OPA1642: Improve noise and DC offset performance

Andrew Gong

Part Number: OPA1642
Other Parts Discussed in Thread: TLE2426, , OPA1656, JFE150, JFE2140, OPA1612, LME49720, OPA891

Dear TI E2E support, Hi there,

Our team made a prototype of a phono-stage amplifier (pic. 1) base on the circuit in OPA1642's datasheet. The amplifier is powered by 2 x 9V batteries for low noise operation. We used a TI TLE2426 (SO8) to split the rail so that it can avoid unbalance rail voltage due to battery voltage difference. Here are some observations and we are seeking for comments / suggestion to enable it can be released as an outstanding design, basing on TI SoundPlus, note that we are using a metal shield can for further reducing interference (if any) to the circuit performance.

If we replace an OPA1656 instead of OPA1642. Noise like scratching something or squeezing paper (which is very low frequency) can be heard if the volume is turned up to max level. But it seems ok for the OPA1642. Is there anything we had overlooked for the OPA1656 ?? (Even the input terminal were shorted)
Little hiss noise still can be heard if turn up the volume to max level in both OPA1656 and OPA1642. S/N ratio is good but not excellent. Is it reasonable to add low noise FET input stage like JFE150 or JFE2140 ??
Is bipolar opamp like OPA1612 easier affected by EMI due to rectifying effect?
If C1 - C6 and C7 - C9 present at the same time, TLE2426 will draw about 25mA current that I think is too much. The idle current of the whole amplifier will drop to about 5 ~ 6mA if I remove C1 - C6 or C7 - C9 either. Is it a normal behavior of TLE2426? (since the C7 - C9 is for noise reduction according to the datasheet of TLE2426) Should I eliminates C1 - C6 for longer battery life ???
If C22 - C25 was shorted, there will be about tens of mV DC present at the output terminal. Is there any way to reduce this offset?
Is it a good idea to remove C22 and C23 ?? And is it true that the distortion performance will be improved if we use Non-polar ECap for C22, C23 instead of Polar ECap ??

Regards

10 months ago

0 Chris Featherstone 10 months ago

TI__Mastermind 23935 points

Hey Andrew,

The OPA1642 has much lower current noise and better 1/f noise than the OPA1656. For high source impedance sensors such as a phono cartridge lower current noise is better. The current noise paired with high source impedance translates into additional voltage noise on the input and this gets amplified by the gain in the design.

For further detail we have a great app note on the subject:

Impact of Current Noise in CMOS and JFET Amplifiers (ti.com)

JFETs offer better 1/f noise performance than CMOS devices and also provide the benefit of high input impedance that doesn't load the sensor in the way that a bipolar input stage does. Bipolar input stages have very low voltage noise but typically have high current noise. For this reason you can replace the front end of a bipolar input stage with a discrete JFET such as the JFE150 or the JFE2140. I have written two application notes and created evaluation boards that are available on the circuits. Here are some links to this collateral.

JFE150:

https://www.ti.com/lit/an/slpa018/slpa018.pdf?ts=1714069240193&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FJFE150

https://www.ti.com/tool/JFE150EVM

JFE2140:

https://www.ti.com/lit/an/sboa563/sboa563.pdf?ts=1714069287386&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FJFE2140

https://www.ti.com/tool/JFE2140EVM\

The discrete JFET solutions are susceptible to power supply noise. For this reason the discrete JFET solutions will need a clean power supply such as the batteries.

Other useful resources:

ti.com/lit/ab/sboa355/sboa355.pdf?ts=1714069348265

For capacitor types and THD effects please see Ti's John Caldwells article on Signal Distortion from high-K ceramic capacitors

https://www.edn.com/signal-distortion-from-high-k-ceramic-capacitors/

What type of capacitors are in the signal path? Any X7R's? For large capacitor values that cannot be found in the COG type you can use electrolytics, tantalum or film.

Stress induced outburts: Microphonics in ceramic capacitors:

https://e2e.ti.com/blogs_/archives/b/precisionhub/posts/stress-induced-outbursts-microphonics-in-ceramic-capacitors-part-2

I hope this information is useful.

Best Regards,

Chris Featherstone

0 Bubble Club 10 months ago in reply to Chris Featherstone

Prodigy 10 points

Hi Chris,
What an attractive product from Andrew !! As I had made a similar project using the LME49720 which is a bipolar opamp, it does have EMI problems especially affected by mobile phones or WiFi routers. Is it true that Bipolar devices are more likely affected by EMI?? Is it make sense to use metal shield like Andrew does?? (For my LME49720)
Best Regards

0 Chris Featherstone 10 months ago in reply to Bubble Club

TI__Mastermind 23935 points

Hello,

EMI is a complex topic. Proper shielding can help. You can see some comparisons in the app brief below. Quoting from the app brief, "Over the past 7 years, Texas Instruments has taken the initiative to include EMI/RF filters in all the op amps making them less prone to errors often emanating beyond the signal chain."

https://www.ti.com/lit/an/sbot041/sbot041.pdf?ts=1714147414438&ref_url=https%253A%252F%252Fwww.google.com%252F#:~:text=Over%20the%20past%207%20years,instrumentation%20amplifiers%20like%20the%20INA188.

Doing a cross reference search on the LME49720 I found the OPA891 which is a modern replacement. Have you considered upgrading to a modern Op amp?

https://www.ti.com/lit/ds/symlink/opa891.pdf?ts=1714148108087&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FOPA891

Best Regards,

Chris Featherstone

Audio

Audio forum

OPA1642: Improve noise and DC offset performance