CC1125: CC1125 SPI - MISO (SO) stays low

Hi Andy,

Could you post the schematic of your custom board so that I can take a look? 

Regards,

HG

It's a client's design, so I'll have to check...

Meanwhile, I found a checklist here: https://e2e.ti.com/support/wireless-connectivity/sub-1-ghz-group/sub-1-ghz/f/sub-1-ghz-forum/581966/cc1125-issues-with-spi-miso

1. Is the xosc running? (Check pin 32)

2. Do you have 1.25 V on Rbias? (Check pin 14)

3. Do you have 1.8 V on the DCoupl pins? (Check pin 6)

We're using an external TCXO (clipped-sine output via 22pF, per datasheet) - that is working correctly.

But I'm not seeing the 1.25V at pin 14, nor 1.8V at pin 6.

Andy Neil said:

In this [dev board] setup, I can successfully read the PARTNUMBER (8F) and PARTVERSION (90) extended registers

PARTNUMBER reads 0x58 - which is correct for CC1125, according to the User's Guide.

PARTVERSION on the BOOSTXL-CC1125 board is 0x21.

Andy Neil said:

I'm not seeing the 1.25V at pin 14, nor 1.8V at pin 6

Actually seeing 0V at both.

Hi,

Make sure that the TCXO is properly soldered. And could you share the part number for the TCXO used?

If the customer do not want to share their schematic, can they share at least the part which contains CC1125?

Regards,

HG

Andy Neil said:

I'm not seeing the 1.25V at pin 14, nor 1.8V at pin 6

- This indicate something fundamentally wrong with the board. In this case it doesn't matter if the TCXO is working or not since the digital domain of the chip doesn't get power and all bias currents in the analog domain is off. 

- What is the current consumption?

- Have you checked that Rbias is 56 kohm and correctly soldered? 

So DCPL is the decoupling for the digital domain.

The User Guide says (section 10.1),

By setting the CSn pin low, the voltage regulator to the digital core turns on and the crystal oscillator
starts.

So does that mean DCPL would be expected to be 0V until CSn is pulled low? Does it also depend on RESET_N?

TheGhostOf said:

What is the current consumption?

Not sure. Not excessive, but hard to distinguish from the STM32.

TheGhostOf said:

Have you checked that Rbias is 56 kohm and correctly soldered?

Yes, it's marked "563"; soldering looks OK to my software-guy eyes...

I'm afraid they don't want any of the schematic shared.

I've cross-checked it against the swrr106a reference design, and all the supply & decoupling connections match

Andy Neil said:

does that mean DCPL would be expected to be 0V until CSn is pulled low?

I tried it on a BOOSTXL-CC1125 board:

Board powered from P2.2 (5V) & P2.4 (GND); P7 jumpered for "Launchpad power" (link 2-3)

This, AIUI, leaves the CC1125 powered but with (effectively) nothing on its digital interface lines - so CSn not asserted.

In this state, DCPL shows 1.8V and RBias shows 1.2V.

Just to make sure, the die paddle of the chip is connected to ground? 

Well, it should be!

The client has successfully used this chip before so would know about it...

Always a problem with these hidden ground pads - how to tell if it really is properly connected?!

I did note that the bottom pad is the only ground connection to the chip.

Do these symptoms sound likely for a missing ground connection?

I tried a similar test on the "problem" board:

I held the STM32 in reset (NRST pulled low);
In this state, all the STM32 GPIOs act as “input floating” – so effectively disconnected, and should have no effect on the CC1125.

 I still get 0V at the CC1125’s DCPL and RBIAS pins.

In this state, the external TCXO is also stopped - would that make a difference?

The CC1125 should start up in IDLE state meaning that at least RBIAS should be 1.25 V. 

It's limited what you can do this not too work:

- Remove the ground (not connect the pad)

- Rotate the chip

- Short Rbias to ground 

If they have successful made boards with CC1125 before they need to compare the old and new design and see where they have done something wrong. A schematic + layout review would normally show these errors clearly. Note that you can request a review through secure channel, no need to upload to the forum. 

Thanks.

TheGhostOf said:

The CC1125 should start up in IDLE state meaning that at least RBIAS should be 1.25 V.

Even with the TCXO off?

TheGhostOf said:

you can request a review through secure channel

How do we go about this?

The clock is not required for the chip to start. You can actually turn on and off the TCXO to save current, see https://e2e.ti.com/support/wireless-connectivity/other-wireless-group/other-wireless/f/other-wireless-technologies-forum/237099/cc112x-and-tcxo-power-enable

Review: www.ti.com/.../SIMPLELINK-SUB1GHZ-DESIGN-REVIEWS

thanks.

Yes, we do (will) turn off the TCXO to save current - that's why it's off when I hold the STM32 in reset!