Lpc2000

I am experiencing some problems with chaining an LPC2138 and LPC2106 on 
JTAG.  The BDI2000 intermittantly has trouble synchronizing the chain 
and takes several attempts before it "locks".  I am wondering about the 
way that I did the chaining and welcome any comments.

What I did was to take the TCK, TMS, TRST & RESET signals and ran them 
from the JTAG header to each processor (in parallel).  The data signals 
I ran in a daisy chain:
    JTAG.tdo --> LPC2138.tdi
    LPC2138.tdo --> LPC2106.tdi
    LPC2106.tdo --> JTAG.tdi

This does work, but about 20% of the time I get synchronization problems 
where the BDI2000 reports a problem reading the ICE BREAKER registers.  
After 2 to 5 successive retries it will link.  In severe situations, I 
have to power cycle the target board to achieve a link.

Once linked, it stays linked.  Rarely (I don't recall but may have 
happened) while debugging will it lose the synchronization and have to 
re-initialize.

Has anyone experienced this?  Is the wiring of multiple JTAG ports as I 
described?  Do I, or should I, consider using the RTCK signal?

Regards,

TomW


-- 
Tom Walsh - WN3L - Embedded Systems Consultant
http://openhardware.net, http://cyberiansoftware.com
"Windows? No thanks, I have work to do..."
----------------------------------------------------

> Tom Walsh:

> Has anyone experienced this?  Is the wiring of multiple JTAG ports as
> I described?  Do I, or should I, consider using the RTCK signal?

isn't the RTCK signal *required* on ARM7TDMI-S cores?

  clemens

clemens fischer wrote:

>>Tom Walsh:
>>    
>>
>
>  
>
>>Has anyone experienced this?  Is the wiring of multiple JTAG ports as
>>I described?  Do I, or should I, consider using the RTCK signal?
>>    
>>
>
>isn't the RTCK signal *required* on ARM7TDMI-S cores?
>
>  
>
No.  Only if you run adaptive clocking do you need the RTCK signal.  
But, if you run at a fixed clocking rate you don't need it.  I do think 
that this may be something I should consider, using adaptive clocking?

My problem is that I have the TCK set to the slowest possible setting 
and I have occassional problems synchronizing.  When the problem occurs, 
it can take 10 or more attempts to synchronize, usually 2..4 attempts.  
Something else I need to try, again, is terminating the signals at the 
far end of the chain.

I finally did get an email reply from Ultimate Solutions (U.S. 
distributor for Abatron) about the daisy chain wiring.  The wiring is 
correct: all signals are parallel with the exception of the obvious, TDI 
+ TDO signals.  I can appreciate that since I didn't spend $4K for thier 
ARM development package, or spend $800 for the telephone support.  They 
seem to be slow at paying attention to email support. :(

A nice product, but the support is not what I expected for the price.

TomW

-- 
Tom Walsh - WN3L - Embedded Systems Consultant
http://openhardware.net, http://cyberiansoftware.com
"Windows? No thanks, I have work to do..."
----------------------------------------------------