to the ESP32 card.

The grounds from the independent supply and from the FTDI card should meet at just one point, possibly at the ESP32 Cam ground input pin.

----------

Make sure the independent power feed cannot get back to the PC ... ie remove the VCC output wire from the FTDI card to the ESP32 board.

A few people have connected 'dubious' power sources to USB connected equipment (not ESP32-CAM), and the supply has for some reason delivered more than 5V to their PC USB socket, and made their PC uneconomic to repair. (Laptops are espeicially vulnerable .. the same 5V line is power to a number of expensive and unobtainable chips!)

The risk is small, but the effect on your wallet will be considerable if it happens!

---------------------

And whilst we are mentioning unexpected power problems, I saw one report of someone who had an ESP32 card with two pins labelled Ground .... unfortunately, only one was actually connected to the electronics on the card. Quality control at its best! Check everything twice!

----------------------- 

Best wishes,

Dave