questions about jtag/bdm

[threepointone]

I'm currently working on a DAP that will run rockbox, and I'm stuck at one important point in the development process: programming the coldfire chip. I've always been a PIC person, and I haven't ever gotten around to working with chips with JTAG/BDM. Hopefully, some of the developers can help me on this: can the coldfire mcf5250 be programmed via JTAG, or is it necessary to use BDM? If I must use BDM, has anyone ever attempted to actually construct the DIY Coldfire BDM interface (the one with the programmable logic IC)? My project is fairly simple and, at least for now, mostly based on the iRiver hardware, so I don't anticipate lots of low-level software debugging--thus, I'm not yet willing to go for the P&E bdm wiggler. Is it possible to use some of the 68k BDM interfaces with the coldfire chips, or are their protocols completely different?

Thanks for the help! I'll update on my project after it gets further along; if anyone's interested, it's an audiophile DAP that's in its planning stages at head-fi.

[LinusN]

No, I haven't tried to build my own wiggler. The 68k BDM variants are not compatible.

[threepointone]

so one last question: what about jtag? I know it definitely doesn't have the debugging capabilities, but that might be fine for my purposes.

Thanks again.

[threepointone]

I have very, very VERY good news!!! After almost two days scouring the Internet, I've found the solution!!!!!

http://forums.freescale.com/freescale/board/message?board.id=CFCOMM&thread.id=624

The one and only easy-to-build, <$10 BDM cable@!!#!!!!!!!!!
I admit, I haven't tried it / gone through all of the documentation yet, but this looks EXTREMELy promising, as almost everything is fully written out

I guessing that this means if anyone's permanently bricked their iriver/iaudio, it might actually be economically feasible to debrick it with a BDM!

[LinusN]

Looks really nice! Thanks for the tip!

[LinusN]

A small question: how will you modify the design to handle 1.8V levels?

[threepointone]

Sorry, I honestly haven't thought of that. I thought I/O was supposed to be 3.3V on the scf5250, while the core was 1.8V, but I assume they might have used the same voltage to remove the need for another regulator.

I'm using this BDM to make an entirely new DAP, and since the first version isn't exactly portable the I/O was going to be 3.3V, which isn't a problem for this programmer.

[edit]
I just checked the P&E website, and it seems that the coldfire bdm they have is also 3.3V. The SCF5250 also lists the minimum I/O voltage at 3.0V. I actually lost my h340 a while ago so I can't check it myself, but is the H3xx I/O voltage really 1.8V? Or is there some BDM line that needs to be at core voltage, 1.8V?

[edit again]
if it's necessary, maxim has an application note on this:
http://www.maxim-ic.com/appnotes.cfm/an_pk/3498
Maxim also has a bunch of ICs dedicated to logic level translation:
http://para.maxim-ic.com/cache/en/results/5043.html
TI and other manufacturers probably also have some, but I haven't gotten to check yet. If 1.8V interfacing turns out to be necessary, the logic IC (74VHC14) could be replaced by a single logic transceiver chip, simplifying and improving the design at the same time. In fact, I'll probably do that for my first BDM--I should be able to get things ready for actual testing in a week or two. If I have time, I'll post (preliminary) updated schematics sometime later this week

[LinusN]

My bad. You are right, the I/O is 3.3V on the 5250. I wonder what made me think otherwise. Maybe I'm getting old...