Part 4 of R&S NGP814 Power Supply - Review

This part of the review focuses on transient behaviour, the Rigol DL3021 electronic load is used to generate the various load conditions used in the experiments.

The table below is from the NGP814 datasheet, the T90 rise time value is below 10ms.

To get a feel of the rise time over the rated power range the following tests were done using Channel 1 at various percentages of the rated power (200W).

[Note: all measurements were taken at 1M samples/Sec and a 16bit ADC]

All the tests were generated using CSV files that were then uploaded to the device as seen in the photo below:

1. Voltage time response

Test A @P = 90%
• P = 180w
• V = 32V
• I = 5.625A
• Tcycle = 4 seconds

Trise = 0.003826 = 3.826 mSec

Tfall = 2.419527 - 2.41781 = 1.717 mSec

Test B @P = 50%

• P = 100w
• V = 32V
• I = 3.125A
• Tcycle = 4 seconds

Trise = 3.994 mSec

Tfall = 2.915 mSec

Test C @P = 1%
• P = 2w
• V = 3.2V
• I = 0.625A
• Tcycle = 4 seconds

T(rise) = 0.497 mSec

T(fall) = 0.213 mSec

• P = 100w
• V = 16V
• Tcycle = 4 seconds

Zoom in is shown below:

Trise = 1.852 mSec

Tfall = 037.65 mSec

2. Current time response

Test E @P = 6.4%
• P = 12.8w
• V = 32V
• I = 0.4A
• Tcycle = 4 seconds

Zoom in is shown below:

Trise = 0.02 mSec

Tfall = 0.075 mSec

Test F @P = 32%
• P = 64w
• V = 32V
• I = 2A
• Tcycle = 4 seconds

Trise = 0.165 mSec

Tfall = 0.073 mSec

Overshoot = 0.524 A

Observations:

• Rise time is proportional with amplitude.
• Total power delivered does not affect rise time (i.e. for the same voltage rise time is roughly the same even if current value changes and vice versa).
• Test D proves the extended fall time documented in the datasheet at No Load condition, the results still fall within "<50 mSec", the changes in fall rate can be seen circled in yellow, this could be due to reactive components discharge response charts.
• Very high current PO is shown in Test E, load was constant resistive load. However, I noticed that at higher currents the PO decreases. This needs further investigation to void damaging sensitive components when powered by the power supply. (When I tested at 2.5A the overshoot burned the fuse of the DMM6500 input, the fuse is rated at 3A, 250V, 235 series, fast acting).
• The same behaviour was observed whether the current value changes were set by the NGP814 or by the load by changing the resistance. (tested with an electronic load with a slewrate of 0.25A/usec, and by simply flicking a mechanical switch between resistors).

Notes:

• These tests do not cover all cases, but aim to show certain selected scenarios and traits.
• Accuracy is limited by sampling rate, and ADC (quantization, bit rate and noise).
• Results will need to be verified again using higher sampling rate oscilloscope, will update in the near future.

Overall time response of the NGP814 exceeds expectations and well within the range documented in the datasheet "<10ms" (highest measured was around 4ms).

Only concern is the percentage overshoot of the current level.