Keyboard problem

[retrofan11]

Hi,
Just flashed the NextNano firmware on the Tang Nano 20k and apparently it works, but I stuck , on the video mode screen, because I can't get any of the usb keyboards to work.
I tried with a powered usb hub, as well as with separate power supply via GPIO pins and direct connection of the keyboard to the usb-c port, but without success.
Which hub did you use and did you test another keyboard, apart from the one in the pictures?
Is this FPGA companion necessary for the keyboard to work or can it be done without it?
I'm pretty stuck here, any advice is appreciated.

[RetroSilicon]

Sorry you’ve run into trouble. I’ve tested several USB hubs, and all but one worked. The only one that didn’t work was the most sophisticated model, which also had an HDMI port.

Here are two hubs that work for me:

• https://aliexpress.com/item/1005007432614449.html
• https://www.amazon.co.uk/Docking-Station-Charging-Capability-Multiport/dp/B0DKFT5VC3
  Even this tiny adapter works:
  https://www.amazon.co.uk/Seminer-Adapter-Charging-Splitter-Chromecast-Gray/dp/B0DCBNGXKG
  However, it doesn’t work reliably on every restart—roughly 50% of boots are successful.

I’ve also tested several wired and wireless keyboards, and they all work:

• EWEADN MK84
• Satechi SM1
• Keychron K3.
• Several other wired USB keyboards.

Flashing the FPGA Companion firmware—both files—to the built-in BL616 is absolutely necessary. USB peripherals will not work without it.
You can find a video showing the flashing process here:
https://youtu.be/odAb_soEHIo

Let me know if you need any more information.

[retrofan11]

Thank you, lots of useful info here, I need to flash the fpga_companion files first and let you know how it goes.
Just one additional question, have you tested the different Tang Nano board versions?
I have board v.3923, have you had success with it in any Hub/keyboard combo?

[RetroSilicon]

I only have v3921 boards, but I’ve just received a reply from Sipeed confirming that there are no significant differences between revisions 3921 and 3923.
According to them, the BL616 section is unchanged, so USB peripherals should work with both board revisions.

[retrofan11]

Ok, I just successfully flashed both fpga_companion files, exactly as you described in the video, but unfortunately still no success,
I've tested with three keyboards (2 wired and 1 wireless), which have been proven to work for me on different retro devices, but I'm still stuck on the video mode screen.
At this point, I'm starting to doubt the board version, because I also noticed similar unsuccessful attempts with it on your Facebook page.
I will try to get one of the usb hubs you recommended (I have different one) , but the only exact check will be when you can personally test this new board revision.
I'm pretty disappointed right now, but I won't give up, there has to be a solution.
Cheers.

[LeonisX]

Hi. I flashed the device exactly as shown in your video. I tried the following devices:

Keyboards:

• Rii i4 Mini wireless keyboard
• Chinese Dell wired keyboard
• Sven wired keyboard
• Cherry G84-4100

USB hubs:

• YX04-X4
• Orico MH4UC-U3 - don't buy, its PC-only device
• UGreen CM431 - don't buy, its PC-only device

Unfortunately, the keyboard still doesn't work. I have a Tang Nano version v3923.

It seems to be a version issue, as Orico and UGreen devices, especially the UGreen, 100% provide additional power.
Also, I also tested the hub you showed in one of the examples.

[RetroSilicon]

Could I ask you to try the following?

• Make sure that FPGA Companion is installed on the built-in BL616.
• Flash your Tang Nano 20K with the appropriate debug core from this repository:
  https://github.com/MiSTle-Dev/bl616debugdisplay
• Connect the display and power the board through the USB hub, but do not connect any USB peripherals yet.
• Let me know what appears on the display. A photo or short video would be helpful.
• Then connect the keyboard and let me know whether any new messages appear on the display.

As I don’t have a v3923 board myself, this information could help me identify and hopefully resolve the issue.

