|
Thanks! Yea Forseti is right Colorlight i5. Its a Lattice ECP5 (LFE5U-25F-6CABGA381), which itself is pretty slow. The RISCV struggles to meet timing at 55Mhz and uses over half the gate capacity of the chip. Gigabit Ethernet (125Mhz) barely meets timing. The DRAM being slow complements it all. If I could share the ethernet PHY with the soft core, I could put DHCP, TCP control, settings/sdcard in the soft core. Putting the UDP "goesouttas" into the FPGA is probably required. The IC7300 display runs at 60FPS, and this soft-core is really struggling to even do 5FPS. Hopefully I can avoid storing the whole framebuffer all together and skip the DRAM, just stream out ethernet. The colorlight i5 seems cool cause I can build a product around it without doing multi-layer impedance controlled PCB design. Just a carrier card with some buffers and connectors?
|
# ? Oct 14, 2021 01:47 |
|
|
# ? May 13, 2024 09:49 |
|
horse_ebookmarklet posted:Thanks! I think personally my favorite thing about it, aside from the form factor which I like a lot, is that you can use SymbiFlow because holy poo poo proprietary toolchains are a pain in the rear end! Is that what you used or did you use Lattice's stuff, whatever that is?
|
# ? Oct 14, 2021 02:03 |
|
Forseti posted:I think personally my favorite thing about it, aside from the form factor which I like a lot, is that you can use SymbiFlow because holy poo poo proprietary toolchains are a pain in the rear end! The open source toolchains are awesome. I've been spoiled as a software developer, tons of compilers/architectures/OS to choose from. The HDL world has crazy bad tools. I am using yosys + prjtrellis + nextpnr is pretty sweet. I am also using LiteX, which lets you define your HDL in Python and lets you dump a softcore in really easily. Haven't used SymbiFlow, how do you like it? Looks like it incorporates Yosys?
|
# ? Oct 14, 2021 02:36 |
|
Symbiflow is yosys + nextpnr/verilog2routing + trellis/x-ray/icestorm. Symbiflow is to those tools much like GNU is to the maintainers of the various parts of GCC. So you're already using symbiflow. That OS tools are available for fpgas now is probably the most exciting thing to happen to OSHW in the last decade. Words cannot describe how much I hate vivado and tcl.
|
# ? Oct 14, 2021 05:09 |
|
horse_ebookmarklet posted:Didn't even touch the lattice OEM software stuff. Having tried, and failed the xillinx and altera tools in the past, I didn't even bother. What CBJamo said, SymbiFlow united those projects I have the i5, but I haven't actually gotten around to messing with it yet. I think I just used one of the off the shelf LiteX configs to give it a test and haven't gotten back around to playing with it. That's awesome to hear it was so nice to work with, that's exactly what I had hoped! I have some Pano Logic G2s also, which feature a beastly Spartan 6 XC6SLX150, a 150K LUT part, in a nice box with the typical computer peripherals like network and video ports and such because they were meant to be thin clients. You have to use ISE WebPACK with them though and I spent a weekend getting that software set up once. I think I got it working, but by the end I was burned out and distracted myself with another project. I hope they reverse engineer the Spartan 6 sometime and get it in SymbiFlow.
|
# ? Oct 14, 2021 15:00 |
|
How feasible would it be to convert the display data to something e.g. a raspberry pi camera input could handle? If that were possible it seems like you'd save a lot of effort on the software side, perhaps even make it possible to use a smaller FPGA or even a CPLD. AFAIK those cameras use annoyingly complex interfaces so might not be as easy as it sounds...
|
# ? Oct 15, 2021 15:37 |
|
longview posted:How feasible would it be to convert the display data to something e.g. a raspberry pi camera input could handle? I've thought about this quite a bit, to my partners annoyance. I am too much of a dummy to do direct CSI2/MIPI. Baby's first FPGA project here. Like my 5th PCB layout. This FPGA can just squirt out low res HDMI, and there are ICs that convert HDMI to CSI2. (Used in this). This is actually pretty reasonable (other than the IC being BGA) and I probably could actually do this. But I can't find a distributor anywhere that stocks the IC. But the raspberry pi is too expensive to build a consumer product around, I think? Not sure if this makes sense but here I go: First option is I sell a tested-working setup including a rpi. Now I've purchased a raspberry pi (for basically retail $$), and I've added labor/shipping, so I have to put markup on it. Who wants to pay markup on a stock raspberry pi? Boo!!! (Looking at you Geochron 4K). Second option is people have to buy my doo-dad and then buy their own $50+ rpi and accessories. This keeps everyone's costs lower but introduces troubleshooting problems.... flashing SD cards, crappy SD cards, bad power supplies. I imagine more returns. I would probably go with both options, if I went that route. Either way, it adds $50 (or more, for #1) to the overall cost. I can either eat that in profit, or I can exceed my target price. I do enjoy diving down the ideas hole, this was fun to think about.
