Post Go back to editing

AD9690 & AD9697 Analog Input max swing and overload recovery.

Category: Datasheet/Specs
Product Number: AD9690, AD9697

Hello,

We would like to use AD9690 & AD9697 in oscilloscope front end. With an oscilloscope you are often looking at a small portion of a large waveform and so the signal will be over range for a significant part of the time and must recover quickly when the signal comes back into range. So I have 2 questions;

1. What is the maximum input range for the ADC which will prevent long term damage and also ensure quick recovery?

2. What is the recovery time from overload to <1 LSB?

1. On max input range we have:

AD9690 data sheet, rev B, page 11 Table 6 ABSOLUTE MAXIMUM RATINGS: VIN± to AGND = 3.2V. Also page 28:

"The absolute maximum input swing allowed at the inputs of the
AD9690 is 4.3 V p-p differential. Signals operating near or at
this level can cause permanent damage to the ADC"

That would be 2.05Vcm+/-(4.3/4)V= 0.975-3.125V.

AD9697 data sheet, rev 0, page 11 Table 6 ABSOLUTE MAXIMUM RATINGS: VIN± to AGND − 0.3 V to AVDD3 + 0.3 V = -0.3 to +2.8V. Also page 23:

"The absolute maximum input swing allowed at the inputs of the
AD9697 is 5.6 V p-p differential. Signals operating near or at
this level can cause permanent damage to the ADC"

- That would be 1.41Vcm+/-(5.6/4)=0.01-2.81V.

So a few issues here:

The AD9697 seems clear that max input range is 0-2.8V from both table 6 and page 23. But for the AD9690 the range is limited to 0.975-3.125V. A smaller range for the device that has a higher AVDD3 (3.3V vs 2.5V)? Is it because the AD9690 input can be 50-400ohm? Whereas the AD9697 is fixed at 200ohm. And if so, what is the AD9690 limit when set to 200ohm?

Second, "Signals operating near or at this level can cause permanent damage". That is a bit ambiguous. What is the safe long term range than can be used?

2. Overload recovery time.

I could not find any spec in the data sheet for overload recovery time on the analog input. It is normally spec'd as time to settle to 0.1%, but would also be good to know time to settle to 1 LSB if that is possible.

Thanks for any help!

Best regards, Ken

Parents
  • A smaller range for the device that has a higher AVDD3 (3.3V vs 2.5V)?

    The AD9690 and AD9697 are different designs and hence the mechanism that causes an overvoltage type scenario is different. This is why the voltage limits are different.

    "Signals operating near or at this level can cause permanent damage". That is a bit ambiguous. What is the safe long term range than can be used?

    I dont see any ambiguity there. the ABS MAX ratings table is a 10year lifetime reliability spec. It means that we can guarantee safe operation of the converter at the level at the max junction temp. however, as you approach this spec or cross it, it could affect the lifetime reliability of the device. 

    I could not find any spec in the data sheet for overload recovery time on the analog input.

    please see Out of range recovery time specification in the datasheet

Reply
  • A smaller range for the device that has a higher AVDD3 (3.3V vs 2.5V)?

    The AD9690 and AD9697 are different designs and hence the mechanism that causes an overvoltage type scenario is different. This is why the voltage limits are different.

    "Signals operating near or at this level can cause permanent damage". That is a bit ambiguous. What is the safe long term range than can be used?

    I dont see any ambiguity there. the ABS MAX ratings table is a 10year lifetime reliability spec. It means that we can guarantee safe operation of the converter at the level at the max junction temp. however, as you approach this spec or cross it, it could affect the lifetime reliability of the device. 

    I could not find any spec in the data sheet for overload recovery time on the analog input.

    please see Out of range recovery time specification in the datasheet

Children
  • Hi UmeshJ,

    Thanks for the answers and pointing me to the 'Out-of-range Recovery Time'.

    On the issue of damage to the device I understand what is meant by 'operating at this level', but what is meant by 'near this level'? How much is 'near'? 10mV, 100mV...

    In other words, how much below this level do we need to operate to be safe?

    e.g. for the AD9690 the limit is 4.3Vpp-diff. How much below this do we need to operate? 4.2Vpp, 4.0Vpp... A clarification of what is the limiting factor here would be helpful.

    I need to know at what level we need to clip the over range signals for safe operation.

    Thanks, Ken

  • On the issue of damage to the device I understand what is meant by 'operating at this level', but what is meant by 'near this level'? How much is 'near'? 10mV, 100mV...

    this is really a difficult question to answer. the adc is already overranged and you will be getting junk data. As for how close to the abs max you can operate, this is very arbitrary and i would recommend that you reduce the overranging as much as possible, using AGC techniques. the ADCs you have outlined have a feature called fast detect, which can be used to quickly control the gain of the amplifier. this will allow you to reduce the exposure of the adc inputs to high voltages. 

    "for the AD9690 the limit is 4.3Vpp-diff. How much below this do we need to operate? 4.2Vpp, 4.0Vpp" -- i cannot recommend any value here as we dont have any data to prove this. I can only recommend that you try to minimize any exposure at or beyond what is recommended in the datasheet. 

    As i have pointed out earlier, you may use the fast detect feature to control the AGC loop and reduce exposure to high voltages. more information here

    RF-Sampling ADC Input Protection: Not Black Magic After All | Analog Devices see figure 9

  • Hi Umesh,

    Thanks for the reply and the suggestion of AGC. We are using these parts for oscilloscope front end and so cannot use AGC, the gain setting is fixed. And the input waveform to the scope can easily be outside the full scale setting so the only option we have is to clip the signal at the diff amp. I was trying to get a feel for what the clip level should be, but sounds like we will have to put is somewhere between the linear range and abs max range.

    Thanks, Ken

  • Thanks Umesh also for the article. So the issue is FET protection. A good read!