Products > Test Equipment

New bench scope - Fnirsi 1014D, 7", 1GSa/s

<< < (41/67) > >>

donwulff:
On the other hand, there's a large body of research on FPGA reverse-engineering (Obviously, I'd say, just the synthesis software will usually visualize the net-lists to level where you *could* reverse-engineer it (due to gates usually being mapped to LUT's), but automated tools better) like https://github.com/emsec/hal - while that doesn't suffice for fully open source firmware (Does it matter if the hardware is "proprietary"?) it would help to get to speed, and perhaps lift some limitations. Reading gate level logic without comments or proper naming is bound to be annoying though, but some large scale architecture should be recognizable enough and perhaps extendable.

Oh and look if there could be self-update mechanism. Another thought is there are actually quite a few published oscilloscope FPGA designs, so perhaps this could be using one with some modifications? Obviously, if the firmware is fully re-written, it could also be adopted to some open source hardware scopes. Although the 1013/1014D price is hard to beat, and have any of the open source hardware oscilloscope projects actually got off the ground?

So now I'll just have to wade through all the threads looking for actual instructions on backing up & replacing the firmware (for the non-complete firmwares), and reading the FPGA etc. Not saying I can find the time to work on this....

pcprogrammer:
The link to the emsec/hal repository is interesting, but will also take some time to get familiar with, and you still need a way to convert the bit stream of your target FPGA into a net list. And there in lies a lot of work, I can attest to. Did it for the 1013D FPGA. (https://www.eevblog.com/forum/fpga/reverse-engineering-anlogic-al3_10-fpga/)

And even though I succeeded in reverting it back into readable verilog it failed to synthesize back into a working bit stream due to timing issues and different versions of the used IDE. The crude verilog did produce a somewhat working bit stream, but still not like the original. It did give insight in the inner working by using simulation. A huge advantage was the already present knowledge about the command set found while reverse engineering the firmware.

So fully reverse engineering a FPGA bit stream is not that simple, and even with a tool like that emsec/hal it will take a lot of knowledge to get things done.

Both tasks were new territory for me and took a lot of time, but even now with the gained knowledge, reverse engineering another bit of similar complex firmware or FPGA configuration would still take a lot of time.

Based on the result of the FPGA reverse engineering, I can state that there is no secret way to update the flash that holds the FPGA configuration bit stream, but I already knew that from reverse engineering the firmware of the 1013D. It does not have any provision in it to write to that flash. Since the 1014D firmware is not that different, and the update process uses the same code it will neither have that functionality.

That is why I won't bother with making something new for it. The new firmware is easy to install, without even writing over the original firmware, so one can easily revert back to it. Some even modified their scope with a switch to be able to go back and forth.

About the price of the scopes, I have not looked at them recently, but I paid about 30 euro more for my DSO2D10, so yes a bit more expensive, but you get a much better scope.

When I decide to do another scope project it will be for the Hantek and not the FNIRSI.  :)

donwulff:
FPGA update might not be done with the stock MCU firmware (software?), but it's odd to me to have upgradeable firmware without ability to upgrade the FPGA, especially as it's usually a standard block in the FPGA design software. Maybe they ran out of LUT space to put it in, I'll take your word for it, but yeah I'm not sure how many people buying 150 EUR scope or so have hardware & desire to flash it manually. Of course, the first FPGA design could probably put the flash-upgrade block in (Bearing in mind I don't really know the exact limitations here, although I've done some FPGA work in the past).

I can only imagine how much work it's taken to reverse-engineer the whole thing. It'll be lot easier for me to start with reading everything that's already been written about it on the threads, but don't know if I can put the time into it. Maybe I can get something working first though, also the idea that the same designs should be partially applicable to other/designed from scratch hardware is interesting, even if people don't get playing with that specific hardware.

Speaking of that, DSO2D10 doesn't seem to currently be available on AliExpress. From what I can see, DSO2C10 without the function-generator sells for around 300 EUR, so it seems also about double the price? Kinda nevermind as this thread is about the 1014D, but still it certainly makes lot more sense to buy better scope than try to one-off improve one! Then again, I personally wouldn't start to tear down even 300 EUR scope for the fun of it, so about 150 EUR is a sweet spot for me, with code & logic that should be applicable for higher performance ones. Still, I ALSO like to have it around for hobby projects & basic indicator. But we'll see, at least it's still available for purchase, which means a lot of people could get it. A shame the F1C200 versions aren't more common, and other hardware upgrades don't really look like an option.

pcprogrammer:

--- Quote from: donwulff on December 29, 2022, 11:12:42 am ---FPGA update might not be done with the stock MCU firmware (software?), but it's odd to me to have upgradeable firmware without ability to upgrade the FPGA, especially as it's usually a standard block in the FPGA design software.

--- End quote ---

The FNIRSI hardware and firmware is full of stupidity, so to me not such a big surprise that the FPGA configuration is not up gradable.

Have not seen something for it in the Anlogic IDE, but have not looked for it either.


--- Quote from: donwulff on December 29, 2022, 11:12:42 am ---Maybe they ran out of LUT space to put it in,

--- End quote ---

There is plenty of room left in the FPGA.

The way I could tell it not being in there is that the pins connected to the FPGA flash are not used in the bit stream.

Programming a serial peripheral interface into the FPGA between the flash and the MCU is not that big of a deal, but they might have had problems with getting their design to work in the first place that adding it made it fail.

They could also have used the MCU to configure the FPGA and have the bit stream sit in the flash the firmware resides in. It would then have been very easy to update it with a better design.

Fungus:

--- Quote from: pcprogrammer on December 29, 2022, 12:17:04 pm ---
--- Quote from: donwulff on December 29, 2022, 11:12:42 am ---FPGA update might not be done with the stock MCU firmware (software?), but it's odd to me to have upgradeable firmware without ability to upgrade the FPGA, especially as it's usually a standard block in the FPGA design software.

--- End quote ---
The FNIRSI hardware and firmware is full of stupidity, so to me not such a big surprise that the FPGA configuration is not up gradable.

--- End quote ---

I don't think their intention is ever to sell user upgradable devices.

The flash card is probably just a convenience for them.

(Maybe it's so they don't have to dump a load of hardware if they find a fatal bug?)

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod