Failure of 68HC908KX8 Monitor Mode

Does your application alter the vectors from 0xf6 to 0xfd? These are also used as security bytes that have to be sent to enable the user flash memory. I am not familiar with the development software you're using: I wrote my own monitor mode programming software that runs on gnu/linux and FreeBSD. My software has a command line switch to load an eight byte security file. How does the software you're using handle the security bytes?

Ian

-- Ian Stirling, G4ICV, AB2GR

oops .. I meant 0xfff6 to 0xfffd

Ian

-- Ian Stirling, G4ICV, AB2GR

Hi Ian

Thank you for the post. The problem I am experiencing occurs before one sends the security bytes to the KX8. When using the MON08 interface in conjunction with P&E software prog08sz, the software executes a series of steps to ascertain whether the KX8 can be accessed.

Step 1 is to check for hardware loopback on the serial channel. Step

In my case the second step fails, i.e. I see only 1 character being echoed due to hardware loopback. Seems like the KX8 has turned deaf - something aint right...

Regards Jack

Hi Jack,

I read your original post again and in it you said that you erased the KX8 after putting the AN2295 software in it. Erasing is a monitor function and is possible without the correct security code being sent. When the security block is written, reset isn't sufficient to recognise the new security data sent by the host - the hc908 has to be power cycled.

Ian

-- Ian Stirling, G4ICV, AB2GR

Thanks for your last response Ian.

Your first statement is correct. However, I suspect that I will not be able to communicate with the KX8 since I do not see any bytes echoed by the KX8 itself when in monitor mode. In other words, it does not seem as if the monitor mode code (which is in ROM) is able to echo characters, let alone accept a Flash erase command.

As far as your second statement is concerned, I have tried communicating with the KX8 multiple times, after having power cycled it several times. The SAME circuit operates flawlessly with another virgin KX8 device.

After I programmed the KX8, the controlling/programming software IMMEDIATELY lost control of the device. It seemed as if the flash program started running right after programming it and took control from the monitor mode even though the monitor mode setup pins were still set up to force the device into monitor mode.

I had the following code in main():

void main(void) {

EnableInterrupts; DDRA_DDRA = 0x1f; // configure as output DDRB_DDRB = 0x00; // configure as input

for(;;) {

__RESET_WATCHDOG(); PTA_PTA0 = PTB_PTB0; // light LED on PTA0 when button on PTB0 is pressed

} }

Now keep in mind that, after having been programmed, the KX8 would still be installed in the MON08 circuit wherein the serial line from the MAX232 would be connected to PTA0. I have configured PTA as an output in my code. Does a program start to run as soon as one has programmed it, even though control pins is supposed to force it into monitor mode? If yes, could it be that the PTA0 line was destroyed because, configured as an output by the flash code, the PTA0 line would have driven into the output from the MAX232 line?

Apologies for these communications, but this is my first attempt at

Regards Jack

Hi Jack,

It is unlikely that PTA0 is destroyed - it is never connected to the MAX232 RS232 lines. On page 258 of the KX8 manual, there is a diagram of a monitor mode circuit. PTB0 should be tied high, PTA1 low, PTB1 low and IRQ1 at the Vtst voltage. These are all necessary conditions with Motorola's KX8 example circuits to enter monitor mode when the reset vector is not 0xffff. When the reset vector is 0xffff, the conditions of PTB0 and PTB1 do not matter (page 257), and you appear to have no problem programming a blank KX8.

Having your test application use the monitor mode configuration pins might be a problem - e.g., is PTB0 high on attempting to go into monitor mode with a non blank reset vector?

With monitor mode conditions applied, user code will not run. I find it easier to use my board as a monitor mode programming board and move the device to another circuit to test application code.

Ian

-- Ian Stirling, G4ICV, AB2GR

Be advised that there are some more-specific Freescale HC08 NGs:

There's also a brand-new NG that leans a bit toward Freescale:

They should be able to help you too.

Gary

Thank you Gary - I will follow up. Ian, I have triple-checked my MON08 circuit to see if the pins are set up correctly for normal monitor mode entry, and all seem fine. I will now test my IC in a target test circuit to see if it actually runs off the flash code. Also, I am in constant communication with a support guy at Freescale who is also in the process of helping me sort out this issue. I will advise as soon as I am able to resolve this. Regards Jack

I cannot offer any direct advice on your problem. My involvement with the HC08/HC12 is redoing Valvano's HC11/HC12 Using C with Hiware, etc. tutorial by adding some CodeWarrior documentation to it plus some of my own rewriting. It's about 1/2 redone at this point.

Gary

Jack, I'm the owner of the MicrocontrollerWorld gropu. There is a guy in my group called Iceman that works a lot with the 908 and probably knows exactly what the problem is. Either try contacting him through the group or just email me and I'll get you his email address after I talk to him. I want to respect him if he wants his privacy. But he's definately the guy.

Michael

Hi Jack,

I look forward to your resolving the problem and I'm interested in what the problem is: I am out of ideas now. My own board with sockets wired up for monitor mode with the GP32, GR8, JL3, JK3, QT, QY, KX8 and JB8, all 0.1 inch pitch PDIP versions, works faultlessly with my own gnu/linux and FreeBSD software. The software identifies the processor on board, except for the JB8 so far because I haven't added a 6 MHz clock to the board yet. When I do that and write the documentation, I'm going to put all the source code on my web site under the GPL. My MC68HC705C8A programming software seems to be still very popular three years after I released it. Given that that processor is obscure, my 908 software should be even more popular. I just wish more than two people out of thousands who downloaded the software would have contacted me.

Best Wishes, Ian

-- Ian Stirling, G4ICV, AB2GR

Hi Ian According to Freescale it might be the absence of the 74HC125 buffers that might be the issue (I am using wired OR on the PTA0 line), as well as possibly timing between the reset pulse and the application of Vtst on IRQ line...I will check it out over the weekend and advise. Regards Jack

Hi Jack,

My board is based on the "Simpler Monitor Mode Circuit", figure 5 in AN2317. It uses a diode instead of the 74HC125. The rest of the wiring is from the individual processors' manuals for configuring specific pins. I made a separate MAX232 board that connects to the board with five wires, the fifth to give the MAX232 V+ voltage for the Vtst voltage. It's all described in AN2317.

Ian

-- Ian Stirling, G4ICV, AB2GR