AVR-Chat

I'm looking for a reasonably inexpensive way to program a bootloader 
into some ATmega324p chips in a TQFP package before I put them on 
boards since I would prefer not to put a programming adapter on the 
boards. All the sockets I've seen seem to be in the $100+ range.  
Anything out there for doing small volumes?

Thanks,

Bob

--- In AVR-Chat@yahoogroups.com, "Bob Gardner" <bob.linda.gardner@...> 
wrote:
> All the sockets I've seen seem to be in the $100+ range.  
I've purchased several TQFP ZIF sockets from http://test-socket.com, 
mostly the Yamaichi brand.  They run about $70 and seem to be good 
quality.

Don Kinzer
ZBasic Microcontrollers
http://www.zbasic.net

On Nov 23, 2007 12:27 PM, Bob Gardner <bob.linda.gardner@verizon.net> wrote:
> I'm looking for a reasonably inexpensive way to program a bootloader
> into some ATmega324p chips in a TQFP package before I put them on
> boards since I would prefer not to put a programming adapter on the
> boards. All the sockets I've seen seem to be in the $100+ range.
> Anything out there for doing small volumes?

The STK-501 was pretty amazing, I remember looking up the TQFP ZIF
socket and finding a $200 price tag.

--- Bob Gardner <bob.linda.gardner@verizon.net> wrote:

> I'm looking for a reasonably inexpensive way to
> program a bootloader 
> into some ATmega324p chips in a TQFP package before
> I put them on 
> boards since I would prefer not to put a programming
> adapter on the 
> boards. All the sockets I've seen seem to be in the
> $100+ range.  
> Anything out there for doing small volumes?
> 
> Thanks,
> 
> Bob

hi;

seems to be a stupid idea, but you may make a pcb for
this TQFP package and place IC on it without soldering
by forcing IC into it using something like a clips or
something, sorry i dont know the exact word of this
<clips> thing in english.

i think it may cost less than 5$

good luck;



      ____________________________________________________________________________________
Never miss a thing.  Make Yahoo your home page. 
http://www.yahoo.com/r/hs

Thanks for the idea - it's not be a stupid idea at all  I'll make a 
board in the next couple of days and try it out. I'll let you know 
how it works.

Thanks again,

Bob


--- In AVR-Chat@yahoogroups.com, Reza <reza_agha@...> wrote:
>
> 
> --- Bob Gardner <bob.linda.gardner@...> wrote:
> 
> > I'm looking for a reasonably inexpensive way to
> > program a bootloader 
> > into some ATmega324p chips in a TQFP package before
> > I put them on 
> > boards since I would prefer not to put a programming
> > adapter on the 
> > boards. All the sockets I've seen seem to be in the
> > $100+ range.  
> > Anything out there for doing small volumes?
> > 
> > Thanks,
> > 
> > Bob
> 
> hi;
> 
> seems to be a stupid idea, but you may make a pcb for
> this TQFP package and place IC on it without soldering
> by forcing IC into it using something like a clips or
> something, sorry i dont know the exact word of this
> <clips> thing in english.
> 
> i think it may cost less than 5$
> 
> good luck;
> 
> 
> 
>       
______________________________________________________________________
______________

> Never miss a thing.  Make Yahoo your home page. 
> http://www.yahoo.com/r/hs
>

Being a test engineer I always cringe when I see a designer lock out 
the ability to upgrade the software, I know you are just programming 
the bootloader and the application gets downloaded. I can't tell you 
how many times engineers have come to me and ask me to fix their BRICK 
after they accidentally overwrote the bootloader.
At least make some test points for either the ISP or JTAG pins then 
you can use a small pogo pin fixture to re-program the device if 
necessary, it takes less than 6 pins for ISP and 8 for JTAG. There are 
some small spring loaded connectors that also could be used, 6 test 
points at .05 spacing won't take up too much space.
Just out of curiosity is this for space or security reasons?
Dyan….

--- In AVR-Chat@yahoogroups.com, "Dyan Fassett" <kv9r@...> wrote:

> ... I can't tell you how many times
> engineers have come to me and ask me
> to fix their BRICK after they ...
> overwrote the bootloader. ...

