Is This Situation Repairable

Question

A while back I posted a blog about building a DSO Shell 15001K Oscilloscope Kit.

https://www.element14.com/community/people/jw0752/blog/2017/04/25/building-the-dso-shell-15001k-oscilloscope

I gave one of these kits to an old friend and he called me a few days later to say that the preliminary test of the Main Board which is done before one starts assembly showed it to be failed We tried a number of diagnostics over the phone and finally I had him send the board to me My tests revealed that there was a low resistance short on the 3.3 volt rail I traced the circuit and removed smd components and cut traces to try to isolate the branch of the 3.3 volt circuit and the component that might be causing the problem Ultimately the fault was in the STM32F103C8T6STM32F103C8T6 microprocessor I removed the microprocessor and decided that the unit was going to be just parts While I can technically replace the uP I have no way to reprogram it As I studied the schematic I noticed that there was in addition to the uP an 8 pin 24LC32A EPROM tied to the SOA and SCL pins of the uP Here is a link to the main board schematic

http://www.jyetech.com/Products/LcdScope/Schematic_150main.pdf

My question is whether there is a possibility that the program for the unit is on the EPROM and not on the uP. Would I be able to replace the fried uP with a raw one and expect it to automatically boot up from the EPROM. At this point it is all just a learning experience but I will probably not waste my money on a uP or my time replacing it if it is impossible for the program to be on the EPROM.

Any insights will be appreciated.

John

jw0752 · Answer

Hi mcb1 Jan Cumps , , gecoz , I want to thank each of you for your awesome response to my question. Based on what you have told me I am sure it will not be as easy as just replacing the uP but your insights and suggested resources seem to indicate that the situation isn't hopeless. I have ordered a raw uP and I will install it when it comes. Then I will use your suggestions and see if I can reload the software. This is all new ground for me so it will be a great adventure. Thanks to each of you for taking the time to research and help me. I have too many correct answers to mark just one so I have marked all your responses as helpful which they were. John

Former Member · Answer

The STM32 micro's come with a bootloader pre-installed. This is activated by making BOOT0=1 and BOOT1=0 when the device starts up. I had to reprogram a dev board in this manner before and used the tx/rx port on a BeagleBone black to flash the program over to the STM32's 2 wire programming header. That being said, many people install a usb bootloader as a way of getting their software and updates onto the micro so you would need to make sure the code you have is for the built in bootloader and not for a user installed usb loader

Former Member · Answer

In the schematic Jumper JP1 goes to BOOT0 and if the jumper is joined it would bring it to BOOT=1 on startup. Im betting that Jumper JP2 connects to PB2 and that pin doubles as BOOT1. Joining that jumper would bring BOOT1=0 on startup and put the device into bootloader programming mode. For many STM32, its then possible to re-program it using one of its UART ports I would check that this option is available on your device before you invest the time though. Judging by the UART header in the schematic I would think it is a possibility.

mcb1 · Answer

Does it still have the short. These EEprom do have a supply, and it would be very close to the uP, so it's possible. They usually are a memory for the processor. It will contain some data that is read and then used by the program in the uP or stored and read at boot-up or re-boot. It would have had to be either preprogrammed or a special program loaded into the processor and then the final version loaded. It's possible .... Shorts on the data lines is unlikely to be causing the short you observed. So time back in the past we had a device that allowed you to inject a signal onto a pcb trace. You had a second device that allowed you to probe for that signal. I can't recall all the details but the purpose was to trace shorts without resorting to the cut and try method. Mark

gecoz · Answer

Jan Cumps wrote: ... It requires a bootloader. It won't work with a fresh µP From the Jye Tech Firmware Update documentation (the link provided in my original reply): The micro-controller (STM32F103C8T6) used in DSO SHELL contains a built-in booloader. This is a small piece of firmware that can communicate with PC via serial port to write firmware into the flash ROM inside the chip. This article describes how to setup the connection and how to perform firmware upgrade step by step. The fresh uP already comes with a bootloader, so in principle, if Jye Tech didn't replace it with a proprietary one, it is possible to flash a fresh uP using the tools and the procedure described in the doc. This is why I think contacting them could remove all the guesswork. Fabio.

gecoz · Answer

Hi John, I don't know this unit, but looking at the datasheet, the uP has 64KB flash memory, while the EEPROM has only 32Kbit (4KB), so my bet would be for the firmware to be stored on the uP, with the EEPROM storing the signal samples. If this is the case, I'll have to echo what Mark already said: the device will not boot if you simply replace the uP without preloading it with the firmware. If you got some time in your hands, as proof, you can still try dumping the content of the EEPROM via the I2C interface, and see if there is any meaningful data. This Application Note from MicroChip gives some information on how to interface a 8051 uP with the memory, which I believe can be easily adapted to any uP. If you have an Arduino board, this Add I2C EEPROM to Arduino gives a nice introduction on how to interface the EEPROM (it uses an 24LC256 chip,but I believe it still applies to the whole of 24LCXX family, so the info can be useful even for different setup). Fabio. P.S: JyeTech does provide a way to upgrade the firmware for the device, perhaps it can be also used to flash a new uP. It is worth tring to contact them and ask if that could work.

