TPS61170: Internal Error amp transfer function

Hello,

Is there any update?

It is urgent matter, we are waiting TI comment.

Regards,

Mochizuki

Hi Mochizuki,

I am not aware of such behavior in the error amplifier. could you share the VIN, SW and VOUT waveform to show your observation.

Hi Jasper,

Thank you for your reply.

We will capture VIN, SW and VOUT waveform to show our concern on TPS61170-280EVM.

Beside of that we will apply VIN 4.9V with +/-100mV 100Hz modulated power source to see VOUT noise.

I hope it will duplicate our customer's problem.

Regards,

Mochizuki

Hello Jasper,

Please find attached waveforms on TPS61170EVM-280 original configuration Vout=24V.

When input 4.8V Vin with 100Hz 100mVpp modulated power source, we can clearly see spike noise which related to Magic Vin voltage.

Our customer is facing noise problem on high gain analog amplifier application with TPS61170 for battery operated power.

If battery voltage is reduced to 4.8V, 20mVpp noise is generated.

Power Supply Ripple Rejection ratio is less than 20dB =20mVpp/100mVpp

It seems this phenomena come from TPS61170 internal function.

How can we solve this EE problem?

Regards,

MochizukiTPS61170EVM eval.pdf

Hi Mochizuki,

thanks a lot for the detail waveforms. firstly of all, I will try to find the root cause. secondly, is the PSRR of this device good enough without such behavior? for example, is the PSRR good enough if the input voltage is 4.6V.

one possible solution to improve the PSRR is to reduce the compensation capacitor. the capacitor connected to the COMP pin.

Hi Jasper,

Thank you for your prompt reply.

PSRR is good enough at below 4.6V or more than 4.8V Vin voltage, just across this Magic voltage range TPS61170 generate spike noise.

It seems this is not normal "closed loop error correction function", this spike noise is created by internal error amp itself.

Would you try to find "Error amp power supply architecture" in TPS61170 design/system specification.

We assume there is error amp power supply rail switching function at this Magic Vin voltage.

Regards,

Mochizuki

Hi Mochizuki,

it seem like the power supply of the EA changes during this special period. I will try to find the root cause, but as it is old device, i'm not sure if I can find the original designer.

even we find the root cause, we still need to find the solution as the IC can't be changed.  small Cc to reduce the spike noise during this period. the spike noise may be not impact the system if it reduce to 5mV (for example). what is the VIN voltage range and the requirement on spike noise?

Hi Jasper,

We knew TI can't change this characteristics from now, and it is within datasheet specification.

At most of your customer could use this device without signal chain noise issue.

Cc value had been optimized, but still have spike noise. Also it create 1msec cycle hunching spike noise phenomena.

Because of power source impedance, Vin is go and back High and Low Hysteresis voltage.

This time we plan to add RC filter (decupling) to Vin pin, to have isolation from Inductor current path to avoid this hunching noise.

Anyway our customer wants to know root cause mechanism of this spike noise on the silicon.

Regards,

Mochizuki

Hi Mochi

it is caused by internal LDO which is used to power the control. The maximum input voltage of this device is up to 18V, but the control circuit is normal lower than 5V. that is why we need this LDO. the LDO may have a circuit change at this voltage range. (in high input voltage, a NFET is good enough. but for low input voltage, PFET is better to regulate the LDO output). this change leads to such VOUT transient.