cancel
Showing results for 
Search instead for 
Did you mean: 

Problems with JTAG

anm
Associate II
Posted on May 14, 2009 at 09:12

Problems with JTAG

This discussion is locked. Please start a new topic to ask your question.
39 REPLIES 39
anm
Associate II
Posted on May 17, 2011 at 13:10

Quote:

On 11-05-2009 at 16:03, Anonymous wrote:

Sorry - since we can't ''out-think'' this problem time for ''brute force.''

ok ;)

Quote:

a) Remove the main xtal from the working Dev Board. test to see if Jtag works without this xtal installed - report.

Yes, JTAG works without the XTAL installed on the devboard.

Quote:

b) If Jtag fails w/o xtal (on your Dev Board) remove the xtal from your custom Bd and install it on the Dev Board. Test and see if this xtal works on the Dev Board. (this eliminates the main xtal as your problem)

c) If Jtag fails w/o xtal - and if works when you install the xtal from your custom board - remove each of the osc caps from the Dev Bd and replace with those from the custom board. Test & report. (this eliminates the xtal caps as your problem)

So it works without xtal, it should not depend on this.

Quote:

d) After a-c - employ the same, remove and ''swap'' procedure with all components used on reset circuit of the STM32.

There is not many on the reset circuit. I changed the parts - same problem occurs.

Quote:

e) should a-d ''not'' solve this - try to ''disrupt'' Jtag opn. on your Dev Bd by connecting your scope probe to each Jtag and reset pin prior to power on. (we seek to find ''if and where'' there are ''Jtag pin sensitivites.'')

I don't understand what you mean exactly. I checked all JTAG Pins of the custom board and the devboard again - without any changes.

Quote:

(I would do some of this for you but am on the road & have trade show end of this week...)

It's not your fault - it's my problem, so I have to solve it. What you do here is very great! Thanks!

16-32micros
Associate III
Posted on May 17, 2011 at 13:10

Hi anm,

May this is coming for you Jlink, could you check and download the latest drivers at Segger web site :

http://www.segger.com/pub/jlink/Setup_JLinkARM_V404a.zip

After correct install , execute ''jlink.exe'' in MS-DOS to check your connection.

Hope this will help :)

Cheers,

STOne-32.

jj
Associate II
Posted on May 17, 2011 at 13:10

Hi -

Thanks for trying ''brute-force'' and reporting. Your report will benefit those of us who encounter this in the future.

I like STOne-32's suggestion.

After trying STOne-32's - how about trying a, ''non-JLink'' JTAG? Sorry it took so long for me to think of this. ALL testing todate involves your JLink - as STOne-32 suggests we don't know if this is a JTAG or JLINK problem. Perhaps you have a Keil or other form of Jtag device? It would be extremely useful if we knew whether this was a Jtag or JLink issue.

Keep trying - let us know - good luck. (Did you tell us the exact STM32 and Rev. which is on your custom bd? - And is the micro & Rev ''exactly'' the same as on your working Dev bd???)

anm
Associate II
Posted on May 17, 2011 at 13:10

@STOne-32:

I tried various Software from Segger within the newest Beta-Versions. On all versions I get the same error.

@jj.sprague:

Sorry, I do not have any other JTAG device.

The STM32 on the dev-board is a STM32F103R8T6 (22036 93 ML
jj
Associate II
Posted on May 17, 2011 at 13:10

My firm has used both 48 & 64 pin STM32 devices with success. (with IAR and JLink) Believe that the odds of ''two'' STM32s being ''bad'' is very, very small. (unless they are being mishandled/misinstalled/ESD damaged etc.)

We have seen cases where clients have confused the pin numbers! (mixed the pinouts for 48 pin devices with those of the 36 or 64 pin versions) Meaning no offense - I am more suspicious of your pin interpretation and pcb than the STM32s. Suggest that you have a 2nd person review your pinout assignment of the STM32. (sometimes the same person ''misses'' something - repeatedly)

*** Eureka! (big breakthrough) ***

''VDDA and VSSA must be connected to VDD and VSS, respectively.'' Cannot tell if your schematic leaves these pins floating! Even if you are not using analog inputs - other STM32 functions require that these pins be connected! Figure 12 of STM's data sheet clearly shows VDDA/VSSA treatment. I am 99% sure ''this'' is your problem.

I've just reviewed your schematic - all key pins look correct.

Should VDDA and VSSA not ''cure'' your JTAG issue I would ''pull'' the robocup RF device - just in case. Hope it is the 2 analog power pins...

[ This message was edited by: jj.sprague on 12-05-2009 09:10 ]

anm
Associate II
Posted on May 17, 2011 at 13:10

Quote:

On 12-05-2009 at 07:59, Anonymous wrote:

My firm has used both 48 & 64 pin STM32 devices with success. (with IAR and JLink) Believe that the odds of ''two'' STM32s being ''bad'' is very, very small. (unless they are being mishandled/misinstalled/ESD damaged etc.)

I've tried three at the moment...

Quote:

We have seen cases where clients have confused the pin numbers! (mixed the pinouts for 48 pin devices with those of the 36 or 64 pin versions) Meaning no offense - I am more suspicious of your pin interpretation and pcb than the STM32s. Suggest that you have a 2nd person review your pinout assignment of the STM32. (sometimes the same person ''misses'' something - repeatedly)

A Professor has checked it also.

Quote:

*** Eureka! (big breakthrough) ***

''VDDA and VSSA must be connected to VDD and VSS, respectively.'' Cannot tell if your schematic leaves these pins floating! Even if you are not using analog inputs - other STM32 functions require that these pins be connected! Figure 12 of STM's data sheet clearly shows VDDA/VSSA treatment. I am 99% sure ''this'' is your problem.

Hmm if that is the problem, I will treat a collegue of mine! He said that he never connected this two, because he does not need them...

Quote:

I've just reviewed your schematic - all key pins look correct.

Should VDDA and VSSA not ''cure'' your JTAG issue I would ''pull'' the robocup RF device - just in case. Hope it is the 2 analog power pins...

[ This message was edited by: jj.sprague on 12-05-2009 09:10 ]

I have made 3 boards - two without the module.

I will check if the VSSa, VDDa is the error - in about two hours. I will report then.

anm
Associate II
Posted on May 17, 2011 at 13:10

jj.sprague - you made my day!!! it works as you said. I think my colleague will get troubles now (when he comes back from australia :-P)

Thanks, thanks, thanks!!! I think the pcb must be rebuild now ;)

lil-vince
Associate II
Posted on May 17, 2011 at 13:10

Hi,

Just to say that VSSA and VDDA are used for the RCs (HSI, LSI) and the PLL´s supply too, so it is often needed. (see ''power supply scheme'' in datasheet)

jj
Associate II
Posted on May 17, 2011 at 13:10

Glad that we persisted - and YOU got it.

You can probably ''tack'' 2 tiny leads to terminate VDDA & VSSA properly.

If yours is a school project I would always want to enable/use the analog inputs. You will find that employment is much richer if you have a ''mixed signal'' background.

Good luck...

anm
Associate II
Posted on May 17, 2011 at 13:10

Quote:

You can probably ''tack'' 2 tiny leads to terminate VDDA & VSSA properly.

I've done this for testing, but on the working pcbs I can't do that. They will work inside a roboter which plays football - so there are shocks when they collide with other robos and so on...

Quote:

If yours is a school project I would always want to enable/use the analog inputs. You will find that employment is much richer if you have a ''mixed signal'' background.

OK, thanks for the advice.