Using the lock bits, you can make it impossible to overwite the 
bootloader.  That is to say, impossible for the bootloader to overwrite 
itself, even if it misbehaves.  I think that putting a bootloader 
(which is actually a Flash loader, not a bootloader in the proper 
sense) in the microcontroller as the only firmware load facility is 
reasonable.

Graham.

--- In AVR-Chat@yahoogroups.com, "Graham Davies" <Yahoo37849@...> 
wrote:
>
> --- In AVR-Chat@yahoogroups.com, "Dyan Fassett" <kv9r@> wrote:
> 
> > ... I can't tell you how many times
> > engineers have come to me and ask me
> > to fix their BRICK after they ...
> > overwrote the bootloader. ...
> 
> Using the lock bits, you can make it impossible to overwite the 
> bootloader.  That is to say, impossible for the bootloader to 
overwrite 
> itself, even if it misbehaves.  I think that putting a bootloader 
> (which is actually a Flash loader, not a bootloader in the proper 
> sense) in the microcontroller as the only firmware load facility 
is 
> reasonable.
> 
> Graham.
>
As long as you remember to program it before you solder it on and 
create a BRICK.
As long as you don't need to change the loader code.
As long as you don't need to debug the application with JTAG. 
All I'm saying is think of all the posibilities that can occur 
before you paint yourself into a corner.
In the last 35 years as a test engineer I've seen a lot of designers 
waiting for the paint to dry. :)
Dyan...

--- In AVR-Chat@yahoogroups.com, "David VanHorn" <microbrix@...> 
wrote:
>
> > As long as you remember to program it before you solder it on and
> > create a BRICK.
> > As long as you don't need to change the loader code.
> > As long as you don't need to debug the application with JTAG.
> > All I'm saying is think of all the posibilities that can occur
> > before you paint yourself into a corner.
> 
> 
> There are millions and millions of credit card terminals out there,
> which load their applications into SRAM, using a loader in ROM.
> From day 1, somewhere back in the mists of 1983, that loader has 
not
> changed.  In this case it's not physically possible to change the
> loader other than removing the ROM.  The user presses one button,
> optionally enters a phone number and merchant ID, and it dials out 
and
> gets new application code.  The loader has to manage a modem, phone
> line interface, scanned keyboard, and VFD display.
> 
> The loader needs to be exactly as simple as it can be, and not any 
simpler.
> 
> I never understood why PC motherboards with their vastly larger
> resources, couldn't get such a simple thing right.
>
Of course loaders work...... That is not the point....
All I'm saying is that if the PCB has NO way of programming the 
device once soldered... and you mount un-programmed devices you have 
a BRICK. 

Dyan...

> As long as you remember to program it before you solder it on and
> create a BRICK.
> As long as you don't need to change the loader code.
> As long as you don't need to debug the application with JTAG.
> All I'm saying is think of all the posibilities that can occur
> before you paint yourself into a corner.


There are millions and millions of credit card terminals out there,
which load their applications into SRAM, using a loader in ROM.
From day 1, somewhere back in the mists of 1983, that loader has not
changed.  In this case it's not physically possible to change the
loader other than removing the ROM.  The user presses one button,
optionally enters a phone number and merchant ID, and it dials out and
gets new application code.  The loader has to manage a modem, phone
line interface, scanned keyboard, and VFD display.

The loader needs to be exactly as simple as it can be, and not any simpler.

I never understood why PC motherboards with their vastly larger
resources, couldn't get such a simple thing right.

--- In AVR-Chat@yahoogroups.com, Reza <reza_agha@...> wrote:
>
> seems to be a stupid idea, but you may make a pcb for
> this TQFP package and place IC on it without soldering
> by forcing IC into it using something like a clips or
> something, sorry i dont know the exact word of this
> <clips> thing in english.

