JTAG multiple devices

Greetings,
  I am new to JTAG and and I am working on a project that that has an
Atmel AVR.  I want to be able to use JTAG but have not bought any
programming/debugging devices yet.  I would like to get something that
is compatible with other products that I see myself using in the
future such as FPGAs or other microcontroller families.  I am on a
tight budget and I want to get something that will work with multiple
devices.

 Additionally, if all the chips that I use adhere to the IEEE
standard, one JTAG device should do, right? I may be missing a major
point and do not understand why is there a plethora of JTAG devices
that are specific to a certain hardware or software.

Any help is greatly appreciated.

Thank you,

--A

"amerdsp" <amerdsp@hotmail.com> wrote in message 
news:1187943709.796145.237760@i38g2000prf.googlegroups.com...
> Greetings,
>  I am new to JTAG and and I am working on a project that that has an
> Atmel AVR.  I want to be able to use JTAG but have not bought any
> programming/debugging devices yet.  I would like to get something that
> is compatible with other products that I see myself using in the
> future such as FPGAs or other microcontroller families.  I am on a
> tight budget and I want to get something that will work with multiple
> devices.
>
> Additionally, if all the chips that I use adhere to the IEEE
> standard, one JTAG device should do, right? I may be missing a major
> point and do not understand why is there a plethora of JTAG devices
> that are specific to a certain hardware or software.
>
> Any help is greatly appreciated.
>
> Thank you,
>
> --A
>
Not a  chance.... Each vendor will have their own software implementation as 
well as their own hardware interface. The only commonality will be the names 
of the signals at the JTAG end.... 

> I may be missing a major
> point and do not understand why is there a plethora of JTAG devices
> that are specific to a certain hardware or software.

 Because many silicon vendors keep JTAG accessible hardware on their
chips which goes beyond boundary scan secret; then, they give the
data some "firm" which is allowed to collect a few thousands per
sold cable - or do it themselves. The cash they collect on that
is negligible, but once you give in they have you under control.
 An astonishingly low number of people seem to object paying
ransoms like that.
 If you can do all you need via boundary scan only, though,
there should be some cheap way to do it - open source or
whatever. (I cannot name any because the tols I use have been
written here and run under DPS which you don't have).

Dimiter

------------------------------------------------------
Dimiter Popoff               Transgalactic Instruments

http://www.tgi-sci.com
------------------------------------------------------


On Aug 24, 11:21 am, amerdsp <amer...@hotmail.com> wrote:
> Greetings,
>   I am new to JTAG and and I am working on a project that that has an
> Atmel AVR.  I want to be able to use JTAG but have not bought any
> programming/debugging devices yet.  I would like to get something that
> is compatible with other products that I see myself using in the
> future such as FPGAs or other microcontroller families.  I am on a
> tight budget and I want to get something that will work with multiple
> devices.
>
>  Additionally, if all the chips that I use adhere to the IEEE
> standard, one JTAG device should do, right? I may be missing a major
> point and do not understand why is there a plethora of JTAG devices
> that are specific to a certain hardware or software.
>
> Any help is greatly appreciated.
>
> Thank you,
>
> --A

On Aug 24, 5:48 am, Didi <d...@tgi-sci.com> wrote:
> > I may be missing a major
> > point and do not understand why is there a plethora of JTAG devices
> > that are specific to a certain hardware or software.
>
>  Because many silicon vendors keep JTAG accessible hardware on their
> chips which goes beyond boundary scan secret; then, they give the
> data some "firm" which is allowed to collect a few thousands per
> sold cable - or do it themselves. The cash they collect on that
> is negligible, but once you give in they have you under control.
>  An astonishingly low number of people seem to object paying
> ransoms like that.

Yes, we object. Especially when we need a few hundred production
testers and programmers.  We don't need stinking Window Point and
Click either (make no sense in production environments),  just connect
and power it up.  We spent hundreds of hours on development (to save
hundreds of hours in production), but it would still be cheaper
overall.  We make the custom device for less than $10 each.

>  If you can do all you need via boundary scan only, though,
> there should be some cheap way to do it - open source or
> whatever. (I cannot name any because the tols I use have been
> written here and run under DPS which you don't have).

