Lost in 52832 land.

RE: Lost in 52832 land.

Ken52832 — Wed, 27 May 2020 14:42:53 GMT

Yes I have a multi-meter and it is actually working when I test on VDD. My dev boards are coming in the morning. It looks like I have a problem with the connection to the probe. I am going to give the board back to the engineer. Once I get my Raytac board in the morning. I should be able to complete the code with all the information you guys have gave me.

Thanks for all your help and I might be back if this doesn't just start working with the new Raytac board.

RE: Lost in 52832 land.

haakonsh — Wed, 27 May 2020 13:46:57 GMT

A battery works great, do you have a multimeter to test the voltage of the battery first?

Don't worry about the voting system, it doesn't really work anyways xD

RE: Lost in 52832 land.

Ken52832 — Mon, 25 May 2020 18:50:38 GMT

Not as of yet. I was supposed to get a couple of the Raytac dev boards but Amazon is playing silly buggers. I will try to read the value straight off the battery on this current board we built. That should be something that makes sense since the board is obviously working. Thanks for the suggestion. By the way. I would like to click on what ever I need to to thank you guys for your answer but my screen reader is not telling me what I should click on. I have buttons on this screen like up, down, verify what should I be clicking on so you guys get ranked.

RE: Lost in 52832 land.

haakonsh — Mon, 25 May 2020 10:24:01 GMT

Have you sampled a known reference voltage and compared the output of the ADC?

RE: Lost in 52832 land.

Ken52832 — Thu, 21 May 2020 05:09:13 GMT

OK I do have another question. I am using the SAADC example from the SDK peripheral section of examples. I have compiled it for my Raytac MDBT42Q module which has an NRF52832 chip as its core. I have read the PDF on the Raytac module and it has chip pin 17 attached to P0.04 AIN2.

I have changed the Saadc to use AIN2. I ran the code and watched the output in the log.

I have a photo resistor attached between VIN of the board to the P0.04(AIN2)

I expected to see values from 0 to 1024 because the resolution is set to 10 bit. This is what I see on my original Arduino INO sketch.

I am not seeing values from 0-1024. I am seing a few values ranging from -1 to -8.

Again I have changed nothing in the SAADC but the AIN2. So my question is how do I get this to give me values I am expecting. Is this some kind of big little indian problem or what am I seeing in this buffer.

Note if I change the AIN to values of other pins that I have nothing attached to. The log shows me values like 16,18,17 in the buffers.

So I am a bit confused at what the samples are. I have read as much as I can find on this and I am confused.

RE: Lost in 52832 land.

Ken52832 — Wed, 20 May 2020 10:33:03 GMT

I want to thank both of you for all your replies. I still would love to pay someone to just do this quicker than I can but after your answers I have now got my custom NRF52832 board running with the GDB server and I can see the logs as well as be able to use the GDB with break points which makes my life better.

I spoke with the J-link folks and because of my company being NOn-profit and the fact that we don't make a dime off of what we build. They said they don't have a problem with me using he Mini. I still will be ggetting one of our J-link pro's from work next week when the boss goes on a delivery run.

Before that happens though since you clearly explained why I can't debug the way I wanted I have broke my project up into all the services as modules and I am going to test all the ones that have nothing to do with softdevice and Ble. If all 4 of the other parts of the project work. I have a good feeling about being able to quickly put them all together and test them with BLE.

In short. Thanks for all the help and unless the world goes even crazier. We can probably close this Q&A. Of course if someone wants to jump in and say they will do it for a few hundred bucks I will step away and run for one of the big Applications I am working on.

Ken

RE: Lost in 52832 land.

Turbo J — Wed, 20 May 2020 10:19:27 GMT

NRF Logs are put on the RTT port 19021 and not on the telnet port 2333.

RE: Lost in 52832 land.

haakonsh — Wed, 20 May 2020 09:46:17 GMT

The J-link's user guide states: " Default [Telnet] port is 2333"

RE: Lost in 52832 land.

Ken52832 — Tue, 19 May 2020 10:37:26 GMT

I will take a look at the board with the NRF52832 on it. With that said my code is now running. It for sure is not done. I will probably remove the need for the softdevice until everything else is working or until my boss swings buy and rops off the other J-link. With that said I should be able to see the logs right? I can step through the code and print values but it would be nice if I can just watch the logs scroll buy on the command line. Are the logs supposed to show up in the telnet localhost 19021? I haven't seen a log yet but then I just got it fully compiling and running just now. If I am right and the logs are supposed to show up in Telnet then I am all good.

I will pop back in here if the logs don't show up.

RE: Lost in 52832 land.

haakonsh — Tue, 19 May 2020 07:42:56 GMT

MMD only works on J-links as far as I know. The debugger changes the breakpoint, single-step, and halt CPU commands into an IRQ, I don't know of any other debugger that does this except Segger, though I have not looked, as Segger is what we use.

For the nRF52832 you can use many of the nRF52 DK's, but I recommend that you get the PCA10040 as that has the 52832 on it.

RE: Lost in 52832 land.

Ken52832 — Mon, 18 May 2020 15:40:05 GMT

I had read their policies. WE purchased a few at work but According to their license if it is for non-profit and I am not making money on it I could use this. I will write them though. Fact is the only reason I am not using a pro one is I have not been into the office since the Corona craziness.