[retrofan11]

I did this test, the first picture is with a connected powered usb hub without peripherals, the other two are with a wired and wireless keyboard attached.
I don't know what this means, but I didn't notice any difference between them, no message appears when the keyboard is connected.
Maybe I didn't do something right or my usb hub isn't good enough, but this is the best I can do.
Maybe someone else will have better results.
https://imgbox.com/aFGLf46w
https://imgbox.com/YOzimzaI
https://imgbox.com/KHVEK1Hm

[RetroSilicon]

Well, it looks like there may be a communication problem between the BL616 and the FPGA.

In that case, please try the following - Flash the BL616 using BL DevCube with the files from this release:
https://github.com/MiSTle-Dev/FPGA-Companion/releases/tag/v1.4.21

Please flash:
bl616_fpga_partner_nano20k.bin at 0x00000
fpga_companion_nano20k.bin at 0x40000

Then try the debug core again. Start with only the USB hub connected, and once it has booted, connect the keyboard.

[LeonisX]

bl616debugdisplay (flashing cursor)

Maintenance BL616 and PiPico: w/o keyboard, attached keyboard, detached keyboard.

[RetroSilicon]

Do I understand correctly that the first photo is from the Tang Nano 20K using its built-in BL616, while the other three are from a Pi Pico connected to the Tang Nano?

The messages seem to suggest a problem with detecting the keyboard. Have you tried connecting a different keyboard?

[retrofan11]

Ok, now I have response from both of my wired and wireless keyboards.
I did two types of tests with both of them, first with a powered usb hub, and then with an external 5V power supply on GPIO pins 1/2 and a regular OTG cable in the usb port, without the hub.
Here is the whole gallery (from left to right), pictures 1/6 are without connected keyboards, pictures 2/7 and 3/8 with connected and disconnected wired keyboard.
Pictures 4/9 and 5/11 are with the wifi keyboard connected and disconnected.
Figure 10 is the OTG cable I used.
https://imgbox.com/g/2A3MGm36cO
So, in both cases the keyboards responded the same or very similarly, what are the next steps?
btw, I have to ask, is it normal for the Tang Nano to heat up quite a bit, already after ten minutes?
It heats up especially when I connect it to the usb port of the computer, less if it is connected to an external 5V psu.

[RetroSilicon]

Thanks a lot for the photos, based on your captures, the USB host itself is working: your powered hub (a full-speed device) enumerates with a completely valid descriptor set. However, both of your keyboards are low-speed devices, and for those the control transfers return partially corrupted data. The pattern is identical on the direct port and behind the hub, with both power schemes — so this doesn't look like a power or keyboard problem, but like a broken/marginal low-speed host path specific to board revision v3923.
I'm contacting Sipeed support about this and will report back.
In the meantime, if you have access to other keyboards, could you test them? A full-speed keyboard may well work — most mechanical/gaming keyboards are full-speed.

[retrofan11]

Thanks a lot for the reply, I appreciate it.
If it turns out that the whole v3923 board revision is broken in that part, will it even be possible to adapt Next core to work on it or will it be useless?
I tried three keyboards and none of them worked, I don't have another one, maybe someone will test the mechanical one you recommended.
So far I haven't seen a single confirmation that someone has had a working keyboard with the v3923 board, which is quite discouraging.

[retrofan11]

I forgot to ask this too.
Does NextNano core also support fpga companion based on rp2040 ?
https://github.com/MiSTle-Dev/FPGA-Companion/tree/main/src/rp2040
I know it already supports the M0S dock method, but this Pico based interface is very easy to build and maybe this keyboard issue can be overcome that way?

[RetroSilicon]

I can't promise anything until I actually get a v3923 in my hands, but I'd say it's very likely that only a minor FPGA Companion adjustment will be needed. And even in the worst case, there's a simple workaround available today: attach an external RP2040 (a tiny RP2040-Zero would do) with a few wires and route your USB peripherals through it. No waiting for firmware fixes required. So don't worry — your v3923 is definitely not useless :)