afe5808 and tx810

it should be OK to put a diode before the AC coupling cap of AFE5808. By doing that, you may see some addtional power consumption if TX810 output has a offset

one thing i want to mention: AFE5808 is designed for high performance systems with extremly low noise. TX810 is designed for ultraportable system, in which noise is not a key spec. the combination of these two parts is interesting. 

Please make sure you use AFE5808A instead of AFE5808. AFE5808A is recommended for new designs. both devices are completely comptable. 

Thanks!

Xiacohen

Can you comment on how the afe5808a will perform differently than the afe5808?  We have boards in eval now with the afe5808 and it's not practical to change them.  We also have purchased afe5808 inventory for the next board build and need to decide if we can use them.  Can they be exchanged for the A parts?

We are currently having problem with a long transient recovery time when the tx810 is switched to receive mode.  Though the actual signal transient is only several hundred nsec, the afe output data has a resulting transient that lasts 5.3usec, which is too long.

Thanks.

afe transient in samples at 12MHz Fs:

I have also noticed that recent afe5808 data sheets have changed the input termination cap recommendation from 100nF to 1uF.  Can you comment on what issues we will see with 100nF terminating caps?

Thanks.

we can compare 1uF to 0.1uF.  Vchange on Cap depends on Cap value. The large cap the smaller Vchange. As you can see, small Vchange is always preferred.  considering the board space, custoemrs may not be able to use too larger caps. so we feel 1uF is a good optimaiton.  when the Vchange is small, it helps stablize the input common mode voltage, which is always good.  Thanks!

David, 

the inductor you plan to use should help the recovery time. 

can you leave you contact info? either phone or email. we can discuss details 

Thanks!

Hi,

We found that with the TGC gain set to max (Vcntl=-1.35V, min atten) the transient pulse is much smaller.  Perhaps the PGA section gets saturated and takes time to come out.

You can email me at dburton@siriusdt.com.

Thanks

How important is the matching between these caps?  i.e. is +/-20% OK or does that cause a significant difference in the ch-ch responses?

Thanks.

matching on caps doesn't matter much. Thanks!

Hi Xiaochen, I made the same choice for my first test board (AFE5808A and TX810). I don`t understand why to put the diode before de Ac coupling cap, what kind of improve this diode cause?

Regards,

Santiago.

as you may know, discrete diodes have lower clamp votlage.  the main purpose is to reduce the large signal reaching the AFE5808A.  

Is this because the clamp diodes inside TX810 are not designed to the 1Vpp linear input range of AFE5808A LNA? So, I need to put externall clamp diodes?

yes. TX810 output can reach 4Vpp including the spikes. thus the external diodes are needed for AFE5808A. Thanks!

Hi,

   Sir,I am working for portable ultra sound system design.

How the working of TX810 IC is controlled with B1/B2/B3 control inputs.Such that the switch will off in transmit mode  and on state in recieve mode.About The control inputs in data sheeet mentioned that to monitor power requirements,So please suggest me to control theTX810 IC.

I am using Transmit Beam former is LM96570 and Analog front end is AFE5808 ,  HV pulser LM96550. 

waiting for your favorable solution,Thank you.

Chandu,

The control switches only control the bias current through the switch to control the output pulse amplitude.  There is no timing control needed with the TX810, you do not need to time the transmit mode and receive mode because they are the same mode.  The TX810 will automatically switch modes.  The control inputs are unrelated to mode switching.

Thanks,

Chuck Smyth

Chandu,

If you are setting the bias current to zero to isolate the afe during transmit pulse you need to be aware that when you turn the bias current on it will inject a bump into the signal that will transmit to your transducer corrupting your transmit pulse.  For this reason we are looking at leaving the TX810 bias on throughout the transmit/receive time period.  I have used all the same parts as you except a different pulser IC.

Regards.