|
# ? Oct 15, 2021 18:40 |
|
horse_ebookmarklet posted:I've thought about this quite a bit, to my partners annoyance. It is quite popular to sell just a hat that someone can put on their RPi, and not uncommon for a fully kitted version to come at a premium price. Instead of worrying about the supply chain process of loading the SDCards just put some information on a website and a link to download the image.
|
# ? Oct 15, 2021 18:51 |
|
Those are fair points, especially if you're cost optimizing for retail as well. If I were doing this I'd probably try to find a FPGA+SoC combo like a small Zynq, but then you're definitely using proprietary tooling, and I know Xilinx generally has a pretty poor price policy for small volume buyers. Benefit would be that you'd get an ARM core with decent performance which could probably do some simple video compression (H.265, or perhaps even just MJPEG), usable DRAM, and a built in gigabit PHY, among others. Downside is those are all relatively high pin count BGAs, you can do a low cost board with one, but I doubt an inexperienced designer could pull it off for a low price without a few attempts. And that's ignoring how you'd solder the drat thing. One option for a Pi board would also be to go for the compute module 4 series, the cost of that is a fair bit lower than the standalone Pi, and it comes with built in eMMC, and built in gigabit PHY (just add transformers!). E: Not knowing the low level details of how hard this is, I think the ideal case would be to find a way to interface a CM4 with the display through something you can hand solder (e.g. a MAX V CPLD is available in QFP packages). Obviously you need something that works properly before that kind of optimization can be attempted. longview fucked around with this message at 19:07 on Oct 15, 2021 |
# ? Oct 15, 2021 19:02 |
|
longview posted:Those are fair points, especially if you're cost optimizing for retail as well. I've had success soldering some BGA packages at home with the squeegee paste and hot air method. Laser cut metal stencils are like $7, its crazy cheap. Placing the drat components with my shaky hands has been the hard part. Buying a pixel pump as soon as they hit the market. Fire up the toaster oven, we're gonna start breathing some lead! At work I've used both the Zynq and the Cyclone V (I do the software) and they're pretty nice. The pricing for low volume is pretty brutal ($50+ for the IC). We've got a Zynq Ultrascale+ in a design, now thats some big bucks. We are forced to use an ANCIENT version of OrCad, he was telling me he had to hand route like 2500 traces with most of them being differential. LOL Compression is gonna be a big thing for getting this working over the internet. The IC7300 is pushing like ~16 MB (128 Mbit) per second of display data (16 bit color, 488x288, 60FPS). So we're already talking needing gigabit just locally and WiFi is simply out of the question without dropping framerate. There is a lot of static content (or just pure black) in there that could get crunched nicely. Even run length encoding would be huge. The compute module 4 is a really good idea. It has a PCIe link that you could just hang a FPGA or dual port memory off of. Framebuffer into software is just a DMA away! Eliminate all of the networking, single target for your software, do h264 in software, HDMI, gige, all built in. drunk mutt posted:It is quite popular to sell just a hat that someone can put on their RPi, and not uncommon for a fully kitted version to come at a premium price. Instead of worrying about the supply chain process of loading the SDCards just put some information on a website and a link to download the image. There is also a reasonable approach for a hat-design. Use the SMI port to do something similar to the CM4 PCIe link, just less documented interface. The CaribouLite does this. Unfortunately both of these require layout skills wrt a FPGA subsystem and DDR memory. Outta my league. Doing breakout for a 300 something pin BGA plus routing DDR memory sounds impossible for me. What I've been telling myself: If I get this working If I get it to market If my partner and I dont go blind hand assembling these in the basement If this generates earnings enough to justify AND THE loving SEMICONDUCTOR SHORTAGE IS RESOLVED THEN: Version 2 will be probably a lot different and I can pay someone to do layout. But a lot of things to go wrong before there!!
