Ucurrent

ucurrent clone

This repository holds a clone of the ucurrent gold from https://www.eevblog.com/projects/ucurrent/

Design changes

• Reimplementation of design in kicad
• exchanged switches for jumpers to reduce cost
• changed shunt usage which reduces accuracy of some measurement regions for simplicity
• used cheaper resistors for a reduced total accuracy which should be good enough for most use cases

Description of range selection

Range selection happens by putting a jumper on the selector panel. To change the range without powering of the device under test, just use a second jumper and remove the first one later. The proposed sense resistors are:

• 10 milli ohms, selected in the most bottom row. (1 mV / mA)
• 90 + 10 milli ohms, selected by the second row from bottom (10 mV / mA, allows more precise measurements in the typical 0-125/250 mA region)
• 10,1 ohms (1.01 mV / µA; so 1 mV/µA with +1% error), selected by the third row from bottom. The measurement error can be mostly removed by setting a second jumper vertically on the bottom right to bridge R5.
• 10010,1 ohms (1.00101 mV / nA; so 1 mV/nA with + 0.1% error), selected by a vertical jumper on the top left)

BOM changes

• Most precision resistors are available as 0.1% / 25ppm/K versions from the german distributor Reichelt.
• R5 has a special value of 90 mOhms and can be bridged if none can be obtained (or use a lower value).
• All shunt resistors are used as 1% versions. You are free to use more accurate ones to improve accuracy. I did rather safe 10 euros...
• The battery indication around the TPS3839 (compatible to TPS3809) is fully optional and activated by a push button. Device needs to be powered on!
• Power can be fed via an external source, power selection (external/battery/off) happens via jumper.

BOM

Specs/Errata

Current usage of other ucurrent clone at 2.7V: 1.004 mA (via Multimeter) or 1.04 mA (via ucurrent clone @ 1mV/µA). Unfortunately, there is a bunch of problems, which also seem to exist in the original ucurrent, but have different levels of manifestation.

• Oscillations of reference voltage: Original ucurrent gold seems to be sensitive about the exact type of U2. I tried two versions and have not been successful. Placing 100nf each over R4 and from GNDREF to V- makes GNDREF stable. This can easily be done by soldering 100nf directly above R4 and another 100nf 1206 from R13 to R4.
• Oscillations/feedback when using nA range: When operated in the nA range the output oscillates with 100kHz-1MHz. It looks like feedback is picked up in the first x10 stage and propagated to the output. Reducing bandwidth by placing 5nF from U3/pin3 (IN+) to GNDREF stops the feedback but reduces bandwidth to ~10 kHz.
• Reduced bandwidth: Output voltage drops significantly above 100 Khz.

Some scope shots, CH1 voltage over a 10 kOhm resistor, CH2 output voltage of ucurrent clone

Case

A draft for a 3d printable case is there as well. There is some space for a 2x AAA batteryholder below the PCB but it does not contain any mounting utilities for that.