For the clip/clamp, you might want to check out destaco clamps: 
http://www.destacoclamp.com/

Available @ McMaster-Carr (www.mcmaster.com)

-Thor

--- In AVR-Chat@yahoogroups.com, "Dyan Fassett" <kv9r@...> wrote:

> All I'm saying is think of all the posibilities that
> can occur before you paint yourself into a corner.

Ok, so, what we have is this ... if you don't provide a means of 
directly programming the chip (JTAG, ISP, etc.) in a product and 
instead rely on a Flash loader to upgrade the application firmware:

1 - You must make sure the Flash loader is in the chip before it is 
soldered to the board.

2 - You must make sure that the Flash loader is working properly and 
does everything you need it to.

3 - You must have a special build of the product with an OCD 
interface in order to use on-chip debug for application development.

I think we can all agree with this.  So, if the OP has all these 
ducks in a row, we shouldn't berate him.  On the other hand, we 
should heed the voice of experience when you remind us that we may 
think we have our ducks in a row but one of them turns out to be 
lame.  Finally, mixing metaphors is not such a good idea after all.

Graham.

As the OP here, I appreciate all of the responses.  My original 
intent was just to save a little space on what I hope to be a fairly 
small board and I have to admit I hadn't really thought through some 
of the possibilities that evryone has brought up.  I don't have any 
problems with providing test points for a 6 pin ISP adapter of some 
type.

Along those lines I have done some internet searching but didn't come 
up with anything in particular.  Can anyone provide more detail or 
even some links for some type of narrow pitch adapter I could make or 
buy to use with some test points I would put on my board?

Thanks again,

Bob


--- In AVR-Chat@yahoogroups.com, "Graham Davies" <Yahoo37849@...> 
wrote:
>
> --- In AVR-Chat@yahoogroups.com, "Dyan Fassett" <kv9r@> wrote:
> 
> > All I'm saying is think of all the posibilities that
> > can occur before you paint yourself into a corner.
> 
> Ok, so, what we have is this ... if you don't provide a means of 
> directly programming the chip (JTAG, ISP, etc.) in a product and 
> instead rely on a Flash loader to upgrade the application firmware:
> 
> 1 - You must make sure the Flash loader is in the chip before it is 
> soldered to the board.
> 
> 2 - You must make sure that the Flash loader is working properly 
and 

> does everything you need it to.
> 
> 3 - You must have a special build of the product with an OCD 
> interface in order to use on-chip debug for application development.
> 
> I think we can all agree with this.  So, if the OP has all these 
> ducks in a row, we shouldn't berate him.  On the other hand, we 
> should heed the voice of experience when you remind us that we may 
> think we have our ducks in a row but one of them turns out to be 
> lame.  Finally, mixing metaphors is not such a good idea after all.
> 
> Graham.
>

You could use pogo pins like these
http://www.mill-max.com/images/products/pdf/121.pdf

I would use this part  0922-0-15-20-75-14-11-0 it is available at
mouser as PN 575-922015 for $0.48 each. 
http://www.mouser.com/search/ProductDetail.aspx?R=0922-0-15-20-75-14-
11-0virtualkey57510000virtualkey575-922015

You would only need a miximum of 10 to even do JTAG debug.

These could be mounted in a proto board and have your UUT push down 
and contact the pogo pins. These pins do require 100 mil spacing.
But all you would need to do is add test pads where you could on the 
bottom or top depending on which is easier to accomplish.

We make this type of adaptor for several CCA's with very good results.

Dyan...

--- In AVR-Chat@yahoogroups.com, "Bob Gardner" <bob.linda.gardner@...> 
wrote:
>
> As the OP here, I appreciate all of the responses.  My original 
> intent was just to save a little space on what I hope to be a fairly 
> small board...

I once had a similar decision to make and chose to provide 6 vias 
spaced 1mm apart and use a 6 pin 1mm Header to connect them for 
programming. I just plug the lins into the holes and cant the header to 
ensure contact. Worked for several hundred boards up to now without 
problem.

You can see the board at one of my my webpages 
(http://www.geocities.com/_wsw_) when you click at "AVR based" 
in "Electronics" and then scroll down to what I call the "TinyTimer".

Stefan