Range · Measurement

Reading microamp leakage on a 300 A channel

A 300 A range cannot resolve microamps. How to route a leakage step to a channel that can, and what to record when it gets there.

· 5 min read · All notes

A 300 A channel is built to move charge, not to resolve it. When a test program moves from a 100 A pulse to a 10 µA standby-leakage step, the instrument does not become more sensitive because the setpoint got smaller. The range, the shunt and the accuracy floor stay where they were. Reading microamps on a high-current channel is therefore a routing problem before it is a settings problem: the measurement has to reach a channel or an instrument whose full scale is close to the current you are trying to see.

Start with the accuracy floor, not the ADC

The BTS-6008A-5V300A-H publishes three current ranges — 75 A, 150 A and 300 A — with current accuracy of ±0.05% of full scale and 24-bit resolution. On the 75 A range the accuracy floor is 37.5 mA. On the 300 A range it is 150 mA. A 24-bit converter spreads 75 A over 16,777,216 codes, which is about 4.5 µA per code, but resolution is not accuracy: the published error band is four orders of magnitude wider than one code. A 10 µA leakage current is below that band by a factor of roughly 3,700 on the lowest range.

This is normal. A high-current channel optimises for thermal stability, cable drop and step response, not for nanoamp offset. The useful question is not “can the 300 A channel see 10 µA?” but “which channel in the rack can, and how does the DUT get there?”

Measurement pathFull scale usedPublished accuracy at that FSTypical use
BTS-6008A, 75 A range75 A±0.05% FS = ±37.5 mAPulses, cycling, DCIR above 1 A
BTS-4008Q-5V100mA-124, R10.1 mA±0.02% FS = ±20 nAStandby leakage, coin cells, micro-cycling
External shunt + DMMSet by shuntSet by shunt and meterOne-off checks, reverse-engineering a suspect cell
Electrometer / picoammeternA–µASet by instrumentInsulation resistance, dielectric leakage

Full-scale figures are from the published datasheets. The external-shunt and electrometer rows are measurement options, not product specifications.

Use the low-current channel for the leakage step

The BTS-4008Q-5V100mA-124 is the relevant comparison. Its first range covers 0.5 µA to 0.1 mA, current accuracy is ±0.02% of full scale, voltage accuracy is ±0.01% of full scale, and the datasheet lists a leak current of 0.005 µA with an input impedance of at least 1 GΩ. On the 0.1 mA full scale, ±0.02% is ±20 nA. That is the same order as the quantity being measured, which is what a leakage test needs.

Two consequences follow. First, a 10 µA step belongs on this channel, not on the 300 A channel. Second, the low-current channel is not a substitute for the high-current one: its minimum pulse width is 500 ms and its recording frequency is 10 Hz, so it cannot run the 100 A pulses that precede the leakage step. The practical configuration is a rack with both channel classes, or a switching fixture that moves the cell from the cycling channel to the leakage channel between steps. The switch itself must be a low-leakage type; a wet relay or a contaminated PCB can leak more than the cell.

Range switching and settling

If the leakage step stays on an auto-ranging channel, lock the range for that step. Auto-range logic is designed for the dominant current in a step; a leakage current that sits just below a range boundary can make the channel hunt between ranges, and every change adds a settling interval to the record. The BTS-4008Q publishes a current response of ≤1 ms (10% to 90%), but the measured settling also includes the cable and fixture capacitance. As an engineering rule, allow at least ten times the published response before taking the first valid sample, and discard the samples inside that window.

On the 300 A channel the published response is ≤3 ms and the minimum recording interval is 10 ms at 100 Hz; average a stable window rather than trusting the first point.

Wiring, guarding and noise

Four-wire Kelvin sensing is published for both channel classes. Kelvin removes the force-lead drop from the voltage measurement, but it does not remove leakage paths around the cell. At a 1 GΩ input impedance, 5 V across an unintended path produces 5 nA. Humidity, flux residue, cable insulation and the surface of the fixture are all part of the circuit at this level.

Keep the sense leads short and separately routed from the force cables. Use a guard around the sense node where the fixture allows it, and earth the guard at one point only. Do not move cables during a leakage step: triboelectric and capacitive currents appear as slowly decaying offsets that look like cell behaviour. Allow the fixture to reach room temperature and record the ambient humidity alongside the data.

Thermal EMF is the other slow error. Dissimilar-metal junctions in the fixture generate microvolts per degree, and a temperature gradient across the terminals turns that into a current through the cell and the shunt. Keep the fixture isothermal, use the same metal at both terminals, and run a zero-current baseline before the step.

What to record

Record the current range, the sampling rate, the ambient temperature and humidity, and the duration of the settling window. A 10 Hz record over 60 s gives 600 samples; averaging the last 30 s reduces random noise, but not drift or a real self-discharge trend. Report the mean, the standard deviation and the baseline offset separately. A single microamp number with no noise estimate is not a measurement you can compare across builds.

Takeaways

  • On the BTS-6008A-5V300A-H, the 75 A range has a ±37.5 mA accuracy floor; microamp leakage is not measurable on that channel.
  • Route leakage steps to a low-current channel such as the BTS-4008Q-5V100mA-124, where the 0.1 mA range gives ±20 nA and the published leak current is 0.005 µA.
  • Lock the range, allow settling beyond the published response, and record temperature, humidity and a zero-current baseline with every leakage result.

Request a configuration review

Tell us the channel count, current range and protocol. We will check the measurement class against the note and return a configuration, not a generic datasheet.

Request a configuration review

References

← All technical notes

Talk to an engineer

Request a quote