Searching \ for '[EE]Arm JTAG in circuit programming.' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/microchip/devprogs.htm?key=programming
Search entire site for: 'Arm JTAG in circuit programming.'.

Exact match. Not showing close matches.
PICList Thread
'[EE]Arm JTAG in circuit programming.'
2010\01\27@162207 by Funny NYPD

picon face
I have a question for ARM gurus who may already know the answer.
We are shifting from PICs to ARMs slowly. One thing I am still confused is:
For PIC'sICSP (In circuit Serial programming), there is only 5 ICSP signal required (Vpp, Vdd, Gnd, PGD, PGC), which makes it easily fit into a small circuits board (In the mean time, I can also use this 5 pin for ICD with ICD2 or PK2); however, all my ARM development kit and programmer requires a "2x10 header (20 pins)" for debugging and programming, 20 pins takes a lot of space on circuit board, not mention it also requires a 2x10 connector. So my questions is: is it possible to have just a few (e.g. 6) signal to program an ARM chip on board if debugging is not possible?


Funny N.
Au Group Electronics, http://www.AuElectronics.com
http://www.AuElectronics.com/products
http://augroups.blogspot.com/



     

2010\01\27@163145 by Bob Blick

face
flavicon
face

On Wed, 27 Jan 2010 13:21:25 -0800 (PST), "Funny NYPD"
<spam_OUTfunnynypdTakeThisOuTspamyahoo.com> said:
> I have a question for ARM gurus who may already know the answer.
> We are shifting from PICs to ARMs slowly. One thing I am still confused
> is:
> For PIC'sICSP (In circuit Serial programming), there is only 5 ICSP
> signal required (Vpp, Vdd, Gnd, PGD, PGC), which makes it easily fit into
> a small circuits board (In the mean time, I can also use this 5 pin for
> ICD with ICD2 or PK2); however, all my ARM development kit and programmer
> requires a "2x10 header (20 pins)" for debugging and programming, 20 pins
> takes a lot of space on circuit board, not mention it also requires a
> 2x10 connector. So my questions is: is it possible to have just a few
> (e.g. 6) signal to program an ARM chip on board if debugging is not
> possible?

The thing about JTAG is that it's a standard like OBDII, so there is a
lot of variation between manufacturers.

I know with TI MSP430 you don't need very many connections for the JTAG
programming. I forget how many, it's more than the PIC. I think seven
total.

Cheerful regards,

Bob

--
http://www.fastmail.fm - IMAP accessible web-mail

2010\01\27@164717 by Funny NYPD

picon face
Thanks bob,
I am wondering if someone has used the "in circuit programming pads (less than 20 pins)" design for ARM products? The real estate on PCB is always a challenge.

Funny N.
Au Group Electronics, http://www.AuElectronics.com
http://www.AuElectronics.com/products
http://augroups.blogspot.com/




________________________________
From: Bob Blick <.....bobblickKILLspamspam@spam@ftml.net>
To: Microcontroller discussion list - Public. <piclistspamKILLspammit.edu>
Sent: Wed, January 27, 2010 4:31:03 PM
Subject: Re: [EE]Arm JTAG in circuit programming.


On Wed, 27 Jan 2010 13:21:25 -0800 (PST), "Funny NYPD"
<.....funnynypdKILLspamspam.....yahoo.com> said:
{Quote hidden}

The thing about JTAG is that it's a standard like OBDII, so there is a
lot of variation between manufacturers.

I know with TI MSP430 you don't need very many connections for the JTAG
programming. I forget how many, it's more than the PIC. I think seven
total.

Cheerful regards,

Bob

--
http://www.fastmail.fm - IMAP accessible web-mail

2010\01\27@171030 by Wouter van Ooijen

face picon face
> is it possible to have just a few (e.g. 6) signal to program
> an ARM chip on board if debugging is not possible?

Most ARM chips have an on-board bootloader, so you can program (not
debug) with just GND, RX, TX, bootload enable, reset.

Newer ARM chips (Cortex only?) have IIRC a 1-wire debug interface, I
don't know the details of that one.

For the 20-pin box header: check the circuit of any board that has this
header: half the pins are ground. Remember PIC ICSP problems with
crosstalk from PGD to PGC? Some magic-value capacitor on the PGC often
solves it. For the ARM, with a 20-wire flatcable, there are lots of
in-between ground lines. Sensible design for medium-length cables.

--

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu

2010\01\27@172927 by Dave Tweed

face
flavicon
face
Wouter van Ooijen wrote:
> > is it possible to have just a few (e.g. 6) signal to program
> > an ARM chip on board if debugging is not possible?
>
> For the 20-pin box header: check the circuit of any board that has this
> header: half the pins are ground. Remember PIC ICSP problems with
> crosstalk from PGD to PGC? Some magic-value capacitor on the PGC often
> solves it. For the ARM, with a 20-wire flatcable, there are lots of
> in-between ground lines. Sensible design for medium-length cables.

Also, some manufacturers who use JTAG program/debug interfaces (not just
ARM -- I'm speaking more generally here) have standardized on a connector
that supports more features than you would need for accessing a single
chip. This sometimes involves breaking into a JTAG chain so that you have
exclusive access to just one chip at a time, and this requires extra pins
on the connector, which otherwise contains a special shorting block to
connect the chip to the rest of the on-board JTAG chain.

In any case, JTAG requires the same number of pins, regardless of whether
you're doing debugging or not. The difference between inward-looking
functionality (e.g., debugging, on-chip flash programming) and
outward-looking functionality (e.g., board test, off-chip flash
programming) is hidden inside the chip.