|
# ? Oct 15, 2021 22:08 |
|
The icom RS-BA1 software has access to enough data to redraw the display over USB. Is sniffing that out not feasible?
|
# ? Oct 15, 2021 22:19 |
|
I would recommend against using a regular pi for anything you need to support. The raspberry pi foundation is willing to sacrifice all other concerns on the altar of cost. Given their mission that's a totally reasonable path, but you need to have the fact in mind when you design around it. If you just want to clone the display, I'd say all your module needs to do is convert to HDMI, let off the shelf parts do the rest. If you want to send it over the internet, then I'd try to get the data somewhere up stream before it's rendered to video. This might be worth the improved flexibility in any case. You mentioned a USB being available, is all the data you want somewhere in there? From the looks of your picture of it, you can probably hit those test pads with pogo pins and avoid soldering. On the topic of USB, you could use a USB-C connector for this without breaking spec. Vendor specific alt-modes are allowed. That would almost certainly not be worth the effort, but it is possible.
|
# ? Oct 15, 2021 22:44 |
|
eddiewalker posted:The icom RS-BA1 software has access to enough data to redraw the display over USB. Is sniffing that out not feasible? How is this different than RSBA1? I think the RSBA1 software simulates the display, and is limited to 44Khz of waterfall, whereas the IC7300 front panel can display 500Khz of waterfall. The USB Bridge IC they use is only capable of 1 Mbps. The USB audio codec is 44Khz@16 Bit (0.67Mbps). The display is ~128 Mbps, so there is simply not enough bandwidth over that USB to recreate the display. So they have to be doing the FFT of the data they got from the soundcard on your PC. The RS-BA1 crazy expensive for software alone at like $152 right? So that informs my pricing, it would be good to be cheaper than it or have better features. Or both.
|
# ? Oct 15, 2021 22:51 |
|
CBJamo posted:I would recommend against using a regular pi for anything you need to support. The raspberry pi foundation is willing to sacrifice all other concerns on the altar of cost. Given their mission that's a totally reasonable path, but you need to have the fact in mind when you design around it. It would be interesting to see what other's experiences are with designing around the rpi? Did you get burned? Or got a good article that talks about being burned? Yea I am not sure what my feature set should be. Just cloning it to HDMI would be a minimum. Over ethernet would be really nice too cause then you can use a pi or your desktop PC to see it. Connector chat: Ethernet now my best friend. You can buy double shielded CAT6 that has minimal boots on it. It can juuuuuust slip through the casting where the fan goes. I'll have to make a 3d printable bracket to stand off the fan but that should be easy. Will be interesting to see how much variation there is in the casting between units.
|
# ? Oct 15, 2021 23:21 |
|
Hi guys, glad to find an active ham radio thread on here. I just started studying for my technician exam and can reliably pass the practice exams so I'm excited to take the real test soon. I will probably end up picking up some DMR hardware to start as I live pretty close to both Brandmeister and DMR-MARC repeaters. Looking forward to being on the air soon
|
# ? Oct 16, 2021 20:53 |
|
Just renewed my license and boy howdy is CORES a byzantine maze.
|
# ? Oct 16, 2021 21:22 |
|
https://twitter.com/jonny290/status/1449524606293790723
|
# ? Oct 17, 2021 00:57 |
|
Get in the robot Shinji, also take this BTECH DMR-6X2 with 4000 channels and 3100mAh battery so we can hear you cry.
|
# ? Oct 18, 2021 11:46 |
|
Anyone know if there's a custom firmware out there for the Retekess V115? I searched but couldn't find one, figured y'all might know of one if it exists though. I ask because I found this video on how to trick it into tuning to frequencies you're normally unable to access: https://www.youtube.com/watch?v=E_fW_lb9O4U I tested it on mine and indeed it did work. I tested 65.8 MHz while using my NES to make an analog TV broadcast signal and was able to hear the audio no problem, so it seems it really will work with those out of range frequencies. If there isn't a custom firmware I'll at least have a look at the official firmware it and see if it might be as simple as changing the boundaries in FM mode or something, maybe even select ranges beyond the ones the AM input allows.
