FNIRSI GC-01: Debugging #14
Comments
Gissio
changed the title
|
FNIRSI GC-01 support is not ready yet (but it won't take long). Once it is, yes, you will be able to log data. In the meantime, watch the project to get notified when a new release comes out. |
|
So quick research on the chip being used, CH32F103 R8T6, and found the following. |
|
Don't worry, Rad Pro for GC-01 is really around the corner! |
|
hi. first of all, I want to thank you for the work on those devices. but there is a boot error on the fnirsi GC-01 device with firmware rad pro 2.0 beta. the device does not boot the firmware, but fortunately Windows can still open it and upload the old firmware to it. I had firmware 1.6 since new, now I have 1.6-2 which I downloaded from the net but it looks like it multiplies the dose. don't you have the original 1.6 uploaded/stored somewhere? thank you |
|
Unfortunately I only have firmware 0.3 and 1.5, which you can find here: https://github.com/Gissio/radpro/blob/main/docs/install-gc01.md. Can you refer me to a link to firmware 1.6? |
|
Ofcourse. Here is it: https://drive.google.com/drive/folders/1XN-d1ZiFq0ur_ae392iQr-pbHZY32Ag5 But I don't have the clean version 1.6,I've had it since buying a new device, but it's nowhere. I have also tried writing an email to support but so far no response... So that's what I'm looking for. There is only 1.6-2 where there is an extended translation and then in 1.5 where fonts are missing after reboot unless I activate them in the settings by selecting the language. The problem is that in version 1.6-2, my background dose jumped double. I only have tungsten-thorium electrodes as a source for measurement, in any case, the computer assembly does not react to them. Even though I know that thorium is primarily an alpha emitter, but from some sources I saw measurements for beta particles from this element and specifically from this source, and they were evident.. I have nothing to compare it with. It would be ideal if someone who has two GM computers tested it, and even more ideally if they compared it to one that is professionally calibrated. |
|
Thanks for sharing! That's odd, the GC-01 contains 64 kB flash, of which 16 kB are reserved for the bootloader. Thus the firmware should never be larger than 48 kB, yet the file you sent me is larger than that. Strange indeed. I'm currently fixing a USB problem with the GC-01. Hopefully Rad Pro will work on the GC-01 after that. |
|
I have added a new release. Can you try now? |
|
Clearly. But there is the same problem as with the previous release. After uploading the firmware, the display just turns off and does not respond to anything, except invoking the transfer mode with a PC. |
|
Thanks for your patience! I'm sending you a special test firmware: After flashing it, unplug from USB. Keep the ok/power key pressed. The pulse LED should flash 13 times. If it flashes fewer times, please let me know how often it flashes. This will let me diagnose the problem. Now you can release the ok/power key. You should see a screen similar to this one: 
If you get to this point, please attach a photo of the screen in your message. pd: One more test. If you release the power key after the pulse LED flashes for the fourth time, does the sequence continue? |
|
Hi, same trouble, my device flashes 13 times, then restarts ... with FW in zip file. |
|
That's actually great news: all the peripherals are initializing correctly! I just found a very silly error with the LCD. Again, thanks for the patience: Even if the screen does not work, let me know whether:
|
|
|
|
It seems the LCD display is receiving commands now, but something is still off. Let's do one more test: as before, the output should be the white screen with four lines of text. The LED and the buzzer should pulse periodically. Again, thanks for your patience. |
|
radpro-beta17.test4 - now there is only a white screen, LED and buzzer that blink irregularly. No characters, no image failure, just a lit panel. That's okay. I guess we can fine-tune it together somehow :) |
|
Let's try one more time. This time I'm sending two tests. Let me know which one(s) work. |
|
When I hold the power, it lights up the panel with no characters, the LED blinks, the buzzer beeps, and in a second it all goes out. After a long pause, it takes longer if it all starts at all (5-6 seconds). it is the same in both cases from test 5 and 6 |
|
Can you make a short video? |
Video.WhatsApp.2024-02-23.o.21.34.02_6231c323.mp4After an even longer pause, it loads for 15 seconds. |
|
I’m attaching photos of the board in case you need to compare it with yours. Does everything work for you when you upload firmware 1.6-2? |
|
That was fast! Thanks for the video and photos. It seems the device is resetting at the moment the screen turns off. I'll get in touch once I have news. |
|
I just put it together when you asked for a video so it went quickly. Great, I’m keeping my fingers crossed for you. |
|
I think I found the culprit! In case you are interested: the code that reads the ADC never exits, so the watchdog timer kept restarting the device before writing to the screen. Let's try once again: If it works, please press the keys (one at a time) and see if GPIOA and GPIOC change. |
|
JantarX They look similar, I also have a version with J321 tube, later versions come with cheaper one, but it also looks like we have a different LCD module and I don't even have the option and pins on the board to connect a smaller and cheaper tube. I have one of the first versions sold by a local dealer, not the current one from ali. I am attaching a photo.
