Courious on getting the OpenMV modules with the OV2640 chip up and running I am here writing a forum post to ask whether or not it is an easy task to get the bootloader installed, get the current firmware matched with the board and get the camera sensor working.
At least I can confirm that the BGA-balling on the OpenMV boards that I have are done just fine.
The only parts missing are the BD33IA voltage regulator, by the way being a very expensive 3.3V regulator which in general doesn’t make sense to use, and the IR LEDs.
Matching lenses and houses have been sourced so I am just eager to get all these 16 boards up and running… Hmm, what kind of fancy vision based robotics projects to make.
Or, I could start selling off these “working” boards at $100 USD a pop
Actually I don’t get the marketing that OpenMV is doing. Leaving all their Kickstarter backers behind, selling off their actually working units at ridiculous expensive prices rather than suppliying the backers, and then degrading their camera module to be supplied in the board for their backers to only 0.3MP instead of the promised 2MP.
What a scam?
I’m honestly curious how you determined that the BGA chips are correctly soldered to the boards if you haven’t yet programmed them with a self-test routine. X-Ray inspection?
As far as I can tell, the bootloader for the OpenMV is the DFU routine that comes as a built-in part of the STM32F427 processor chip. At least I assume that is what it is, because that is how it works on MicroPython, and Ibrahim contributed heavily to developing the DFU tools in the MicroPython project.
As for the change in sensor and the other issues mentioned here- I’m writing as a Kickstarter backer who also sprung for a Firesale Board, not as one of the project members. And I have to say that given the disaster that they had to deal with after purchasing 1000 mishandled camera sensors, I’m impressed that they found a way to raise enough money to keep going. Some projects fold after a setback like that; it’s almost like people forget that in a Kickstarter you are actually backing a product that doesn’t yet fully exist, and that the risks are openly presented up front. Michael, Ibrahim, and Kwabena seem fully committed to seeing this through, to eventually get the cameras to all of their backers, and have found a get around the hurdles that they have faced.
(While I’m not a project member, I do design circuit boards as part of my job. So far I have avoided designing anything based on a BGA, though I won’t get away with that much longer. Reading about the OpenMV team’s experience does not make me eager to jump into that hornet’s nest…)
As for the sensor ‘downgrade’, I was initially disappointed as well, even though it seems that the 2MP resolution was never going to be utilized by the machine vision algorithms anyway. What is lost is the ability to take high-resolution snapshots; it appears that what is gained is a sensor that is faster and more light-sensitive at the resolutions used for the machine vision system. At this point I take it as a reasonable tradeoff to get the product working.
Or, I could start selling off these “working” boards at $100 USD a pop >
Actually I don’t get the marketing that OpenMV is doing. Leaving all their Kickstarter backers behind, selling off their actually working units at ridiculous expensive prices rather than suppliying the backers, and then degrading their camera module to be supplied in the board for their backers to only 0.3MP instead of the promised 2MP.
What a scam?
Or is this just me?
That’s kinda a cruel man. I haven’t made any money on this project. I’ve spent a lot of time/energy/money trying to make the company go and I could just say F’it and let the KickStart fail. I joined after the fact too - I didn’t take your money initially. I’m only still working on this because I believe in the project.
Seriously, I’ve stated more than once that we’re low on cash. That means we’re low on cash. Like < $20K left which we reserved for shipping. We only have enough let to do a new order of boards. Not to ship them. Without raising some funds no-one gets anything. The current cameras we have are coming out of a batch with a 50% failure rate. So, we build 2 cameras and only 1 works. Each camera board costs about $30 after paying for a $5 reballing cost. That means right off the bat we’ve lost $60 bucks to pay for materials. So, selling at $100 is pretty fair. Shipping costs for folks range from $6 and order domestic to $22 internationally. We’re looking at $12K+ at least to ship orders.
Sigh, anyway, the actual target price is $75 for these things normally, that’s what we calculated after adding in a 40% distributor overhead so we don’t have to sell them ourselves (yes, if you want a company like SparkFun to stock your product you will pay). If we were to sell them ourselves they’d cost less (about $50), but, that would require us to run an online store which is difficult if you only have limited time to spend on the project.
Anyway, I understand your anger. But, please try to be civil, I tried to be upfront and honest with everyone. So, please do me the same.
Moving on… I’ll post later about how to get those things working. Too busy right now to get into the technical depth.
Those look like the very first revision (we’re at the 5th rev now), someone made a batch a while ago, did you happen to buy those on ebay ? Anyway, the design has changed a lot since then and we don’t really have time to support every homemade cam out there, so I’m sorry but you’re on your own here… I’d start by buying some regulators
If you can get the thing working fine, more power to you. We’re constrained to using only no lead solder which is not as good generally. But, it makes our stuff RoHS. Distributors won’t touch you unless you can verify this.
As for my previous post, we’re selling anything at all right now because MacroFab did indeed build quite a few working boards. We also ended up with 120 reballed sensors that cost of $600 to get and we decided to build out 120 boards with those reballed sensors - 50% of them worked, so, we’ll have another 60 boards available for sale in addition to the original 40 or so that worked. Of course, that also means that we lost 60 boards with a price of ~$33 each in raw materials. The $100 price is based on us trying to make enough to cover the cost of 2 lost boards from the KickStarter order and enough money to build another board.
Anyway, these 100 OV2640 boards are the only ones we’re ever going to make. If you don’t buy one you’re not going to get an OpenMV Cam with a 2MP sensor.
.