BQ51221 Wireless Power Receiver

Is your customer using the same coil as EVM? Can you capture the VRECT signal during the power up? Do you have the schematic?
Since you mentioned that in WPC mode, bq51221 does power up before it shuts down, using a Qi-sniffer to read the actual commands would give a good answer.

Thanks,
Wenjia

Thanks Wenjia,

Yes, they are using the same coil. They also have an alternative coil that they're trying (and the other coil works on the EVM without modifying any components) but since they're having trouble getting it working consistently they're focusing on using the same coil that came with the EVM right now.

They have considered that maybe they need to get a Qi sniffer. I think it would help with the reason for cutting out in OtherCoilNoLoad.bmp. I suspect that one may be due to VRECT being too high (side question - what components would we vary to get VRECT lower?). In the screenshots with the 27ohm load is there enough communication that the sniffer can tell us what's going wrong there? I wasn't sure if it's even communicating at that point.

Something else to note: their board always has that short startup attempt and reset whether it works or not (e.g. 27ohmLoadDelayed.bmp, NoLoad.bmp), but the EVM never does that.

Since I cannot post customer design files on the public forum, I’ll send them to you via my local TI contacts. The zip file contains a few screenshots, a ReadMe file, and the schematic. U2 is not populated so that they can focus on the BQ51221, and they switched R10 and R12 (R10 is populated, R12 is not populated) to match the EVM. Let me know if there's anything else I can provide to help.

Text from ReadMe in the zip file:

All screenshots:
Channel 1 = VRECT
Channel 2 = VOUT

Using the same coil as the evaluation board:
NoLoad.bmp - Our board with no external load (only the load on the board ~120k) 27ohmLoad.bmp - Their board with a 27ohm load (should be 185mA) on the output.
27ohmLoadDelayed.bmp - Added the 27ohm load about 1.5 seconds after the output is stable. Output stays stable.

Using a different coil (IWAS4832FFEB9R7J50) - coil does work on EVM OtherCoilNoLoad.bmp - Communication often cuts out after second burst of data, even with no load.
OtherCoil27ohmLoad.bmp - Output won't start, similar to 27ohmLoad.bmp

Thanks,

Hello,

I've reviewed your schematic and plots, below some initial comments :
1. Communication issue is a possibility. So please try increasing the COMM1/2 (C6 and C22) capacitors to 120nF.
2. Please remove R8/R9 for now for debug purpose.
3. Remove the resistors on LPRB1/2, leave them float.
4. Make sure R10 is not populated.
5. Make sure all the capacitors that are connected to AC1/AC2 has >25V voltage rating.
6. Is this a 2-layer or 4-layer board? 2 layer may have more noise.
7. Can you provide the zoom in plot of the failure point that after a few seconds of powering up, the VOUT dropped from 5V to 0V? Also a zoom in plot right at the point of OUT goes high. Please capture VRECT and VOUT, capture TS too if you have a spare probe.
8. Make sure the soldering is all good
9. What do you mean by sandwiched the NTC thermistor between the coil and PCB?

Thanks,
Wenjia

Hi Wenjia,

Thanks for the suggestions. Unfortunately, still no luck. Details below:

 

1. It decreased the size of the communication spike on VRECT a bit, but didn't make any difference to the overall behaviour of not starting up with a load on the output. I had also previously tried decreasing the value to 47nF because the datasheet specified a value from C1/8 to C1/3 (and 82nF is C1/3) but that also didn't help. Spikes were much bigger on the board with the other coil.

 

2. Done.

 

3. Done.

 

4. Confirmed - not populated.

 

5. Thanks for mentioning this, we discovered that despite our schematic specifying 50V for all capacitors, C5 and C21 (on AC1/AC2) were mistakenly populated with 10V capacitors and C17 and C18 (on VRECT) were mistakenly populated with 6.3V capacitors. Turned out there was a part number mistake in our stock library. We were really hopeful that this was the problem but after replacing the capacitors the behaviour hasn't improved.

 

6. This is a 4-layer board.

 

7. I will send you a number of screenshots because I wasn't sure how zoomed in or what exactly you wanted to see. I didn't have a spare probe so I took a couple separate screenshots with TS, not that it mattered (it's always 0V)- a difference I have noticed is that TS is always grounded on the board that shuts down after 2 seconds. On the EVM and the board that has the same coil as the EVM (the one that doesn't shut down) there are 500mV "blips" every ~1.7sec while it is connected. Could that mean that maybe this board's TS pin isn't soldered correctly? All of the screenshots are from board 1001001 - the one that is not using the same coil as the EVM. (though as I mentioned previously, the EVM does work with this coil. Vishay IWAS4832FFEB9R7J50)

 

8. We did initially have a problem with our CM not properly soldering the BQ51221. We sent the boards back to them for fixing and sent a few boards to a more expensive place that could guarantee it was done properly. We are currently working with 3 boards (2 from new CM, 1 from our usual CM) that are all exhibiting the same problems and soldering looks okay.

 

9. That's just referencing how we will mount the thermistor in the enclosure. Right now we have a normal through-hole 1/8W 10kOhm resistor in its place.

 

Any other suggestions that we could try or other documentation to read? We feel like it must be a simple problem but we have no idea what it could be.

 

Thanks,

What you observe on TS pin of the working board is correct - TS is internally driven and should have a voltage pulse above 380mV every 1.8s. If the TS voltage is below 380mV which is the TS hot threshold, the IC will send EPT to transmitter and the system will stop sending power to the receiver. So can you please try connecting 10k fixed resistor from TS pin to ground? Please make sure the pin is not shorted to ground.

Thanks,
Wenjia

The layout looks mostly good. Just one recommendation: put some vias on the GND copper of the top layer so that it is connected to the bottom layer. It will improve the thermal performance.

Regarding to the original issue, TS is still most likely the issue. The TS voltage has to be above 350mV to enable the part. With 10k resistor connected between TS and ground, the TS pin voltage should be above 500mV if the pin connection is good. Section 9.3.6.1 on pg# 15 of the datasheet described various EPT conditions and please note when TS/CTRL < Vts-hot, EPT will happen. If the customer has a Qi sniffer, it will return “Over Temperature” in the GUI and this is the quickest way to tell what is the reason for power shut down.

Thanks,
Wenjia

Thanks for the board review.

Please don't be distracted by the TS issue; that is one board with a different problem.

The other boards we have are showing the proper > 500 mV pulses on TS; we see 640 mV.  When the load is applied after linking, it works.  It's only when the load is present prior to linking that the board shuts down immediately.

 

It is really hard to tell what the issue could be. To get a Qi sniffer and read the information directly from the software is the quickest way. You can also send the board to us to look at from our side.

Thanks,
Wenjia

Hi Wenjia,

They have started using the BQ51020 (Qi only) and it is working.  They are procuring a Qi sniffer, but just looking at what the BQ51221 is doing with a scope, it appears that the output is collapsing before significant communications exchange takes place.  (This is in contrast to what they see when the TS pin is shorted to ground--they see the receiver and transmitter communicate several times and then decide to shut down.)

 

They will use the BQ51020 for now, but they would still like to get the BQ51221, and would like to take you up on your offer to look at their board.  They will prepare two boards to send to TI sometime next week.  Please let me know the protocol for getting these boards to you.

 

Thanks,