Does anyone know how to successfully upload new firmware in linux ? BOOT partition is visible, readme.txt is visible, fw can be uploaded, but initialization and restart is not done, after disconnection the old fw is still active. What is the difference between uploading in win and linux, does windows have any special device driver ? I don't have a chance to reboot into win for another test today :( |
|
Photo of one of the latest versions from aliexpress. The same box, different hw, it seems like fun for debugging fw, wouldn't it be useful to use some debug port to get data ? Trial and error will take ages. |
It is restarting without any changes compared to tests 5 and 6. |
Have you tried version 1.6-2 that I shared above? Does it work on both of your devices? As for Linux, I personally can’t help you. The only thing I know is that you literally need Windows 10 and a USB 2.0 port to run the bootloader. I tried it on USB 3.0 but it didn’t launch. I tried it on Ubuntu and the whole bootloader crashed. I had to disconnect the battery and flash it according to the conditions above. |
|
I tried 1.6-2 and all subsequent test versions only on a device with J321
tube, with the same result as you have. I don't have the other device
available, I only have a photo.
More info for linux ( I have no experience with STM32 devices, maybe it is
trivial, it would help me in further testing. ):
$ lsusb
Bus 001 Device 026: ID 0483:002a STMicroelectronics AIR01
$ fdisk
Disk /dev/sdc: 100 MiB, 104857600 bytes, 51200 sectors
Disk model: DETECTION 1.0
Units: sectors of 1 * 2048 = 2048 bytes
Sector size (logical/physical): 2048 bytes / 2048 bytes
I/O size (minimum/optimal): 2048 bytes / 2048 bytes
105 MB — 82 KB free
FAT (12-bit version)
so 24. 2. 2024 v 8:59 odesílatel JantarX ***@***.***> napsal:
… JantarX <https://github.com/JantarX> They look similar, I also have a
version with J321 tube, later versions come with cheaper one, but it also
looks like we have a different LCD module and I don't even have the option
and pins on the board to connect a smaller and cheaper tube. I have one of
the first versions sold by a local dealer, not the current one from ali. I
am attaching a photo.
Does anyone know how to successfully upload new firmware in linux ? BOOT
partition is visible, readme.txt is visible, fw can be uploaded, but
initialization and restart is not done, after disconnection the old fw is
still active. What is the difference between uploading in win and linux,
does windows have any special device driver ?
I don't have a chance to reboot into win for another test today :(
Have you tried version 1.6-2 that I shared above? Does it work on both of
your devices?
As for Linux, I personally can’t help you. The only thing I know is that
you literally need Windows 10 and a USB 2.0 port to run the bootloader. I
tried it on USB 3.0 but it didn’t launch. I tried it on Ubuntu and the
whole bootloader crashed. I had to disconnect the battery and flash it
according to the conditions above.
