Skip to main content
Alexius P. BY
Alexius P. BY
Associate
May 28, 2015
Question

How to reduce FLASH noise on STM32F205VET6?

  • May 28, 2015
  • 11 replies
  • 2320 views
Posted on May 28, 2015 at 23:12

Hello!

I am work with uC STM32F205VET6. Its flash is very noisy!

How to reduce this noise with software methods? I tried turn off HSI, prefetch and slowdown core clock. Also I turned on spread spectrum clock generation. But there is no signify help from my actions.

P.S.: Thanks for answers and sorry for my english.

#flash #noise #stm32f2
This topic has been closed for replies.

11 replies

Tesla DeLorean
Tesla DeLorean
Guru
May 28, 2015
Posted on May 29, 2015 at 00:10

When executing or when erasing/writing?

What does this noise look like running from FLASH vs running from RAM?

Tips, Buy me a coffee, or three.. PayPal Venmo (See Profile) Up vote any posts that you find helpful, it shows what's working..
waclawek.jan
waclawek.jan
Super User
May 29, 2015
Posted on May 29, 2015 at 07:52

What noise are you talking about? Where and how do you experience that noise?

JW
Alexius P. BY
Alexius P. BYAuthor
Associate
May 29, 2015
Posted on May 29, 2015 at 08:41

In my device uC placed near embedded GPS antenna. I connected my laptop directly to RX/TX pins of GPS module. With special software I saw that noise of FLASH jamming GPS signals.

The noise produced by FLASH in erasing/programming mode is not important for me as the noise in executing mode.

The noise dissapear with program code placed in RAM.

For example, I attached two screenshots.

________________

Attachments :

FLASH.jpg : https://st--c.eu10.content.force.com/sfc/dist/version/download/?oid=00Db0000000YtG6&ids=0680X000006I143&d=%2Fa%2F0X0000000bj0%2FXVYf44Wp3_UBzr9OkaoaKarzrAuDxbCJFEe1LNvrthw&asPdf=false

RAM.jpg : https://st--c.eu10.content.force.com/sfc/dist/version/download/?oid=00Db0000000YtG6&ids=0680X000006I0jU&d=%2Fa%2F0X0000000bix%2Fivh6BFJleSGl8IGW21Vpz37P_KevCMC0BW4aVzTeSrg&asPdf=false
AvaTar
AvaTar
Senior III
May 29, 2015
Posted on May 29, 2015 at 09:26

Except of not running from Flash, I can't see a method of avoiding it.

But you could shield the MCU (faraday cage), and/or filter the power supply and communication lines.

Alternatively, you can try a low-power MCU, with smaller Flash and lower max. clock frequency. Catching NMEA sentences doesn't require much performance.

stm32forum
stm32forum
Associate III
May 29, 2015
Posted on May 29, 2015 at 09:41

Looks more to me for a pour designed pcb.

Try shielding the antenna, to see if it's cumming trough the air or over the power supply.

waclawek.jan
waclawek.jan
Super User
May 29, 2015
Posted on May 29, 2015 at 09:55

While I agree with others that the real solution is in hardware, you can try to play with ART and prefetch as these appear to have influence on the amound of digital/supply noise, see AN4073.

JW
Hamid.Wasti
Hamid.Wasti
Associate II
May 29, 2015
Posted on May 29, 2015 at 10:01

I also think that this looks like poor layout and or poor power distribution architecture.

Do you have the micro properly bypassed? I am talking about looking at the actual layout, size/shape/location of power/ground traces, not just number of bypass caps on the schematics.

Where does the return current flow in relation to the GPS antenna and the antenna trace? Is the noise from the processor traveling on the ground plane used by the antenna or its trace?

Is the noise from the processor, both power and ground, flowing through the GPS module?
Tesla DeLorean
Tesla DeLorean
Guru
May 29, 2015
Posted on May 29, 2015 at 15:57

What is the frequency and characteristics of the noise? ie spectrum analyzer vs anecdotal.

What is the IF for the GPS receiver, or what make/model? I haven't seen an STM32 locally jam a GPS receiver, and I've built a few systems combining the two and cellular radios.

Can you change the PLL comparison frequency from 1 MHz to 1.5 or 2.0 MHz?

As others have suggested you'd want to look at the supplies, bulk/decoupling capacitors, and caps/voltages on VCAP pins.

Tips, Buy me a coffee, or three.. PayPal Venmo (See Profile) Up vote any posts that you find helpful, it shows what's working..
Alexius P. BY
Alexius P. BYAuthor
Associate
June 10, 2015
Posted on June 10, 2015 at 21:39

Oh, sorry, I had no ability to access to forum in last week.

It is to be regretted that there is no simple software solution.

Unfortunately, I can't answer to all our questions, because I have not enough information about noise. But thanks for hints. I will take into account all of them.

Hamid.Wasti
Hamid.Wasti
Associate II
June 10, 2015
Posted on June 10, 2015 at 23:01

>

It is to be regretted that there is no simple software solution.

What is really regrettable is that people think that poor hardware design should be fixable in software. 

Hardware design is a lot more than slapping together a few chips on a board. Do it without understanding and insight and bad thinks are likely to happen. If the design happens to include sensitive elements like RF or high precision analog, these bad things become evident quickly.
Show more replies
Terms & ConditionsAccessibility statement