BUF802: Spice Model error

Jean-Jacques Aubry

Part Number: BUF802
Other Parts Discussed in Thread: TINA-TI,

After loading the "BUF802 TINA-TI Reference Model" from the website into TINA 9, and trying to run the analysis (DC or AC) of the circuit, I get an error related to the Model

Is there an update to the Spice Model of the BUF802 circuit?

over 2 years ago

0 kai klaas69 over 2 years ago

Guru 140200 points

Hi Jean-Jaques,

I get the same error message

I have the version: "Version 9.3.150.4 SF-TI"

Kai

0 Shreenidhi Patil over 2 years ago in reply to kai klaas69

TI__Expert 3126 points

Hello All , can you send me the TINA file

0 kai klaas69 over 2 years ago in reply to Shreenidhi Patil

Guru 140200 points

Hi Shreenidhi,

here it comes:

jean_buf802.TSC

Kai

0 Jean-Jacques Aubry over 2 years ago

Prodigy 10 points

Deleting one of the two lines starting with "R_R2" allows ( seems) the model to work properly!

0 kai klaas69 over 2 years ago in reply to Jean-Jacques Aubry

Guru 140200 points

Hi Jean-Jaques,

but the frequency response seems to look different:

Kai

0 Jean-Jacques Aubry over 2 years ago

Prodigy 10 points

There is another problem with the model concerning the maximum output voltage.
I have modified the schematic with
Vch = 6V, Vcl = -6V and VG1 = 3.4V sine
With the transient analysis the output signal is clipped at about +2.9V and -2.9V, in contradiction with the datasheet.

0 Sean Cashin over 2 years ago

TI__Genius 14066 points

Hi Jean-Jacques,

At the last minute, I had added a 1G ohm dummy resistor at the input pin to help the model converge for a current input, specifically for a capacitor bootstrap application. I added it to the netlist directly as "R2", and unfortunately did not see that there was already a reference designator "R2" in the PSRR circuit. You can change

R_R2 MID IN+ R_NOISELESS 1E9

R_R100 MID IN+ R_NOISELESS 1E9, or use the attached.

BUF802_GWL.TSC

I am updating the BUF802 model based on updated measurements I have taken using our low-parasitic third EVM revision, which should be available on ti.com next month. This model already contains some of those updates, so please disregard the previous models' validation curves. I am updating them as they become available.

Kai,

The original 2GHz peaking was found to be largely due to the first EVM's trace inductance, not device input inductance. The new model doesn't include that ~1nH input inductance, so its frequency response looks less resonant than the response in the embedded image.

Thank you for you patience as I compile my latest testing data for this new device, I assure you that the finished model will be great.

Best regards,

Sean

Amplifiers

Amplifiers forum

BUF802: Spice Model error