—
Reply to this email directly, view it on GitHub
<#14 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJPKQUZNSBNAQ3S2SFXPL3YVGMWLAVCNFSM6AAAAABBZQJGQKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNRSGI4TCMZZGM>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
|
Hey folks … have been watching this thread for a week or two .. am
available to do some testing….
…On Sat, Feb 24, 2024 at 3:36 AM Michal Matouš ***@***.***> wrote:
I tried 1.6-2 and all subsequent test versions only on a device with J321
tube, with the same result as you have. I don't have the other device
available, I only have a photo.
More info for linux ( I have no experience with STM32 devices, maybe it is
trivial, it would help me in further testing. ):
$ lsusb
Bus 001 Device 026: ID 0483:002a STMicroelectronics AIR01
$ fdisk
Disk /dev/sdc: 100 MiB, 104857600 bytes, 51200 sectors
Disk model: DETECTION 1.0
Units: sectors of 1 * 2048 = 2048 bytes
Sector size (logical/physical): 2048 bytes / 2048 bytes
I/O size (minimum/optimal): 2048 bytes / 2048 bytes
105 MB — 82 KB free
FAT (12-bit version)
so 24. 2. 2024 v 8:59 odesílatel JantarX ***@***.***> napsal:
> JantarX <https://github.com/JantarX> They look similar, I also have a
> version with J321 tube, later versions come with cheaper one, but it
also
> looks like we have a different LCD module and I don't even have the
option
> and pins on the board to connect a smaller and cheaper tube. I have one
of
> the first versions sold by a local dealer, not the current one from ali.
I
> am attaching a photo.
>
> Does anyone know how to successfully upload new firmware in linux ? BOOT
> partition is visible, readme.txt is visible, fw can be uploaded, but
> initialization and restart is not done, after disconnection the old fw
is
> still active. What is the difference between uploading in win and linux,
> does windows have any special device driver ?
>
> I don't have a chance to reboot into win for another test today :(
>
> Have you tried version 1.6-2 that I shared above? Does it work on both
of
> your devices?
>
> As for Linux, I personally can’t help you. The only thing I know is that
> you literally need Windows 10 and a USB 2.0 port to run the bootloader.
I
> tried it on USB 3.0 but it didn’t launch. I tried it on Ubuntu and the
> whole bootloader crashed. I had to disconnect the battery and flash it
> according to the conditions above.
>
> —
> Reply to this email directly, view it on GitHub
> <#14 (comment)>,
or
> unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AAJPKQUZNSBNAQ3S2SFXPL3YVGMWLAVCNFSM6AAAAABBZQJGQKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNRSGI4TCMZZGM>
> .
> You are receiving this because you are subscribed to this thread.Message
> ID: ***@***.***>
>
—
Reply to this email directly, view it on GitHub
<#14 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADKN6MUVPKSTXM6XQDUBDJTYVGRCHAVCNFSM6AAAAABBZQJGQKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNRSGI4TSMRVGA>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
I Tested the Voltage of both of my GC-01 for the Tiny Tube I got stable 293.0V, for the J321 Tube it was first 300.0V but colapsed fast. PWM 10kHz PWM 2.5kHz |
|
Great! radpro-fnirsi-gc01-2.0beta17-test3.zip works but the display's power-saving mode needs to be adjusted. After the set time, the display does not darken but lights up white. I really appreciate the tube selection, beeping without HW upgrade and logging. Brilliant! |
|
When the GC-01 is charging, do you see a lighting icon on the battery indicator? |
No. I see only how the display slowly fills up and shows full within minutes. It's the same when you remove the device afterwards and within minutes the display goes back to showing empty. The red light next to the USB port is permanently on when charging. |
|
Does the radiation alarm work? It should flash the screen (in the current version, when the screen becomes white after sleep, it should briefly show the screen), beep the buzzer and trigger the vibrator. |
Yes, it works all. The screen flash after sleep, vibration and buzzer |
|
Does the dead-time measurement work? Check the statistics, the dead-time should diminish as time goes by (the measurement should converge after a few hours). Also in the statistics, does the device temperature show a reasonable value? Does the random generator work? Try "coin flip", it takes two pulses to generate one bit of data. |
|
I think I have found a strange bug. I tried the power button, it did not show "RADPRO 2.0 Beta" as it should, it showed the Instantaneous screen and a value for a fraction of a second and then went blank again. I tried the hardware reset button... Now it works again I will leave it running for now, hopefully I can reproduce the behaviour and film it next time. |
Can you tell me approximately how many hours that was? Mine ran for about ~9 hours without any problems with 2 charges via USB for about 1 hour.