|
# ? Oct 20, 2021 02:52 |
|
I did realize one thing that's really nice and can be done with the existing exploit. I was able to tune into my local weather band station at 162.40. I'm sure that it's using the full wide band broadcast FM channel bandwidth, I have to turn the volume up all the way to hear it. I'm lucky that there are no other signals nearby here though so no other channels were trying to break through. So not ideal, but it's great to have it at all and personally I'm very excited because that makes this a much more reasonable thing to pack on a backpacking adventure instead of the much smaller AM/FM/WB radio I also have. It's still Regarding the hacking, I couldn't find any firmware images out in the wild or mentions of how one would actually go about updating it, so for the average user they may just be stuck with whatever firmware they got from the factory. I did dump the SPI Flash, but binwalk did not detect anything interesting: code:
So I assume that the 48-pin QFP chip is an STM32 clone of some sort and probably the code lives on there. I didn't see any debug headers or dig into the pinout to confirm this though. I didn't have any luck looking up the part number stamped on it, which I believe reads FCK660R.1 but it's very difficult to make out. Gonna cross reference the layout with the pinout of the STM32F103 48 pin package later, but haven't dug that deeply yet. Reinstalled the flash chip and mine is back to stock, so at least I didn't damage anything afaik. I'll also make sure my binwalk has things like ARM microcontrollers in its database of interesting patterns because I'm using whatever it installed by default right now. Edit: Correction. I misremembered what I got when I weighed the V115. Forseti fucked around with this message at 21:31 on Oct 20, 2021 |
# ? Oct 20, 2021 21:20 |
|
I don't have my licence yet (exam on November 8th) so I've just been listening. A bunch of guys on a local 2M repeater were taking about getting RF burns from home made antennas. Is this really a thing? How can this be avoided?
|
# ? Oct 21, 2021 11:27 |
|
yoloer420 posted:I don't have my licence yet (exam on November 8th) so I've just been listening. A bunch of guys on a local 2M repeater were taking about getting RF burns from home made antennas. Is this really a thing? How can this be avoided? I thought this was part of the technician exam. Yes, RF emissions can heat up the meat you are made of. There are distances you need to keep from various frequencies at various power levels. They're published in a table. When I was studying for my license, i remembered this as "don't stick the antenna up your nose while transmitting".
|
# ? Oct 21, 2021 13:19 |
|
lol good ol’ hams I guess. The easiest way is dont touch the exposed metal wire portion of antenna. Commercially made antennas are usually insulated for exactly this reason. stopping you from touching the burn’y part. Something to be aware of though as a license holder. You are allowed to experiment and that’ll bring you closer to a scenario where you can get zapped.
|
# ? Oct 21, 2021 13:29 |
|
yummycheese posted:lol good ol’ hams I guess. The easiest way is dont touch the exposed metal wire portion of antenna. It didn't even occur to me people might be touching bare metal on a transmitting antenna. Holy crap.
|
# ? Oct 21, 2021 13:42 |
|
Yeah just WTF, "hurr durr hey Elmer, watch what happens when I touch muh haym raydio antenny!"
|
# ? Oct 21, 2021 13:51 |
|
Wrap a steak in antenna wire and transmit on 2m for two minutes or until well done.
|
# ? Oct 21, 2021 14:38 |
|
You'll probably want to add a ground strap to the steak for maximum flavor transfer.
|
# ? Oct 21, 2021 14:48 |
|
America posted:You'll probably want to add a ground strap to the steak for maximum flavor transfer. You definitely want an SWR (Sizzle to Wrap Ratio) of 1:1.
|
# ? Oct 21, 2021 15:43 |
|
they also couldve been talking about rf in the shack. radiated rf can build up potential in metal stuff in your shack; touching that metal can shock/burn you that seems much more likely to me - grabbing a transmitting antenna is really dumb, but rf in the shack is much less so rf in the shack isn't dumb at all - just a mistake. and it doesn't always occur, so as i understand it it's one of those problems you don't know you have until you get a zap. whether it happens often or painfully enough to actually _fix_ is up to an individual's taste and situation) Achmed Jones fucked around with this message at 15:59 on Oct 21, 2021 |
# ? Oct 21, 2021 15:57 |
|
cruft posted:
|
# ? Oct 21, 2021 16:10 |
|
Achmed Jones posted:they also couldve been talking about rf in the shack. radiated rf can build up potential in metal stuff in your shack; touching that metal can shock/burn you I could def see that, when I was learning about my NanoVNA I came across a very good video warning about just picking up the long coax from your antenna and screwing it right on. I hasn't thought of it but yeah, a coax is essentially a big capacitor and they can apparently accumulate quite a high potential. I remember I used to get shocked pretty good occasionally pressing the switch on top of our old TV to change the RF input from the satellite receiver to the roof mounted antenna, I'm guessing that's why. For the NanoVNA, all you have to do is discharge it by shorting the conductor to the shield first at least.
