BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

AmbystomaLabs — Sat, 23 Dec 2017 17:06:58 GMT

In mixed signal design it is not uncommon to have separate digital and analog grounds. It is one method of solving the EMI/EMC issues which I pointed out earlier.

The nRF doesn't provide for separate grounds, however Sweety may have isolated grounds for other devices in their custom design.

Generally I don't like separate grounds. I have been called in to fix many products where the original design team chose to isolate grounds inappropriately.

Sweety should really be approaching this as an EMI/EMC problem. All the indications are as such. If they are truly running separate grounds then they should also post their schematic and layout artwork so that the problem can be debugged properly.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Matt — Fri, 22 Dec 2017 19:35:07 GMT

I guess I'm on shaky grounds here... there should only be 1 "ground" in your circuit, correct?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Sweety — Fri, 22 Dec 2017 14:21:06 GMT

Is there anything to do with analog ground and digital ground?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Matt — Wed, 20 Dec 2017 14:08:17 GMT

Can you ask Sandeep to take a look then?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Sweety — Wed, 20 Dec 2017 14:00:52 GMT

I am using Arduino Core for Nordic Semiconductor nRF5 based boards written by sandeep mistry. Link: github.com/.../arduino-nRF5 I am using the above-mentioned firmware for NRF52DK and my custom board.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Matt — Wed, 20 Dec 2017 13:56:49 GMT

Are you running the same firmware in your custom board and in the nRF52 DK? Don't you define your own custom_board.h to replace the bsp.h and remove all the button and led code specific to the DK? Also, are you using the UART in your custom board? If so, can it be the interrupt handler there causing the issues?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Sweety — Wed, 20 Dec 2017 13:54:07 GMT

Yes, your understanding is correct.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Matt — Wed, 20 Dec 2017 13:51:05 GMT

So the current situation is (a) with custom board, when the motor is turned on, the BLE connection is disconnected within 1 minute (b) with nRF52 DK, when the motor is turned on, the BLE connection is fine, and stays fine after 1 minute of motor running.

No EMI/EMC fixed has been applied to either (a) or (b).

Is my understanding here correct?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

AmbystomaLabs — Wed, 20 Dec 2017 12:49:06 GMT

Ok, so you are back to your EMI/EMC problem. Are you actually going to fix it now?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Sweety — Wed, 20 Dec 2017 07:10:49 GMT

I used led Indications for resetting and ble connection. With this, I confirmed that ble is disconnecting.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Matt — Tue, 19 Dec 2017 12:37:23 GMT

You cannot "agree" that nRF is resetting until you verify it is truly resetting, and one way to "verify" that is to add debug trace into your main() and see if that trace message repeats again and again. You can use the serial debug (debug message pumped out through uart), or use jlink real time trace (RTT) in which the debug message is sent out through the SWD interface. You can still use the RTT with your custom board, you just need to use the nRF52DK as the JLink to connect to your custom board (or just use your stand alone Segger JLink if you already have one). Search in this forum "RTT" and you'll find code examples that you can use in your project.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Sweety — Tue, 19 Dec 2017 04:22:00 GMT

For nRF52DK, I am using external 3.3V power supply. I am using similar power supply with my custom board. I agree that nRF is resetting, only ble is disconnecting. Can we eradicate that? What do you mean by adding rtt to my custom board image ?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Matt — Tue, 19 Dec 2017 04:15:39 GMT

For the nRF52 DK are you using CR2032 as the power source? Also if you have some spare cycles I suggest you add some simple rtt to your custom board image, just to convince yourself the nRF is not resetting when it's disconnected...

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Sweety — Tue, 19 Dec 2017 04:00:44 GMT

Thank you for your valuable suggestions. I tried the same vib motor setup with nRF52DK. Ble Connection is not getting disconnected on starting vibe motor with nRF52DK. What might be the reason for this?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

AmbystomaLabs — Mon, 18 Dec 2017 17:57:00 GMT

The schematic for the LC network is on the first page of the doc. Please actually read it thoroughly before commenting. See here: www.precisionmicrodrives.com/.../nokia-noise-filtering-example.original.png

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Sweety — Mon, 18 Dec 2017 17:54:04 GMT

In the best practices guide, he mentions to place a capacitor across motor terminals to counter attack radiate and conductor noise. But he didn't tell any thing about LC filter. Is it required to put LC filter or a capacitor is sufficient?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

AmbystomaLabs — Mon, 18 Dec 2017 17:35:55 GMT

Yes, just follow the guidance in the best practice article.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Sweety — Mon, 18 Dec 2017 17:29:51 GMT

Okay. I have to connect LC network across motor terminals. Did I understand correctly ?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

AmbystomaLabs — Mon, 18 Dec 2017 17:20:47 GMT

No, RC is first order network LC is second order network. You need a second order network fully isolating the vibe motor.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Sweety — Mon, 18 Dec 2017 17:09:33 GMT

Thank you for your valuable suggestions. So, to eradicate radiated noise, I have to put a filter across motor terminals. Am I right? Can I go with RC filter rather than LC filter.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

AmbystomaLabs — Mon, 18 Dec 2017 16:53:29 GMT

Steppers have similar EMI/EMC problems but a more complicated drive mechanism. Since they sit more than they move and don't have brushes, likely you are dealing with a low frequency problem.

I'm sure you can apply it conceptually, but the final solution will look different due to how steppers work. I would start with guidance on filter designs from the stepper manufacturer.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

Matt — Mon, 18 Dec 2017 16:46:46 GMT

Thanks for that best practice article! Should we follow the same practice for driving a stepping motor?

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

AmbystomaLabs — Mon, 18 Dec 2017 16:42:31 GMT

And, your comment on using "strong decoupling" caps is rather erroneous. There is no one "strong" decoupling cap. A single cap covers a narrow band of frequencies around its SRF and forms a first order network based on the source resistance. In this case the source resistance is really low since it is a LiPo, so I doubt they are doing much.

You need a comprehensive EMC/EMI solution not just a decoupling cap.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

AmbystomaLabs — Mon, 18 Dec 2017 16:08:31 GMT

A comprehensive LC plan will solve all of it. But don't forget the zener and series resistance. The turn off spike from the motor could fry the LiPo and FET over time and turn on surge from lack of a series resistance could fry your vibe motor quickly.

You can use a spectrum analyzer and an o-scope to figure out precisely what noise is the worst to guide your LC filter design.

If you wanted to really engineer the turn on/off side of things, put an RC network on the FET gate and treat the nRF control signal as a PWM. Then you can have a soft turn on/off which will eliminate the surge and spike.

RE: BLE is getting disconnected when the motor is turned on in NRF52832 circuit.

AmbystomaLabs — Mon, 18 Dec 2017 16:04:09 GMT

As I said the noise could be radiated and/or conducted. Radiated (ie, off of power lines running between motor and LiPo) would likely be a 2.4GHz problem, conducted through power, gnd and switchers could be anything from mid megahertz (affecting the low IF in nRF, causing NVIC chaos through false pin asserts) through to 2.4GHz.

You should be able to easily come up with an LC solution that covers that range. Just make sure it covers both legs of the power connection.

This is a nice, though dated, article on the subject: www.precisionmicrodrives.com/.../ab-008-vibration-motor-best-practices-from-mobile-cell-phones