I'm not exactly sure what you mean but as you can see in the picture: Tube Life Time: 10:14:09 Are these good values or not? I'm only asking so that I know whether the software would be usable in an emergency despite the beta
The value in the ~9 hours was between ~26-30.xdegrees Celsius. It had 30.x after about 60 minutes of charging (Maybe a bit more, I didn't look at it immediately after loading, but I will do so today). After Start and a little bit random generator it has 27.3 degrees Celsius. After ~12 Min Pulses it has 28.4 degrees Celsius.
Yes, that works. Tested Coins and a few others (but not all). Sometimes faster, sometimes slower. But that has to do with the pulses. I will run the GC again today for several hours and see if Tig3rch3n's problem also occurs for me.
|
|
Tig3rch3n: thanks for the report. Fortunately I had the same problem with my device some days ago, so I should be able to debug it. Probably a new bug. What microcontroller does your device have (WCH or Geehy)? w01fdev: Schaut gut aus (looks great)! Let me explain what dead-time is: Geiger-Müller tubes are saturated for a brief time after each pulse (usually 50-500 µs), which affects high-rate measurements. The dead-time measurement lets you measure this time, and dead-time compensation lets you compensate it. You can read more about it here: https://en.wikipedia.org/wiki/Geiger%E2%80%93M%C3%BCller_tube#Quenching_and_dead_time You also mentioned you required two USB charges in a 9 hour period. How long would the original firmware last? It should last way longer. One more question: could you turn on the device, connect through USB and see if a serial port device appears on your computer (under Windows, check the Device Manager)? If that works, can you use a serial port terminal (like Putty) to send some commands and see if you get a response? Please let me know the microcontroller of your device (WCH or Geehy). Thanks. |
I have to say: I have set the display so that it is permanently on. It doesn't help with the white screen, as it still draws power
Wunderbar. Das freut mich
Thanks for information
I have geehy. So under Linux I can't find any device with lsblk, lspci.. I'll have to have a look at Windows later. I use it very rarely. |
I completely forgot about this nasty fact. Now it makes sense!
Under Linux, try |
So I just had the device running for 3 hours without crashing. Then I restarted it and now I'll let it run until this evening. It probably won't crash because display sleep is deactivated. I've also set the brightness to the lowest level, but that's probably not the problem. Brightness doesn't work either.
With lsusb and also otherwise it shows nothing with Linux. You mean when the GC is running, right? Overall, the values for the j321 tube seem a little high to me. All the gmc on the website have lower values within a few kilometres of where I live (but you can't see the tube on many of them). E.g.: https://gmcmap.com/historyData.asp?Param_ID=48541894204 My average after 43:45 and time 05:19 today are 0.215. with original firmware it was ~0.16-0.18 average (If I remember correctly) You also said that there is a 5% chance that it will go above +-50%, but for me this is more the rule with the 'instantaneous' screen. Often ~+-51%-57% is displayed. More rarely+-78 and hardly ever values below+-50%. I have only seen values below +-50% twice so far. At average, for example, it is just at +-4.9% |
|
Sorry to comment again under my last post. But I may have found the problem of the high values and peaks: I saw in your code that you gave the j321 the same values as the j305. But the j305 comes in 2 different sizes. ~90mm and ~107mm. I suspect that the data for the calculation comes from the 107mm. On many pages the m4011 is mentioned as being virtually identical to the j321. Now I have selected the tube. And as you can see in the pictures, the values are now much closer to those of the fnirsi firmware and are also very close to the GMC values near me on the website. You can also see from the pictures that the line at 60 min is similar to that of the fnirsi firmware and no longer makes so many strange peaks in both directions. The pictures are probably self-explanatory.
