THS4551: Linearity test issue

Vincent Chen64988

Part Number: THS4551
Other Parts Discussed in Thread: TINA-TI

Hi Team,

My customer is evaluating THS4551's linearity performance but we found some issues and need team's comment.

1. Below is customer's circuit, this gain is 1k/6.05k and is that correct?

2. Please find below measurement result, the gain error is quite large when Vin voltage is small, say about lower than 1V. Is the result reasonable? How to improve the linearity? (Asked customer to use multimeter to measure the real Vin voltage, when DC supply is 10mV, the measured Vin voltage is 10.185mV. A small error won't cause large gain error)

THS4551線性測試.xlsx

3. Since it is not tested on TI's EVM but connected by long wire, is it possible to cause the error?

Thanks a lot.

VIncent Chen

over 6 years ago

0 Jacob Freet over 6 years ago

TI__Expert 8680 points

Hi Vincent,

Yes, your board may certainly be causing issues. The THS4551 is a high speed device and should follow high speed layout guidelines as listed in the datasheet's layout section. The board you are using very likely may be causing the part to oscillate, and the actual circuit may be making it somewhat unstable as well. I would suggest running a stability analysis in TINA-TI

Regards,

0 Vincent Chen64988 over 6 years ago in reply to Jacob Freet

TI__Expert 4480 points

Hi Jacob,

Thanks for the comment. Yes, I did a DC analysis and found there just is a litter bit gain error. I will keep working with customer to figure out this issue.

Vincent Chen

0 Michael Steffes over 6 years ago in reply to Vincent Chen64988

Guru 48395 points

Also, running the THS4551 as an attenuator is fine but risks low phase margin even with good layouts - easy to fix with a cap across the inputs as I explained in section 9.1.6 in the datasheet.

Those long wires also will cause a problem, to run LG phase margin sims - use the approach shown here, This approach includes the interaction between the open loop output impedance and the feedback network - what is shown in the Precision labs for stability sims breaking the loop at the output misses that important effect.

https://www.planetanalog.com/author.asp?section_id=434&doc_id=564934&

0 Michael Steffes over 6 years ago in reply to Michael Steffes

Guru 48395 points

I had just finished a LG stability article with TINA circuits using the THS4551 in an MFB filter, just modifying that to your circuit shows 44deg phase margin, which is good - no changes required, just tighter layout - try to do this circuit on EVM board. He is that file I just ran showing how to set this up (and the MFB elements zeroed out).

THS4551 attenuator loop gain.TSC

0 Vincent Chen64988 over 6 years ago in reply to Michael Steffes

TI__Expert 4480 points

Hi Michael, Jacob,

My customer got the EVM board and test it, the result is quite better than last experiment. But we encountered Vout clamp issue.

Below is customer's test result. In the sheet of "0~+12V", you can find the value of Vout+(green) is clamped at about 4.377V but datasheet shows the min. voltage output(high) is (VS+) – 0.23, here the VS+ is 5.0V. Could you please help to explain why we encountered this issue.

THS4551 EVMBD 線性度測試.xlsx

Thanks very much.

Vincent Chen

0 Vincent Chen64988 over 6 years ago in reply to Vincent Chen64988

TI__Expert 4480 points

Hi Team,

Is there any update for this question?

Thanks very much.
Vincent Chen

0 Michael Steffes over 6 years ago in reply to Vincent Chen64988

Guru 48395 points

Hey vincent, Just running a TINA DC test on this would be helpful looking at the input summing junction pins. with 5V supply and output at 2.5V, a 12V input to the input resistor is right at the edge of the upper limit on the input voltage range. so it might be limiting at the input. the sim would show that easily for you.

Amplifiers

Amplifiers forum

THS4551: Linearity test issue