LOST... XTAG2 Not detecting my device

[jonecm]

Hi all, I have a little issue here. when I run xburn -l this is what I get

Available XMOS Devices
----------------------

ID Name Adapter ID Devices
-- ---- ---------- -------
0 XMOS XTAG-2 KUa3cTcL None

I have the device attached to a custom xmos board I had made and I am really stumped as to what the problem might be.
Here is how things our connected from the XTAG to the processor


TARGET : XTAG2

GND -> GND
RST_N+TRST_N -> RST_N
MSEL -> TRST_N (played around with this one allot including just grounding it etc...)
TMS -> TMS
TCK -> TCK
TDI -> TDSRC
TDO -> TDSNK

Voltages are all fine 1.01 and 3.29 and the pll supply is nice and smooth
Using a 20mhz clock

Is there something I'm missing?
Is it a mistake to join RST_N and TRST_N together?

Jon

[Bianco]

Hi jonecm, I assume you mean 3.29 instead of 2.29?
Are you using a G device?
Does xrun -l work or not?

[jonecm]

Thankyou Bianco for the quick reply!
Yes I ment 3.29, I've updated it now.
The device is a XS1-L1
when I do xrun -l it returns the same thing like xburn -l It lists the XTAG2 but shows there is no devices attached to it.
I put my logic analyzer on the signals from the XTAG2 I see a short pulse on the jtag clock and the TMS is pulsed twice but didn't see any data coming back from the processor.

Jon

[Bianco]

Can you perhaps share the schematic?

[jonecm]

Here is the schematic, This is the second XMOS board I have made that dosn't seem to work with XTAG, I have checked the Power On Reset timing and everything looks perfect.

[Bianco]

I don't think you have a proper power supply sequencing (3V3 must ramp first to within spec and after that 1V0 must ramp). What regulator chip do you use?

[jonecm]

TC1303A It has built in sequencing as well as open drain PG output (integrated 1.0 switching and 3.3v linear)

[Bianco]

Thats a nice regulator, just the datasheet doesn't mention in what sequence..
You are using a TC1303 which DOES NOT have power supply sequencing:

Sequenced Startup and Shutdown (TC1304)

This does not mean that there is no sequencing, you might want to check it with a scope.

With a 10K resistor on RST you are not able to pull the reset line down properly.
What you get is a potential divider.

Code: Select all

                [XMOS RESET INPUT]
                        |
                        |
GND---[10K]----------------------------[30K internal pull-up]----3V3

Lets assume that the internal pull-up is 30K and you pull it down with a 10K resistor with the PG open drain output, you will get 0.825V on the reset input. You can drop the resistor completely (replace it with 0 ohm or something very low).

Furthermore there is a risk that you short the reset output of the XTAG2 connector because it is directly wired to the PG output of the regulator. Say the power is not good and the PG port will sink to GND and the XTAG2 does not know this and pulls the reset high, then the XTAG2 reset output is shorted to GND.
Can be solved with buffers here and there.

Other things you might want to check: whether the oscillator works and the chip has a clock, preferrably measured on the XMOS device package pin; and whether the PLL AVDD is ~1V behind the resistor. It is very easy to make the mistake to put a 4K7 resistor there.

[jonecm]

the 10k is only a pullup to 3.3v to keep the thing from resetting under normal conditions, After I had boards made I realized that the device had internal pullups and I never populated it. The PG output from the TC1303A dose not go through a 10K resistor

Thanks for pointing out the thing about power sequencing, I thought I had the chip that did sequencing (Cruddy manual tries to cover too many different products) but I guess not, I did test though and it dose have enough sequencing to meet the requirements in the datasheet.

I did test for clock and it's a nice clean 20mhz square wave

I also replaced the L1 just to make sure it wasn't bad, Didn't see any reflow curves listed in the datasheets for this device so I just stuck with a 60 seconds at 260, Is that going to kill the chip?

What our your thoughts on the best way to buffer the RST lines so there not shorted by the XTAG?

Are there any drawback to combining the cpu Reset and the Jtag reset's together?

Well play with this one a bit more but I'm going to have more PCB's made and just want to make sure there done right the next time

Thanks for your help!

Jon

[Bianco]

Nevermind about the reset resistor, I was wrong, its just a regular pull-up indeed.
I will get back to you later (tomorrow).