THS4541: Noise Analysis Results THS4541

Thomas Behnke

Part Number: THS4541
Other Parts Discussed in Thread: THS4551

Dear E2E Team!

I simulate AC, Transient and Noise Analysis of THS4541 using LTSpice. DS provides 2.2nV/sqrt(Hz) spectral density.

All resistor less than 1KOHM. Simulation shows a noise spectral density of 90nV/sqrt(Hz) which is much higher as expected.

in the spice model i add 10MEG resistors according "THS4541:ThePSPICE model for ths4541dose not work" discussion.

I found also a THS4551 spice model. But simulation provides error message in any simulation.

the log file out put is gieven:

Circuit: * H:\MyAppNotes\Simulation\LTSpice\CircuitAnalysis\MBOLO\fda_mbolo&ADC_THS4551.asc

WARNING: Node U2:I0:I0:I21:1:O is floating.
ERROR: Node U2:I0:I0:I32:NET8 is floating and connected to current source G:U2:I0:I0:I32:RCMFBAVG1

WARNING: Less than two connections to node U2:I0:I0:I21:1:O. This node is used by R:U2:I0:I0:I21:1:3000.
u2:i0:i0:i32:vprobe: Missing value, assumed 0V @ DC
u2:i0:i0:vprobe1b: Missing value, assumed 0V @ DC
u2:i0:i0:vprobe1a: Missing value, assumed 0V @ DC
u2:i0:i0:vprobe2a: Missing value, assumed 0V @ DC
u2:i0:i0:vprobe2b: Missing value, assumed 0V @ DC
u2:i0:i0:vprobe4: Missing value, assumed 0V @ DC
Direct Newton iteration for .op point succeeded.
Heightened Def Con from 0.000363366 to 0.000363366

Date: Mon Jan 09 15:07:20 2023
Total elapsed time: 49.607 seconds.

tnom = 27
temp = 27
method = modified trap
totiter = 812984
traniter = 812967
tranpoints = 406490
accept = 398937
rejected = 7554
matrix size = 357
fillins = 744
solver = Normal
Thread vector: 42.3/18.1[8] 5.0/8.8[1] 14.0/15.5[1] 1.7/7.1[1] 2592/500
Matrix Compiler1: 50.14 KB object code size 14.5/7.9/[4.0]
Matrix Compiler2: off [5.5]/6.1/54.2

Any help is welcome! with my best regards Thomas

over 2 years ago

0 Michael Steffes over 2 years ago

Guru 48395 points

Hey Thomas,

You are operating at pretty high gain, so that 2.2nV input noise is getting a lot of gain to the differential outputs. Also, you should be looking at the differential outputs, not just one side

And putting in your numbers to a noise calculator tool I have, should see about 135nV/rootHz across the output pins,

Which if you take 1.41*90nV you are seeing now on one side, about 127nV - pretty close.

It is likely the THS4551 model has statements not recognized by LTspice, this comes up regularly.

0 Thomas Behnke over 2 years ago in reply to Michael Steffes

Prodigy 60 points

Dear Michael

With RF = RG = 1KOHM the gain is near 1V/1V. I double checked it and add my plot.

Also the thermal resistor noise is not so high, all resistors <= 1kOHM.

I add also my spice model (annex) for further investigation. As I mentioned I add R1000 to R5000 in the used model.

8865.THS4541_Pspice.CIR

Best regards

Thomas

0 Alec Saebeler over 2 years ago in reply to Thomas Behnke

TI__Genius 17306 points

Hello Thomas,

Thank you for adding your SPICE circuit to the thread. I will make sure I take a look at your circuit this week; thank you for providing the updated information.

Best,

Alec

0 Thomas Behnke over 1 year ago in reply to Alec Saebeler

Prodigy 60 points

Dear Alec

do you have any news about the noise performance of the HS4541? Please contact me for any input you need!

best regards

Thomas

0 Jacob Freet over 1 year ago in reply to Thomas Behnke

TI__Expert 8400 points

Hi Thomas,

Apologies for the slight delay, but it is a holiday in the US today. We will send over a response to you question by tomorrow.

Best,

Jacob

0 Thomas Behnke over 1 year ago in reply to Jacob Freet

Prodigy 60 points

Hello Jacob, Alec

Could you make any investigation concerning my simulation results of THS4541 noise floor ?

Do you have a alternative simulation model of THS4541 ?

Any input is welcome for clarification. Thank you in advance!

Best,

Thomas

0 Alec Saebeler over 1 year ago in reply to Thomas Behnke

TI__Genius 17306 points

Hello Thomas,

I am working with your TI FAE contact, Adrian, to confirm our solution and advice to you on this matter. We are meeting to discuss this Monday January 30th.

I have files from Adrian to look over today & tomorrow.

Best,

Alec

0 Michael Steffes over 1 year ago in reply to Alec Saebeler

Guru 48395 points

Just out of curiosity, I tested teh July 2018 model update for noise, here is the PDS plot we did back at release,

Here is the input current noise, the flatband is correct, but the 1/f corner is too low, getting 5pA at 10kHz instead of 10pA

And then the input voltage noise, looks pretty darn close to Figure 39 measured data,

There is an initial release model from 2014 that I also have, but this looks pretty adequate to me.

And this last file,

THS4541 2018 model input voltage noise sim.TSC

+1 Thomas Behnke over 1 year ago in reply to Michael Steffes

Prodigy 60 points

Dear Michael

thank you for the valuable information! I will continue my investigation based on your input!

I will try to extract the THS4541 spice model from your tsc file.

best regards

Thomas

0 Thomas Behnke over 1 year ago in reply to Thomas Behnke

Prodigy 60 points

Thank All for any help and clarification. Issue resolved completly. best regards Thomas

Amplifiers

Amplifiers forum

THS4541: Noise Analysis Results THS4541