-- Dave Tweed

2010\01\27@190753 by Matt Pobursky

flavicon
face
I can tell you what we've done -- and this applies to every microcontroller
that we design in.

In the case of the ARM parts, I looked at numerous ARM based design and
development kit schematics, ARM chip datasheets and ARM JTAG connector
pinout from ARM themselves and built a "transition board" to plug in the 20
pin ARM JTAG cable and carry it's signals to another connector that I can
then make a small programming adapter cable to the specific connector used
on the PCB assembly. This allows us to use any connector on the PCB we
like, although for a given processor type we will use a standard
connector/transition cable/pinout. The target board cable/connector can be the
minimum number of pins that support the subset of the full JTAG pins your
microcontroller requires.

We always keep the transition cable very short, less than 6" and have had
zero problems with noise, crosstalk or any other signal related problems.
These adapters get used every day for production line use as well as
development use by several software guys so I would definitely hear of any
problems.

We have standardized on the Hirose DF3 (2mm) and DF13 (1.25mm) families of
connectors for most all of our PCB assemblies regardless of microcontroller
type. We have the tooling to terminate the pins (plus DigiKey sells pre-
terminated wires!) but JST and others make similar connectors which are
very inexpensive and small. They are inexpensive enough (less than $.10)
such that we almost never leave them off a production assembly.

Here is a quick snapshot of the ARM JTAG Adapter board schematic and PCB
that we use:

www.mps-design.com/misc-images/ARM_JTAG_Adapter_Schematic.gif
http://www.mps-design.com/misc-images/ARM_JTAG_Adapter_PCB.gif

We've used these on everything from simple ARM7 to ARM920 based designs
successfully. We do the same thing for MSP430 and PIC designs -- an adapter
board for the "standard" programmer pod cable (14 pin, dual row for the
MSP430 and the god-awful RJ jack that Microchip chose eons ago). All of our
MSP430 designs as well as PIC designs have either a 2mm or 2.5mm
programming/debug header and all use a standard (for us) pinout.

The major advantage to doing it this way is that any standard
debugger/programmer pod will plug into our adapter with it's standard
cable. That allows us to use tools from any manufacturer with no additional
changes to the setup.

Matt Pobursky
Maximum Performance Systems



On Wed, 27 Jan 2010 13:21:25 -0800 (PST), Funny NYPD wrote:
{Quote hidden}

2010\01\27@191111 by Xiaofan Chen

face picon face
On Thu, Jan 28, 2010 at 5:46 AM, Funny NYPD <EraseMEfunnynypdspam_OUTspamTakeThisOuTyahoo.com> wrote:
> Thanks bob,
> I am wondering if someone has used the "in circuit
> programming pads (less than 20 pins)" design for ARM products?
> The real estate on PCB is always a challenge.

Yes. You only need a few pins. We use board edge connector
(gold finger).

--
Xiaofan http://mcuee.blogspot.com

2010\01\27@191555 by Matt Pobursky

flavicon
face
That's correct -- you can possibly get by with 6 signals depending on your
design but 7 is safer since the one optional signal is your target board
voltage. Most MSP430 JTAG pods use this pin to set the I/O buffer voltage
based on the target voltage or optionally some pods cadn supply limited
target power. If you are running your design at (or very close to) the pod
voltage you can get by without it.

Our adapter includes this signal and also a jumper that switches the target
VDD to either the sense pin (target supplies voltage during
programming/debugging) or VDD supply pin from the JTAG pod (JTAG pod
supplies target power).

I almost never use pod power to power a target as most JTAG pod power is
not so great to start with, but the option is there.

Matt Pobursky
Maximum Performance Systems

On Wed, 27 Jan 2010 13:31:03 -0800, Bob Blick wrote:
{Quote hidden}

2010\01\28@114509 by Funny NYPD

picon face
Gold finger is definitely a good idea for 10~20 pins, do you have a good edge connector for this purpose?

Funny N.
Au Group Electronics, http://www.AuElectronics.com
http://www.AuElectronics.com/products
http://augroups.blogspot.com/




________________________________
From: Xiaofan Chen <@spam@xiaofancKILLspamspamgmail.com>
To: Microcontroller discussion list - Public. <KILLspampiclistKILLspamspammit.edu>
Sent: Wed, January 27, 2010 7:10:28 PM
Subject: Re: [EE]Arm JTAG in circuit programming.

On Thu, Jan 28, 2010 at 5:46 AM, Funny NYPD <RemoveMEfunnynypdTakeThisOuTspamyahoo.com> wrote:
> Thanks bob,
> I am wondering if someone has used the "in circuit
> programming pads (less than 20 pins)" design for ARM products?
> The real estate on PCB is always a challenge.

Yes. You only need a few pins. We use board edge connector
(gold finger).

--
Xiaofan http://mcuee.blogspot.com


'[EE]Arm JTAG in circuit programming.'
2010\02\01@090311 by Xiaofan Chen
face picon face
On Fri, Jan 29, 2010 at 12:44 AM, Funny NYPD <spamBeGonefunnynypdspamBeGonespamyahoo.com> wrote:
> Gold finger is definitely a good idea for 10~20 pins, do you have a good
> edge connector for this purpose?

We use Samtec female connectors. They have plenty of these type of connectors.
Then you need a small adapter PCB to get the lines connected to the 20pin
JTAG connectors. We just solder the wires there and then use some drops
of resin to secure the pins.

--
Xiaofan http://mcuee.blogspot.com

More... (looser matching)
- Last day of these posts
- In 2010 , 2011 only
- Today
- New search...