Capacitor values on DEC pins

RE: Capacitor values on DEC pins

George555 — Tue, 05 Apr 2022 06:02:37 GMT

I'll go with 1uF + 100nF (internal) and see how ti goes...

Thank you !!

RE: Capacitor values on DEC pins

JONATHAN LL — Mon, 04 Apr 2022 21:24:14 GMT

[quote user="giorgos3924"]Should I modify the external capacitor on DEC4 to something like 0.82uF (+ 100nF internal) to get closer to the right value?[/quote]

It is difficult to say, using a 1pF or a 0.82pF both are better options then not adding any of them. The performance might that is effected might be the radio, as the value of the decupling will effect the filering of spurious that effects the RF. I would test both values and see if you see any difference on radio performance. I dont think it will have a major impact on current draw.

[quote user="giorgos3924"]Regarding DEC3. Unfortunately I cannot change it.[/quote]

It is what it is, you will just try to make the best out of the situation and if it cant be changed than so bee it. Test and see of you gett acceptable results.

If you have any other questions feel free to ask and i will do my best to help.

Regards,
Jonathan

RE: Capacitor values on DEC pins

George555 — Thu, 31 Mar 2022 12:07:54 GMT

Thank you Jonathan.

The high current of 150mA I measured (as I mention in the other ticket) was from a damaged module.

After finding the errors in the internal schematic of the modules I'm trying to do the best I can to improve the design.

In my understanding no matter the issues I've faced it's very crucial to fix the errors and get closer to Nordic's reference circuit. So, I'm now focused on that.

Currently only DEC4 (1.1uF) and DEC3 (100nF) are different than Nordic's design.

Should I modify the external capacitor on DEC4 to something like 0.82uF (+ 100nF internal) to get closer to the right value?

Regarding DEC3. Unfortunately I cannot change it. I have already planned on replacing the beacons with a better design but for now I'm trying to make some room for improvements in the current design if possible. I need to buy me some time.

RE: Capacitor values on DEC pins

JONATHAN LL — Thu, 31 Mar 2022 11:29:23 GMT

[quote user="giorgos3924"]Are those capacitors used only in DCDC mode?[/quote]

The capacitors are there regardless of DCDC or not, Only difference is the L2 and L3 inductors.
This is the LDO reference schematic and here is the DCDC

[quote user="giorgos3924"]Would it be better to operate the chip with its internal LDO?[/quote]

Due to your power requirements and supply it would be best to use DCDC.

Ideally the DEC4 should be 1uF and not 1.1uF, you can try and see how the performance is but the capacitance value on the DEC pins are closely related to the regulators so changes in value will result in changing behaviour of the regulator.

So DEC4 and DEC1 should be as they are in ref design, and you can gett get DEC1 to be as in the specs, DEC4 is more difficult but 1.1uF is better then 0.1uF. DEC3 is related to the radio peripheral so deviation on this value can change the RF performance for the worse. But it should not affect the current usage in any significant way.

I the other ticket you have you say that you measure 150mA, this is a indication to that something is fried in the SoC. So maybe there has been a short to ground on one of the decupling capacitor pins ?

Only a few of the devices is experiencing this current draw issue, so this might be a quality control issue from the module provider, not clear exactly what is the cause.

As it stands now i would recommend that you change to a different module supplier or get the broken modules replaced, as they do not perform as expected.

Here we have a list of 3rd party module providers: https://www.nordicsemi.com/Nordic-Partners/3rd-party-modules

Regards,
Jonathan

RE: Capacitor values on DEC pins

George555 — Tue, 29 Mar 2022 14:44:20 GMT

Notice the DEC3 has an internal capacitor 100nF instead of 100pF. The manufacturer of the module have modified that value too and I cannot change it.

I'm going to remove all the external capacitors I added but keep only the 1uF on DEC4.

Below you can see how the schematic will look like

So, according to the schematic there will be actually two deviations in the reference circuit.

DEC4 1.1uF instead of 1uF

DEC3 100nF instead of 100pF.

Are those capacitors used only in DCDC mode?

Would it be better to operate the chip with its internal LDO?

Unfortunately I cannot do further modifications on the circuit because I cannot intervene on the module itself. I can't think of any better solution any suggestions are welcomed of course.

Also, I measured the voltage on the DEC1 pin with and without the external capacitor

NOTE: There is a 10 Ohm series resistor 1% as Rsense during the measurements below that I use for current measurement. I have already bought a PPK2 but before I got it on hands I'm trying to debug the circuit with what I have.

2x 100nF (one internal, one external)

1x 100nF (Internal)

Voltage on DEC4 pin with 1uF + 100nF (internal)

The noise in the middle is during packet transmission.

Voltage on DEC3 pin with 100nF (internal)

The shape during transmissions is exactly the same with the shape I measure across the Rsense (10 Ohm) during the transmissions.

RE: Capacitor values on DEC pins

JONATHAN LL — Tue, 29 Mar 2022 14:36:10 GMT

Hi,

Seems like there is a deviation in the module from the reference circuitry you are using for the nrf52810, as stated in the previous Devzone thread.

The 100nFon DEC3 is not recommended so changing to 100pF, the correct value is something I would recommend.

So it look's like you have added decupling caps on DEC1, DEC3 and DEC4, since the module all ready seems to have components mounted internally on these pins this will result in the decupling cap values being outside of the recommended values, DEC4 is the DCDC component and for optimal performance you should stick to the reference circuit values.

So if i understand correctly this is how the device looks, the RF matching network is not the same in this illustration but it does not matter.

Have I understood correctly ?

Regards,
Jonathan