Following Atmel specs, we could SPI program the AVR and JTAG boundary
scan them.  We are unable to do JTAG programming as stated in the AVR
specs.  So, we end up with both interfaces on every production board.

> ......
> >  An astonishingly low number of people seem to object paying
> > ransoms like that.
>
> Yes, we object. Especially when we need a few hundred production
> testers and programmers.  We don't need stinking Window Point and
> Click either (make no sense in production environments), ....
> ....

Hmm, I did hope you would come in with this sort of a message :-).

I know I am not alone in that - but I think my "astonishingly low"
still applies...

Dimiter

------------------------------------------------------
Dimiter Popoff               Transgalactic Instruments

http://www.tgi-sci.com
------------------------------------------------------

On Aug 24, 8:56 pm, linnix <m...@linnix.info-for.us> wrote:
> On Aug 24, 5:48 am, Didi <d...@tgi-sci.com> wrote:
>
> > > I may be missing a major
> > > point and do not understand why is there a plethora of JTAG devices
> > > that are specific to a certain hardware or software.
>
> >  Because many silicon vendors keep JTAG accessible hardware on their
> > chips which goes beyond boundary scan secret; then, they give the
> > data some "firm" which is allowed to collect a few thousands per
> > sold cable - or do it themselves. The cash they collect on that
> > is negligible, but once you give in they have you under control.
> >  An astonishingly low number of people seem to object paying
> > ransoms like that.
>
> Yes, we object. Especially when we need a few hundred production
> testers and programmers.  We don't need stinking Window Point and
> Click either (make no sense in production environments),  just connect
> and power it up.  We spent hundreds of hours on development (to save
> hundreds of hours in production), but it would still be cheaper
> overall.  We make the custom device for less than $10 each.
>
> >  If you can do all you need via boundary scan only, though,
> > there should be some cheap way to do it - open source or
> > whatever. (I cannot name any because the tols I use have been
> > written here and run under DPS which you don't have).
>
> Following Atmel specs, we could SPI program the AVR and JTAG boundary
> scan them.  We are unable to do JTAG programming as stated in the AVR
> specs.  So, we end up with both interfaces on every production board.

Ok, so if you have different components on your board sharing the same
JTAG connector, you have to have different JTAG devices as well?  And
you connect which ever JTAG device that corresponds to the chip you
are testing?  Is this how it goes?

On Aug 24, 1:18 pm, amerdsp <amer...@hotmail.com> wrote:
> Ok, so if you have different components on your board sharing the same
> JTAG connector, you have to have different JTAG devices as well?  And
> you connect which ever JTAG device that corresponds to the chip you
> are testing?  Is this how it goes?

Yes, we poll the Jtag/Isp IDs and decide how to test/program the board
automatically.

On Fri, 24 Aug 2007 10:56:58 -0700, linnix <me@linnix.info-for.us>
wrote:

[Snipped]

>Following Atmel specs, we could SPI program the AVR and JTAG boundary
>scan them.  We are unable to do JTAG programming as stated in the AVR
>specs.  So, we end up with both interfaces on every production board.

I have had no problem programming Atmel devices using the JTAG
interface. I use an altera byteblaster and my own software.
AFAIK there is some sort of issue if the AVR is not the only device on
the chain, but IIRC it stops one accessing the other devices, and not
the AVR itself.

Regards
  Anton Erasmus

On Aug 25, 7:29 pm, Anton Erasmus <nob...@spam.prevent.net> wrote:
> On Fri, 24 Aug 2007 10:56:58 -0700, linnix <m...@linnix.info-for.us>
> wrote:
>
> [Snipped]
>
> >Following Atmel specs, we could SPI program the AVR and JTAG boundary
> >scan them.  We are unable to do JTAG programming as stated in the AVR
> >specs.  So, we end up with both interfaces on every production board.
>
> I have had no problem programming Atmel devices using the JTAG
> interface. I use an altera byteblaster and my own software.

Good to know that it could work.  I might get back to it later.   For
now, ISP works fine for us.  However, it would simplify our interfaces
if we can do everything in JTAG.

> AFAIK there is some sort of issue if the AVR is not the only device on
> the chain, but IIRC it stops one accessing the other devices, and not
> the AVR itself.
>
> Regards
>   Anton Erasmus