TPS544C25: Current Telemetry Accuracy

Actually you do not specify the accuracy at less than 3A output current. Can you tell me what that is too?

Hi Christopher,

For currents below 20A, you can use a fixed offset of +/-3A, including less than 3A output current. And above 20A, it will be +/-15%.

The datasheet accuracy is without calibration. With calibration, the accuracy will be more accurate, but that accuracy ultimately depends on the accuracy of the calibration. There will still be some tolerance for uncompensated temperature coefficients of up to +/-300ppm/°C.

Can you tell me if there is drift over time and/or temp coefficient drift over time? If I calibrate it initially but the value read back from the TPS544C25 READ_IOUT drifts over time, then my calibration won't do much good some years in the future.

Hi Christopher,

Due to US holiday, our US expert will reply you by Tuesday.

Hi Christopher,

The current telemetry accuracy is not expected to drift over time, and we have not seen appreciable shifts in our accelerated aging qual testing. So you can expect it to be stable over time. However, the behavior versus temperature will not be the same on every unit, so if you are doing a temperature calibration, it needs to be done on each unit separately to improve accuracy. 

I have a few more questions regarding the current telemetry:

1. From our readings at a 0.65A output current, the drift over temperature is much more than 300ppm (see table below). At this low current (0 - 2A output) what is the maximum possible temperature coefficient?
2. What are the maximum differences in temperature coefficient from unit to unit? Could a unit have a positive slope instead of a negative slope? Could the slope be alternating positive / negative over the operating temperature range?
3. Our current readings at +65C show the READ_IOUT values hit 0 at 0.5A, so anything below 0.5A is bottomed-out at 0A. We would like to use the IOUT_CAL_OFFSET register to give the readings the maximum positive offset to avoid this and then correct that offset in our software. Will this strategy work or will it just apply an offset to a bottomed-out value? For example if we applied a 3A offset, would the readings for 0.5A and 0.25A be 3.5A and 3.25A, or just 3A and 3A?

I will check with the team and get back to you tomorrow. 

Hi Christopher,

Sorry for the delay. I could not get the information you are looking for today. So please give us another day.

Christopher Arntzen 

The current sense circuitry in the TPS544C25 is not well suited for accurate current measurements from 0 - 1A for a number of reasons.

1) The current sense error at this level is dominated by offset, not gain errors.  In fact, the current sense offset error could be greater than the full signal value.  Since these are offset errors, a parts per million per degree C characterization of the offset is impractical.

2) The current sense circuitry is only designed to measure positive (Current flowing from GND to SW during the low-side FET on-time) current and reports all current measurements, including analog front-end and ADC conversion noise, less than 0A as 0A.

3) When combined with sample averaging, the clipping of negative current values will produce an offset values that vary significantly over the range of load currents from 0-3A.

As a result of these items, some TPS544C25 devices will report 0A for all 0-1A load currents.  This is a limitation of a 30A device attempting to measure current accurately at 3% of its full load.

For improved current telemetry accuracy in a 1A application, I would recommend considering a TPS40422 controller based design, which uses an external differential current sense that can be set to a much higher value than the internal MOSFET Rdson of the TPS544C25.  The TPS40140 can use precision current sense resistors and accept current sense elements upto 15mΩ.  Even higher values could be used for greater accuracy in a 1A application, but it would require additional post-processing of the PMBus value response to READ_IOUT.

http://www.ti.com/product/tps40422 

This controller with external MOSFETs would also be able to achieve much better efficiency than the TPS544C25 for a 0-1A application by selecting external MOSFETs optimized for lower current applications.

Thank you for that information, it is very helpful. I need to better understand the temperature dependence on the readings and my first two questions (in my list of 3 questions) haven't yet been answered. Can you look back at #1 and #2 and let me know if you have any information regarding those?