mcb1 · Answer

Not with an Arduino UNO because it's 5V output. Some of the Arduino have a 5v/3v3 switch. I use the itead studio one for development to allow interfacing to 3v3 parts without having the extra level converters. mark

jw0752 · Answer

Hi Jan,

As I was searching the internet for the shortest easier solution last night I remembered that I had a UART that I used a couple years ago to program an Arduino Mini Pro. Here are pictures of it:

As you can see it is just like the one you recommended without the cord. It is based on the Silabs CP2102.

NOTE: I have need of a clarification:

As long as we are talking I have a clarification question on these UARTs. Most of them have a way of choosing 5V or 3.3V. I am assuming that this is strictly for supplying power to the unit that is being programmed and has nothing to do with logic levels that the UART supplies. This one has 3.3V logic levels and it worked well with my 3.3 V STM 32 microprocessor. If however I was trying to use it on a chip that used 5V logic levels I might have a problem and would have to use a 3.3V to 5V logic converter. I have written this as statements but I need reassurance that I have the right idea about what I have said.

Thanks

John

Jan Cumps · Answer

Fabio Origlia wrote: ... P.S: JyeTech does provide a way to upgrade the firmware for the device, perhaps it can be also used to flash a new uP. It is worth tring to contact them and ask if that could work. It requires a bootloader. It won't work with a fresh µP

Jan Cumps · Answer

... it requires a USB to UART converter. They are available in the market (some mentioned in the guide). They can also be made with a RPi, BeagleBone, MSP43x, ... Not with an Arduino UNO because it's 5V output. That would require an extra level translator. edit: on some dev boards, there's no programming needed because the UART compatible signals from USB are directly accessible and can be decoupled from the on-board microcontroller I have an MSP432 and MSP430F board here with a 3V3 UART that can be tapped directly. The cheaper MSP430G board has RX and TX that can be tapped. That's also enough for the setup as shown in the upgrade guide.

dougw · Answer

Outstanding! Great intuition and curiosity, good sleuthing of the root cause, good research on potential solutions, including very healthy interaction with members, critical confidence to try your solution, and perfect success. It seems simple enough when reading about it, but how many would actually get it working? Great work.

Former Member · Answer

Yes I remember activating the uart on a beaglebone and then using minicom to flash the file over. I did try a USB to UART first but that never worked out for me. jw0752 If you feel confident enough to solder a new STM32 in place then its definitely worth a shot. If you do use a Raspberry Pi or similar to do the flashing, remember to add your user to the dialout group so you can access the serial ports. (and dont forget to remove the jumpers after programming )

Jan Cumps · Answer

Here's a USB to UART that would work from a PC: FTDI Cable

jc2048 · Answer

Here's another one that doesn't cost much - you even get a free, snap-off, PSoC 4 processor board with it! It's a CY8CKIT-049-42xx but the I/O is all 5V so it would need level conversion and some of the other suggestions are probably better

jw0752 · Answer

Hi Mark, The 3.3 V rail had less than 1 ohm resistance to ground. I removed all the easy stuff and didn't loose the short until I pulled the uP. It is possible that we mishandled it and it got hit with a bit of static. My friend thought that he might have damaged it by using the wrong power supply or by hooking it up backwards. I told him that there was a diode and a voltage regulator that would have protected it from these mistakes. Thanks for your ideas. John

Jan Cumps · Answer

You may be right, Lucie. Your explanation matches with the procedure they have in the upgrade guide . They certainly don't use a USB bootloader, it's via UART. And they mention: The micro-controller (STM32F103C8T6) used in DSO SHELL contains a built-in booloader. This is a small piece of firmware that can communicate with PC via serial port to write firmware into the flash ROM inside the chip. That would be good news.

gecoz · Answer

Hi John,

Your statement is absolutely right. UART devices can use different voltages for their logic levels, while still implementing the same communication protocol. In your case, the logic levels of the STM32 UART are 0V low, 3.3V high, so your USB to UART converter was OK. But if you were to use it to interface USB with a TTL UART device (like the Arduino UNO UART for instance), you would have needed a logic level converter, as TTL logic works with 0V low and 5V high levels.

Fabio.

Is This Situation Repairable

Top Replies