Re: [avr-gcc-list] Homebrew ICE?

[Neil Johnson]

Joerg,

> Yes, i know, and that would certainly be a nice solution for some
> prototypes.  OTOH, i hesitate to make this the official policy, say
> someone wants to produce 100 PCBs for it, and each of the hundred
> potential customers will have to request their own samples...

In cases like that I have contacted local distributors and negotiated
deals, where I get a number of interested people to pay up front for PCB +
special chips, and then do bulk orders, which saves everyone money.

Of course, if you're unable or unwilling to do that then it gets a bit
trickier.  For example, I bought a tube of Analog Devices chips from a
local distro, and sold them faster than I thought I would :-)  I'll have
to buy some more now, for my own projects!!

> >Also, when you consider the Maxim chip handles both directions, then it
> >would do the work of two 74HC chips, which would take up more board space
> >than this one Maxim chip.
>
> Board space would not be much of a concern for me (it's small anyway),

But if you're going to make 100 boards, then it does become an issue.  Not
to mention the larger number of solder joints, each of which is a
potential for trouble.

> and the bidirectional functionality is useless for the JTAG interface
> (even sRST which is a bidirectional line needs two separate level
> shifters).

True, but as you said in a later email, going bidirectional means you get
level conversion over a wider range of target voltages.  The Maxim design
is rather neat, using transmission gates as the level shifters.  Nice.

> I thought about using two unidirectional chips, one for the signals
> going from ICE to target, one for the opposite direction.

I'd still tend towards the single-chip solution.

Also I guess (from the PCB photos) you're using MAX232 for the RS232
(judging by the close proximity of the charge-pump capacitors).  The Maxim
MAX233 which doesn't require external caps (other than the usual
decoupling cap) would save you a few caps (i.e. less soldering and cost).

I think this is an interesting project indeed, and would like to see the
schematics when its done.

Cheers,
Neil

--
Neil Johnson :: Computer Laboratory :: University of Cambridge ::
http://www.njohnson.co.uk          http://www.cl.cam.ac.uk/~nej22
----  IEE Cambridge Branch: http://www.iee-cambridge.org.uk  ----