|
# ? Oct 21, 2021 16:52 |
|
Forseti posted:a coax is essentially a big capacitor and they can apparently accumulate quite a high potential. I remember I used to get shocked pretty good occasionally pressing the switch on top of our old TV to change the RF input from the satellite receiver to the roof mounted antenna, I'm guessing that's why. i meant things getting a charge during tx from basically turning into impromptu antennas, not charging up and storing like a cap. it hadn't even occurred to me that the shield and conductor could act like plates in a cap. thanks for bringing that up before i move to a place that lets me have big cool antennas and zap myself! for tvs i'd expect it was just the crt doing its whole "im a crt this is how i work" thing and putting static everywhere more than the coax but im wrong a lot, especially about electricity and radio and that's just a guess. fuckin electricity. fuckin radio. always surprising me
|
# ? Oct 21, 2021 23:31 |
|
I just got my two old rear end 27mhz walkie talkies working again, but the telescoping antennas are in bad shape. I wanna buy new ones and I got some questions. -SWR, I just heard about this, apparently this is how much of the energy that is sent to the antenna is actually radiated out and the rest is sent back to the unit, which means with badly adjusted SWR it could destroy the circuitry? How do I adjust the SWR, do I need a special unit for it, do I measure the antenna connected to the transceiver or separately before connecting it? Do you adjust it for a particular range or is it meant to compensate for random environmental oddities in your local transceiver? Should I go for a telescoping antenna, or would one of these work, the telescoping one looked like it was over a meter or 3 feet extended. I don't need that big a range I think. I see most antennas you can buy has a BNC connector, I am wondering if I should fit a BNC connector so I can easily swap antennas, if that is possible. Again coming back to the SWR thing, not really sure if one adjusts the antenna or the transceiver, if the latter I guess swapping out antennas without recalibration would not be a good thing. https://www.ebay.com/itm/154077276357
|
# ? Oct 22, 2021 04:57 |
|
His Divine Shadow posted:I just got my two old rear end 27mhz walkie talkies working again, but the telescoping antennas are in bad shape. I wanna buy new ones and I got some questions. SWR is basically the "fitness" of the antenna to the the frequency being used, and is effected by a ton of stuff, up to and including matching networks and antenna tuners. If you're buying a CB/27mHz antenna specifically, then any will work as long as the connector is right, as it will be purpose-made for that band and a low, or at least acceptable, SWR. If the old ones can come off, post a picture and any of us can identify a connector.
|
# ? Oct 22, 2021 14:12 |
|
This one has no connector, it's a soldered connectio and the thing goes into the body of the walkie talkie: I am thinking of fitting a BNC connector on the top instead. I was looking at those foldable antennas.
|
# ? Oct 22, 2021 17:43 |
|
Officially I don't think you're allowed to modify anything in the case and keep it part 95 accepted for CB if you're in the US. IDK how the law works elsewhere, check your local listings.
|
# ? Oct 22, 2021 18:51 |
|
Not been able to find anything specific except talk about difficult to interpret regulations, but also that nobody cares as long as you keep within specs.
|
# ? Oct 22, 2021 20:16 |
|
Are GMRS licensing questions OK in this thread?
|
# ? Oct 22, 2021 20:28 |
|
|
# ? Oct 22, 2021 20:45 |
|
|
# ? May 13, 2024 09:49 |
|
Alrighty then. I'm interested in getting a GMRS license but I'm not super comfortable with my name and address being available in a publicly searchable database. I've noticed that a lot of the people that use GMRS around me are far-right nutjobs that brag about their gun ownership and love of Drumpf. So my question is, if I use a name and address other than my own, am I running afoul of some federal law? Or should I just use whatever I want and not worry about the FCC? sporkstand fucked around with this message at 21:44 on Oct 22, 2021 |
# ? Oct 22, 2021 21:41 |