Error in LTspice with ADA4932

Hi 

May I ask if this is where you were able to get the ADA4932 spice model
ADA4932-1 Datasheet and Product Info | Analog Devices?

I downloaded the model from the specified URL, but the result is the same.

Hi John Kevin,

Does your post mean my problem doesn't occur with the latest models?
I am simulating with a 5V power supply specification, but it seems to work with a ± 5V specification.

Please check it.

Hi  Takahiro,

Thank you for your patience. I asked some colleagues who are experts o this subject matter to look into it. here's what I got.

"The model is functioning as intended. In the customer configuration there is always a large differential input voltage due to the +0.6V and -0.6V offsets on Vp and Vn. This is duplicated in attached circuit “ADA4932_0.6Vrefin.asc.” This shifts the midpoint of the differential output voltage to +1.2V and causes the saturation to +Vs-1.4V on Out+."

20211108_Error_in_LTspice_with_ADA4932.zip

Based on p. 25 of the datasheet and Figure 65, it seems like the ADA4932 is designed to handle single-ended input voltages centered at 0V. With the standard feedback network, the voltages at the input terminals are biased up to a level in the allowed common-mode voltage range and the differential input voltage is centered at 0V. This allows the output to swing symmetrically around Vocm and achieve the maximum output swing range. The attached circuit “ADA4932_0Vrefin.asc” gives an example.

Thanks for the answer.

I am asking about the case of 5V OPERATION.
According to the DataSheet (Rev. E | Page 5), Output Voltage Swing is defined as +Vs-1.15V for 5V specification.
If I use "ADA4932_0.6Vrefin.asc." at 5V, shouldn't it be saturated at +Vs-1.15V?

Hi again  Takahiro,

I got more inputs from the team.

You are right. The ADA4932 model is matching the Min instead of the Typ.

It is being adjusted to match the Typ for both 5V and 10V supply.

The updated model will soon be available in the library. Once available, you will need to do Sync Release.


Some details of the discussion about the model.

The model has two areas of improvement: 

• Output voltage swing is modeled to the minimum, not the typical which we usually target
• Output voltage swing does not change with supply voltage

The model was adjusted to handle this better. Now the output swing is within 1.2V of the rail at 10V supply and within 1.02V of the rail at 5V supply. Other valid supply voltages from 3V to 11V are linearly interpolated.

Example with 5V supply (±2.5V) – output can swing within 1.02V (white lines)

Example with 10V supply (±5V) – output can swing within 1.2V (yellow lines)

Hi John Kevin,

I understand that a model that matches each power supply specification with standard values will be released.
Thank you for your cooperation.

Takahiro

Hi  Takahiro,

please be advised that an updated model of ADA4932 is already available in the LTspice library. Please do Sync Release to update your library.