|
Maybe it would be a good idea to crowdsource some PCB Pictures to see if there are different PCB Iterations :P
About the Time to getting stuck. As mentioned, Brightness Controll does not work atm and Sleep only turns the Display White. If you are Interested, a friend of mine has a Nordic Power Profiler II, I could run a Comparison between the Original Firmware and the last Beta
|
|
Rad Pro 2.0beta18 was just released! |
|
Here's mine just for reference (If you're looking for more PCB examples):
|
|
Rad Pro 2.0beta18: The display brightness setting and sleep mode are now working, but I have a frozen charging icon after disconnecting from the PC, even after restarting and resetting the device. |
|
w01fdev:
JantarX:
To all:
|
Yes, that has improved considerably. Only in brightness low would it be good if it were a little bit brighter
The battery icon is really annoying, because you no longer know how full or empty the device is.
I can test it, but preferably only as soon as the battery indicator is working again. By the way, the other profiles are now gone. There are now only factory default 1, 2 and custom I'm not quite sure if that's the reason. Yesterday at about 19:00 I had loaded for about 1 hour. With a standard USB charge and 1100 mAh battery, that should mean at least 50% battery. Display set to switch off after 30 seconds. Brightness set to low. Nevertheless, the device was empty around 0130 hrs. I think I've read that the fnirsi with standard firmware lasts approx. 11 hours (but I haven't checked this myself yet). Then it should have lasted until 0130 with the display off. I have a few general questions. Downloaded quite a lot yesterday for developing the boards. It would be cool if you could give me a brief answer. I have to familiarise myself with it first and that will take some time:
As soon as the icon is fixed I will test it thoroughly (HV profiles, different modes, display on/off, brightness and then post the values here). A function to switch the vibration on and off might be useful at some point. In addition, perhaps icons so that you can see what is active and what is not. In general, wouldn't it be better to slowly create a separate issue for each problem, or do we want to keep it that way? It's just not quite as clear when it gets into the finer details and becomes more complex. |
There is no "smallest mode", just ways of using power more intelligently.
USB requires 72 MHz. And it is not easy to change CPU speed on the fly.
IMHO, cost-benefit doesn't cut it.
Read the developer's manual.
The charge controller stops charging the device once it reaches a target voltage. Don't worry.
Beta19 goes back to the previous behaviour. It turns out that the GC-01 hardware reports "charging" when USB is not plugged in, so I decided to remove the indicator.
Agreed. Please post each new issue as a separate thread. |
|
I have one more request for everyone in this thread. In order to help people who are having trouble installing Rad Pro on GC-01's with a CH32F103R8T6 processor (WCH), I need volunteers who own a GC-01 with a WCH processor and do the following:
fnirsi-gc01-bootloader-dumper.zip
If somebody else already uploaded the dump, it would be interesting that you repeated the process, just in case there are different versions of the bootloader. |
Gissio
added
bug
|
May I extend this request to bootloaders from Geehy APM32 - F103RBT6 devices, too? |
and others
I want to run a script on my PC that would email and SMS me data log of when the measurement goes +X%.
Is it possible to grab the readings or logs from the GC-01.
This is the one I got.
https://www.amazon.com/Geiger-Counter-Nuclear-Radiation-Detector/dp/B0BHH9X1WG/ref=sr_1_4?keywords=nuclear+radiation+detector&sr=8-4&ufe=app_do%3Aamzn1.fos.18ed3cb5-28d5-4975-8bc7-93deae8f9840
Also wanted to hook it up to my android phone to map rad levels while traveling on plane.
Thanks for the help.
The text was updated successfully, but these errors were encountered: