Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Einar Thorsrud — Wed, 10 Dec 2025 13:55:30 GMT

Hi Charlie,

This is a confirmed issue but there is no errata for this yet. I agree with your comment though, and have forwarded it internally.

Br,

Einar

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

erwallc — Wed, 10 Dec 2025 13:21:34 GMT

Hi Einar,

Intended to enable noise shaping in our project, also using nRF54L15DK, but it seems I should avoid that. Can''t see this issue in the SoC Errata nor the SoC Datasheet. Would be nice to see them updated if it is a confirmed issue.

Regards,

Charlie

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Einar Thorsrud — Fri, 05 Sep 2025 06:10:39 GMT

Hi Joseph,

We have identified issues in the implementation of the noise shaping filtering in the SAADC peripheral. We are continuing to look into this, but regardless of hte outcome this is issue is part of the SoC itself that as far as we can see cannot be worked around in software. So you either need to limit the input signal as mentionned in my previous post or avoid using noise filtering.

Br,

Einar

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Josephchiang — Fri, 05 Sep 2025 02:03:14 GMT

Hi Einar,

Thanks for your previous support on this topic. I’d like to provide some updates and also ask for clarification:

Kyle, who initially raised this question, was our summer intern. He has now returned to school, so I will continue the discussion on behalf of our team.
Regarding the unexpected spikes in SAADC output when enabling noise shaping (NS=1, NS=2), we’d like to confirm whether Nordic will continue investigating the root cause. From our perspective, it seems unreasonable that enabling the noise shaping function would introduce such a usage limitation.

We look forward to your feedback and any further insights you can share.

Best regards,
Joseph

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Einar Thorsrud — Fri, 29 Aug 2025 08:18:54 GMT

Hi Kyle,

We see a problem in the impelemtation of the noise shaping filtering in the SAADC. It is still being analyzed, but a possible workaround is to ensure that the maximum amplitude of the signal is less than half of the maximum (so that the highest bit is never used). An alternative is to use differential mode with negative input tied to GND. Or you can of course avoid using noise shaping.

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Fri, 22 Aug 2025 01:30:48 GMT

Hi Einar,

Thank you for your reply and for confirming the behavior.

I appreciate you looking into this further, and I'll looking forward to hear any updates once available.

Best regards,

Kyle Anderson

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Einar Thorsrud — Thu, 21 Aug 2025 10:28:47 GMT

Hi Kyle,

I am sorry for the late reply. Thank you for the detailed explanation and code. I am seeing the same as you on my end, with the seemingly flipped signal. I am looking more into this and will get back to you when we have an update.

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Mon, 18 Aug 2025 06:53:47 GMT

Hi Nordic team,

I just conduct new experiment based on the suggestion that Hieu give, for example, Turn on oversampling and use higher resolution and higher sample rate.

But i still have some question want to be solved.

The first question is when i reproduce the result that Hieu reply me.

In NS=0 condition, the result be the same as Hieu gave me:

However, in NS=1 and NS=2 condition, the signal still looks like flipped, i not sure the result is right.

I found when the signal over 1.8V(half full scale), the raw code will be negative, but i don't know why the result will be negative.

The raw code file is attached below

NS = 1

devzone.nordicsemi.com/.../ADC_5F00_SINE_5F00_3V0Vpp_5F00_1V5offset_5F00_83_5F00_14Bit_5F00_NS1.csv

NS = 2

devzone.nordicsemi.com/.../ADC_5F00_SINE_5F00_3V0Vpp_5F00_1V5offset_5F00_83_5F00_14Bit_5F00_NS2.csv

Thank you for your help.

Best regards,

Kyle Anderson

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Hieu — Fri, 15 Aug 2025 09:00:50 GMT

Hi Kyle,

Thank you for the wish.

If you can consistently reproduce the issues after trying the recommendations from my last reply, please simply reply here with the details of the experiment. We have got your application, so you only need to include information about input signals and any changes you made.

I fully trust that the next support engineer will be able to help you.
Then, best of luck to both of us.

Best regards,

Hieu

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Fri, 15 Aug 2025 08:04:11 GMT

Hi Hieu,

I just wanted to sincerely thank you for the time and effort you’ve dedicated to helping me investigate the issue with the nRF54L15 SAADC noise shaping.
Although the problem hasn’t been fully resolved yet, your insights, testing suggestions, and guidance have been very valuable in moving the investigation forward.

I understand that you will now be assigned to other tasks, and I truly appreciate your support up to this point. Your assistance has made a real difference in narrowing down potential causes, and I’ll continue to build on the work we’ve done together.

Thanks again, and I wish you the best in your upcoming projects.

Best regards,

Kyle Anderson

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Hieu — Thu, 14 Aug 2025 20:49:18 GMT

Hi Kyle,

In setting up the measurement today, I was reminded of an issue: the nRF54L15 DK VDD is 1.8V by default.

Have you changed this when you measured?

The issue is that the maximum voltage the SAADC can measure is VDD, and the absolute maximum rating of the pin would be VDD + 0.3V.
Therefore, we can expect voltage above 1.8V to be measured incorrectly.

Here are the results of the few tests I run today.
There was no strange spike like you had at first.
There was also no weird graphs like the ones you last sent.
There are only the clipped off peaks that is expected.

I was a little careless with note taking, so I lost track of the 2nd and 3rd result and am not sure what exactly the input was. However, consider they have the same characteristics of interest as the other results, I hope that is not a big problem.

For now, I would say, if you want to measure the signal correctly, reconfigure the nRF54L15 DK to 3.6V supply, which is the highest recommended operating voltage.
Then you can reattempt with the 3.0Vpp, 1.5V offset signal.

Please also make sure to enable oversampling and 12-bit or higher accuracy, as mentioned in the datasheet.

As for the graphs you last sent, I initially thought it would be aliasing. That's why I asked for your input and sampling frequency.
After I reproduce the setup today, I am now fairly certain that is the case.

I actually was assigned to something non-DevZone from yesterday and will be away for a few weeks. Another engineer will continue to support you, so please feel free to reply.
I will see if I can find time to reproduce my result some day, because they don't make total sense. 1.6Vpp shouldn't have been clipped off. I didn't notice that when I was just gathering data.

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Thu, 14 Aug 2025 06:51:13 GMT

Hi Hieu,

Of course, I will attached my test application along with the usage.

Environment: nRF Connect SDK 3.0.2

Base Configuration: prj.conf

Extra Devicetree Overlay: boards/nrf54l15_nrf54l15_cpuapp.overlay

UART baud rate:1000000

Usage: the AIN4 port will be ADC input

devzone.nordicsemi.com/.../adc_5F00_test_5F00_ns.zip

Thank you for your help.

Best regards,

Kyle Anderson

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Hieu — Wed, 13 Aug 2025 21:48:31 GMT

Hi Kyle,

I am trying to setup a test application but met quite a number of hiccups. Do you mind sharing your project to speed things up?

Best regards,

Hieu

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Tue, 12 Aug 2025 02:59:06 GMT

Hi Hieu,

OK, I will try it.
Could you please give me some example of correct result of noise shaping to let me compare my result

Thank you for your help

Best regards,

Kyle Anderson

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Hieu — Mon, 11 Aug 2025 22:46:27 GMT

Hi Kyle,

Could you try using oversampling? I asked for this a bit ago.

Best regards,

Hieu

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Mon, 11 Aug 2025 07:17:32 GMT

Hi Hieu,

Yes, I have the correct looking graphs from the reply on Aug 8, and have these latest graphs with an input of 3.0V Vpp and 1.5V offset.

The HSA_1M is mean high sample rate 1M Sa/s.

The input frequency is 1k Hz and sampling rate is 1M Sa/s, and i didn't use oversampling.

Other parameter stay the same as 1.6V Vpp test

Thank you for your help

Best regards,

Kyle Anderson

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Hieu — Mon, 11 Aug 2025 07:07:46 GMT

Hi Kyle,

Do I understand right that:

With an input of 1.6V Vpp and 0.8V offset, we have the correct looking graphs from the reply on Aug 8?
With an input of 3.0V Vpp and 1.5V offset, we have these latest graphs?

What does the HSA_1M in the file name stand for?

Did you change sampling rate and/or input frequency? What are they now? What is the oversampling configuration?
Is there any difference with those parameters between this 3.0V Vpp test and the last 1.6V Vpp test?

Best regards,

Hieu

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Mon, 11 Aug 2025 06:20:34 GMT

Hi Hieu,

I adjust my input signal to 3.0Vpp 1.5V offset, the output waveform still look incorrectly.

The output waveform attached below

devzone.nordicsemi.com/.../ADC_5F00_SINE_5F00_3V0Vpp_5F00_1V5offset_5F00_1k_5F00_HSA_5F00_1M_5F00_12Bit_5F00_NS1.csv devzone.nordicsemi.com/.../ADC_5F00_SINE_5F00_3V0Vpp_5F00_1V5offset_5F00_1k_5F00_HSA_5F00_1M_5F00_12Bit_5F00_NS2.csv

I want know the result is correct or incorrect.

Thanks for your help.

Best regards,

Kyle Anderson

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Hieu — Fri, 08 Aug 2025 12:05:25 GMT

Hi Kyle,

The spikes seem gone. The troughs seem quite sharp and rough compared to before, but that seem to be related to sample rate and number of samples graphed.

Are you satisfied with the current results?

The background behind this suggestion is that noise shaping is recommended to be use with 12-bit or higher resolution, and such resolutions requires oversampling.

Best regards,

Hieu

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Fri, 08 Aug 2025 02:19:35 GMT

This is the result.

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Fri, 08 Aug 2025 02:04:57 GMT

Hi Hieu,

Okay, thank you.

I would try with oversampling.

I currently change my sample rate to 1M Sa/s and resolution to 14 Bit to check what would be different.

Thanks again for your help.

Best regards,

Kyle Anderson

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Hieu — Thu, 07 Aug 2025 11:11:54 GMT

Hi Kyle,

We are still investigating. Could you please try with oversampling?

Best regards,

Hieu

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Sat, 02 Aug 2025 06:59:55 GMT

Channel configuration: P1.11(AIN4),
NRFX_SAADC_DEFAULT_CHANNEL_SE(SAADC_INPUT_PIN, 0)
Signal: Sine wave 3.0V Vpp, 1.5V Offset 2.44Hz
Connection Setting; P1.11 and GND
The value associated with the CSV file: unsigned_val

This is the configuration.

Thank you for your help.

Best regards,

Kyle Anderson

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Hieu — Fri, 01 Aug 2025 20:34:55 GMT

What is the channel configuration you are using?

What is the input signal? I understand it is from a function generator, but what is its configuration?

How is the function generator connected to the DK?

In this code snippet:

//printk("ADC Samples:\n");
            for (int i = 0; i < p_event->data.done.size; i++) {
                current_value = ((int16_t *)(p_event->data.done.p_buffer))[i];
                uint16_t unsigned_val = (current_value < 0)?(uint16_t)(current_value + 4096):(uint16_t)(current_value);
                average += current_value;
                if (current_value > max) max = current_value;
                if (current_value < min) min = current_value;
                printk("%d\n", unsigned_val);
            }
            average /= p_event->data.done.size;
            //printk("\nAVG=%d, MIN=%d, MAX=%d\n", (int16_t)average, min, max);

Which value is associated with which CSV file?

I looked at your raw data but there are a few things that I can't understand.

First, if the input is a sine wave, then the ADC measurement should be a sine wave too, but... it isn't quite one. Looks like something is flipped?

How did this data point happen, I don't even understand:

T	Val
690	-1868
691	-1772
692	-2017
693	-184
694	1961
695	1807
696	1856

It's these data point like the -184 above that cause the spikes in your graph.

Secondly, I sorted the raw value from lowest to highest, and plot it

Notice how there is a complete absence of value around 0?

Usually, you don't miss value around the average. You would miss value at min and max.

At least, it looks like there are no spike from the SAADC peripheral. The issue must come from the setup, configurations, and calculations.

It must be something simple, but for I can't imagine it for some reasons...

RE: Unexpected Spikes in SAADC Output When Using Noise Shaping on nRF54L15

Kyle Anderson — Thu, 31 Jul 2025 06:28:51 GMT

I want to know the raw code result when i enable the noise shaping is correct or not.

And know how to fixed the spike found after the conversion process.

Thank you.