Re: gEDA-user: Announcement : gTAG schematics released

[Stefan Petersen <spe@stacken.kth.se>]

Hello Andrew,

On Mon, 5 Feb 2001, Andrew M. Dyer wrote:

> I looked over the schematics and I had a few comments in no particular
> order.  You may consider it overkill, and that's okay, they're just
> suggestions.

Yes!


> 1) I would consider opto isolating the jtag from the USB side to keep
> ground loops from doing strange things - you could conceivably have a
> loop between the gTAG board, target PCB, wall socket, wiring, PC and
> back to gTAG spanning many meters that can cause a lot of problems.
> This might be especially important in a production line environment.

Considered...rejected. Good point, in fact. I think we have to consider
thos as lab equipment. And if you have ground loops in the jtag
programming chains I think you have more severe problems in other places.


> 2) pull-up/pull-down the lines from the micro to the 7414 to keep them
> from floating and doing weird stuff before the micro gets around to
> initializing them as outputs

OK, good point. Yet another four resistors added.

> 3) consider re-ordering the pinout of the JTAG connector to allow use
> of the NC pin as a keying pin, so if somebody made a nice cable, you
> could be sure the (l)user would plug it in right.

To be able to mount the cable wrong now you simply have to dislocate it by
one step. If someone does that, they're worth it. It hurt to make
errors:-)

Seriously, you mean that I should put the NC pin in the middle of the
connector?


> 4) USB connector shows no connections for the metal shield

Since I am a lazy but I didn't draw any special usb_connector
symbol. Maybe I should do?


> 5) you might consider making the connectors into two-row versions to allow
> for ribbon cables.

Huh? Confused. The extension connector is two rows. The JTAG will remain
on row (Xilinx has that on their JTAG-programmers). USB looks like it
always looked. Missed anything?


> 6) maybe add some leds for the programmers on PA4/PA5

Blinken lights are for Stephen and he has said no.


> 7) I see no decoupling for the 7414

Ok yet another resistor added.


> 8) maybe add a two pin jumper on reset to allow shorting it to ground for
> development

Whats wrong with a screw driver:-)


> 9) current limiting resistors on the outputs of the 7414 might be a good
> idea and/or protection diodes.  Slew rate limiting caps might be good if you
> are using a higher speed variant of the 7414.

Creeping featurism. I'm lazy. Sorry.


> 10) you have the wakeup# pin on the micro pulled high - if you aren't
> planning on using suspend mode the docs say to hold the pin low.

"Holding WAKEUP# LOW inhibits the EZ-USB chip from suspending". As I read
it you can have it high, but if you go into suspend mode (programmatic
error?) you can't get out unless you pull wakeup# low. But if you pull it
low you can never reach suspend mode. Either the programmer always does
the right thing or I change the default...I change the default. :-)


> 11) what type of 7414 are you planning on using?  

Looking at 74HCT14 seems plausible (I cheated, see below ;-) ).


> If the 7414 is running off the jtag_power at a +5V rail will the 3.3V
> logic on the microcontroller be able to swing to a high enough voltage
> to trigger the schmitt trigger inputs?

From data sheet of 7c64603:
Voh Min 2.4V 
Ioh 1.6 mA

From data sheet of 74hct14 (assume 5.5V):
Vt+  Min 1.4V Max 2.1V
Vt-  Min 0.6V Max 1.4V

The micro driving the 74hct14 seems ok then. 

I first looked at the 74hc14 data. It has better voltage tolerance, but
then you can't guarantee the trigger levels. What you loose is that you
probably could have run the jtag chain on 3.3V.

But...if I add the pull up resistors (to 5 volt) you suggested between
micro and 7414, wouldn't they make the trick so you can use 74hc14 anyhow?


> Is the micro 5V tolerant on the input?  I haven't looked...  

Yes, according to the data sheets it should be.


> You might also consider using "open collector" outputs on the micro
> with pull-ups tied to jtag_power.

I added the 74xx14 because of:
1) schmitt trigger.
2) protective


Gee, talk about keeping me on the toes. This was really fun. Hope you
don't get angry on me for rejecting some of your ideas with not so good
rejection motivations from my side.


/----------------------------------\
! Stefan Petersen, MSc EE         !  \    
! http://www.stacken.kth.se/~spe/ !    \_________________________
! spe@stacken.kth.se              !    /  ! umop apisdn 'sdoo !
! stefan.petersen@home.se         !  /    ---------------------
\----------------------------------/