With that said I guess I could be using the raspbury pi to do this rather than Jlink.

My question is which DK should I get for NRF52832 and where is the best place to buy it?

RE: Lost in 52832 land.

haakonsh — Mon, 18 May 2020 13:18:37 GMT

The SoftDevice will assert if any timings are missed, I don't see how you can get around that when a CPU gets halted in a breakpoint.

You can evaluate MMD with the nRF52 DK's onboard J-Link, but you can't use it to develop a product. You can ask Segger directly whether "evaluation" covers your use-case or not. Anyways I suggest that you try out MMD on our DK as it is a good debugging tool in general that works with all Cortex-M processors.

RE: Lost in 52832 land.

Ken52832 — Mon, 18 May 2020 12:34:21 GMT

It doesn't look like that will work. According to the Secure link you sent it only works for the pro and up versions of JLink. Iam using the education version right now. I will have to see if I can get one from work. They might have a JLink pro that is sitting idle.

Will the old way with PRIMask and changing break points work with the JLink education version?

RE: Lost in 52832 land.

haakonsh — Mon, 18 May 2020 12:22:46 GMT

There is a way to debug with SoftDevice, it's called Monitor Mode Debugging, or MMD for short.

See this link to SEGGER for an explanation: https://www.segger.com/products/debug-probes/j-link/technology/monitor-mode-debugging/.
And this link to a github repo for a tutorial with code: https://github.com/NordicPlayground/j-link-monitoring-mode-debugging

RE: Lost in 52832 land.

Ken52832 — Mon, 18 May 2020 12:02:28 GMT

Thanks that actually did it. I have it running and can step into main. I am still a long way from being done but at least I am moving again. Is it still true that you can't debug with softdevice? I saw something about having to break before any soft device function and then jump it and break after with a flag called PRIMASK?

At least I know my software is running. Now to start trying to figure out where it is dieing and why.

Thanks to both of you that answerd. I still think I should have paid a 14 year old to do this. I might be getting to old *GRIN*.

RE: Lost in 52832 land.

haakonsh — Mon, 18 May 2020 11:21:25 GMT

The standard output from gcc is an .out file that is actually an ELF file. Try debugging the .out file.

RE: Lost in 52832 land.

Ken52832 — Mon, 18 May 2020 03:40:49 GMT

First I want to thank you because I forgot to in my last reply. Thanks for the command line debugger. I had not seen that one. I still have not got it to work yet for either the logs or GDB. I think if I can get the debugger working the logs will work. Here is what I ahve tried

I have got the jlinkgdbservercl working with:

jlinkgdbserver -device nrf52832_xxaa -if swd -port 2331

I usd telnet with the port you posted. All seems to connect correctly. Other than the connection messages I get nothing else.

I Then try to compile my program with debugging

I changed the opt line from -O3 -g to

-O0 -ggdb

I also tried just g alone but I found -ggdb as an option on one of the documents I was reading on line. Neither of them seem to create a .elf file and the hex file it does produce does not have debugging symbols at least when I do

arm-none-eabi-gdb SALSProbe.hex

It says there are none. Everything I read says it should make an elf file but it doesn't. Can you or anyone tell me what my possible problem is? My Makefile is in the original zip file of the question and all I changed was the opt lines. Is there something else I need to change?

ken

RE: Lost in 52832 land.

Ken52832 — Sun, 17 May 2020 19:35:34 GMT

Ok. I went out and read a page on the jlink server gdb. I am used to the command line gdb which I use for a C++ server game project Irun in linux. I have also ran the jdb server for a JAVA project I work on. I ran into a slight snag here though.

I am compiling at the command line with make. It is attached to the original question. It looks like they did not make a default debug target in the Makefile of the SDK project I used as a base line. I figure if I was using an IDE it would probably add it but I am not. When doing regular C projects I just add -g to a target and use that. I am a bit confused as to how I would add it to this make file. It is using ooptimisation and -g3 which I will ahve to go read on. If you know how to make a debug .elf using the Makefile in my project that would help.

RE: Lost in 52832 land.

Turbo J — Sun, 17 May 2020 08:27:09 GMT

[quote userid="90399" url="~/f/nordic-q-a/61478/lost-in-52832-land"]For example, the new J-link tools are not accessible with my screen reader[/quote]

There are command line versions of these SEGGER programs, e.g. using JLinkGDBServerCL.exe should get rid of those unreadable UI elements.

[quote userid="90399" url="~/f/nordic-q-a/61478/lost-in-52832-land"]I think that is because I can’t see the log warnings telling me to expand the ram in the config.[/quote]

Use the above JLinkGDBServerCL to flash and run your program (with arm-none-eabi-gdb, target is localhost:2331). Then connect to the RTT port via telnet client, host is "localhost" and port is 19021 (telnet localhost 19021).

Both nordic command line tools and SEGGER are available for Linux, this might be simpler to setup with reader or Braille display.

There is unfortunaltely no quick way for converting your .ino into an SDK program (that I know of). And no, you don't want to program an ARM Cortex-M in Assembler. These are meant to be programmed in C.