AFE5809: Ghost signals.

Hi Poul,

I'm sorry to hear about your issue.

Based on your description, its a little difficult for me to point out the issue right away. I would appreciate some more information:

1. Can you share the signature of the signal, an image would be better to understand the issue better. 

2. Do you have the raw data? if yes, please share, we can look at any pattern or code breaks. 

3. Also, can you share the operating conditions, like are you using TGC path or CW path, gain settings, etc.?

4. Is the issue reproducible on multiple devices?

Thanks & Regards,

Abhishek

Looking at the image. From the left to the right. Dark green waterpath, then a white plastic film and then a set of light green ghosts of the plastic film. 

We use TGC. *The same level all the way. We don't use CW and we have PGA set on 24 dB , 30 MHz LP filter. 

We have this issue on all units. 

We have measured input level: < 120 mvpp.

I don't have direct acces to raw data. Can be extracted if needed.

Best regards

Poul Pedersen

New information. 

It looks to start when input signal i over 60 - 70 mvpp.

Best regards

Poul Pedersen

Hi Poul

Thank you for the information.

Let me talk with my team on this issue and shall get back to you.

Regards,

Abhishek

Any news?

Hi Poul, 

Abishek is out since there is a national holiday in India. You should get a response by Friday. 

Kind regards,
Nick Z

Hi,

How are you?

According to what you just mentioned at the beginning,

"Using ch8."

Does it mean you are using only Channel 8 and other channels are not using right now?

If this is correct, when you are using AFE5809 EVM GUI to setup AFE5809 device registers,

could you please click and enter "Power Down Settings" tab

and then do the followings?

1) Please click on 
PDN LNA by Channel buttons for CH1 to CH7 but keep CH8 un-clicked.

2) Please click on

ADC PDN by Channel buttons for CH1 to CH7 but keep CH8 un-clicked.

3) Please click on:

LVDS PDN by Channel buttons for CH1 to CH7 but keep CH8 un-clicked.

After that, please run your image again,

and see if you can get better result or not?

Thank you for using TI AFE5809 device.

Best regards,

Chen

Only CH8 is in use. We run a single element.

We are not using the AFE5809 EVM GUI for setup. We have a FPGA with a Picoblaze micro to setup the AFE5809 via the SPI bus. 

We just found our old first "pre prototype" of the board. Here we hardly can see the problem, if it is there at all. Here the AFE 5808 has a production code: 75ZNF43. On our newly produced prototypes, where we clearly have a problem, we have a production code: 81ZLR93. We also suffer from less gain in the on the boards using chips with production code: 81ZLR93.

Hi,

Please look at our TI AFE5809 Data Sheet page 52, page 58.

you can see these register maps.

for ADC, you can set register 0x01[8:2] to be '1'.

for LVDS, you can set register 0x0x02[9:3] to be '1'

for LNA, you can set register 0x35[6:0] to be '1'

Thank you!

This did not solve my problem. 

Hi Poul, 

I have a few questions to better understand your problem - 

1. You mentioned that your first prototype did not have this issue in the exact same use case (same setup, same film, same transducer, same register settings). Is that correct? Which part did you use in that system - AFE5808, AFE5808A, or AFE5809? What are the other differences between your first prototype and the current system? 

2. When you performed this experiment, how far was the plastic film? Are you sure this is not a reverb? One way to check this would be if receive time of actual signal is T, then first reverb would be at 2T, second reverb would be at 3T. 

3. Can you tell me your LNA setting (gain, input clamp setting, active termination setting, HPF setting) and V2I_CLAMP setting. Can you try changing these settings one by one and see if it improves your performance. That might also give us a clue on the root cause. 

Thanks,

Karthik