AFE5832: The outputs of all odd channels are abnormal. However the ones for all even channels are okay.

Addtional information is that it seems the data of all odd channels are okay at the very beginning of data-receiving process. 

Hi Yinhui,

Thanks for reaching out to us.

Sorry to hear about your issue with AFE5832. 

Straightaway, I don't recall any know issue with odd channel data being abnormal. I would like some more information to be able to comment further:

1. Can you describe your setup?

• Are you using customer EVM or custom board? What is the capture card, is it TSW1400?
• What is the sampling speed? Can you try reducing the sampling speed and check if the problem persists

2. Can you share the data of odd and even channels? Based on the data, we could see if there is any pattern or something?

3. Is it happening on multiple devices?

Thanks & regards,

Abhishek

Hi Abhishek,

Thanks for your support. This is Yinhui. I don't know why my name appears as "?? ?".

For your questions, 

1. We build our own board for the use of AFE5832 for ultrasound imaging. Currently the sampling speed is 80MHz. We will try reducing it and see whether the problem is stiil there.

2. To receive the data, we removed the transducer and make the 7332 not transmit any signal. Then we observe the received data. At the very beginning, the data of channel 1 (appear as "0") and channel 2 (appear as "1") are as following:

We bypassed the digital processing. Therefore, the offset is there. Both signals are okay in our opinion.

However, after 5 seconds of performing the data-receiving process, the data of channel 1 (appear as "0") and channel 2 (appear as "1") becomes as following:

After another 5 seconds of performing the data-receiving process, the data of channel 1 (appear as "0") and channel 2 (appear as "1") becomes as following:

3. Yes, this situation happens for multiple AFE5832s. 

Hi Yinhui,

Thanks for sharing the data. Its very useful.

It looks like some sort of systematic issue and we haven't seen such kind of noise from the ADC in the past. It's quite unique and I think something else outside the ADC is coupling at the same instant when odd channels are getting sampled. 

I am saying this because the same ADC samples the odd and the even channels. 

So in you system could you identify something which is running at a lower frequency and could couple with the signal chain.

Looking at the data it looks like some periodic noise is entering the system. And from the waveform looks like it is a factor 64 with respect to sampling.

Can you do the following:

1. Put the device in test pattern mode and observe if the test pattern (say ramp) is getting distorted.

2. Are you sure digital is bypassed, can you verify specifically the High Pass Filter, is it enabled or not?

3. Reduce sampling speed and check if the frequency of this distortion changes.

4. Can you also probe the ADC inputs, both odd and even channel. 

5. Are you doing any SPI operation after t = 0, i.e after you start the capture. Is there any settings getting written to the device.

Thanks & regards,

Abhishek

Hi Abhishek,

Thanks for the information. The received data for test patterns (all including ramp) are okay. We further perform the measurements on the board and found out sth. For the input of AFE5832, the schematic design is as following:

Since we remove the transducer and make 7332 not transmit any signal, the correct signal measured here should be around zero-intensity. The green numbers in the figure denote the measured points. The measurements of 1&2, 5&6 and 7&8 are similar as following, which is correct:

However, for the measurement of 3&4, the obtained signal is like this:

This waveform is like the noise we obatined from the system. Is there sth wrong for the interface of AFE5832?

We checked the detailed part number of our used AFE5832. It is found that our previous work is based on AFE5832LPZAV. However, the currently used one is AFE5832ZBV. The details of register writing are based on AFE5832LP. Could this be the reason of such unusaul noise? 

Can you help provide us the document for AFE5832ZBV?

Thanks,

Yinhui 

Previously, our work is based on AFE5832LP 9CZLE53. The one for the current board is AFE5832 OBZCGY9.

Hi Yinhui,

Thank you very much for sharing the data.

I suspect that the reason for this unusual behavior is coupling of noise from the high impedance node of the TR Switch. And somehow this is only affecting the odd channels. So may be the odd outputs of the TR switch are floating and hence this couples with the ADC. 

Can you try one small experiment. Disable all the TR switch outputs by selecting all the TR_SW_DIS bits in TX7332 and observe if the noise couples on the even channels as well. If that happens we can root cause it to TX7332.

Regards,

Abhishek

Hi Yinhui,

The devices AFE5832 and AFE5832LP have common ADC die but different VCA die. 

The AFE5832LP's VCA die is different and is optimized for low power consumption, hence the name LP. So the complete VCA register map shall be different with respect to AFE5832. 

So, please modify your settings accordingly. 

But I don't believe that this might be the reason of the noise.

Thanks & regards,

Abhishek

Hi Abhishek,

Thanks for the suggestion. However, after we tried to disable all the TR switches, the problem we encountered was still the same. We still don't observe the noise from the signals of even channels, which is still there for odd channels.

Yinhui

Hi Abhishek,

We are already in the process of replacing the AFE5832 with the AFE5832LP on our board. I will let you know the result once we have further update.

Thanks

Yinhui  

Hi Yinhui,

Thanks for performing the test.

Let me check with my team on this and get back to you.

Thanks & regards,

Abhishek

Hi Yinhui,

That's great. 

Glad to hear that it finally is solved.

Thanks for your patience and understanding. Do reach out again for any future support.

Regards,

Abhishek