Blog #2: A Current-Sensing PSU Upgrade - Unboxing

23 Apr 2022

Upon receiving the gear, obviously the next thing that needs to be done is to unbox it and prepare it for use. It’s probably no big surprise as to what we received – as we were told all about this in the official challenge, but I will be talking a bit about the design of the gear.

Unboxing

Because of supply issues, contestants received just the Mikroelektronika Pi 3 Click Shield and Current 6 Click. This combination should allow the evaluation of the MAX40080 with the Raspberry Pi platform with minimal work, while the Click board itself could be reused with nearly any microcontroller supporting I²C on its own.

The use of Click boards on the mikroBUS is a time-saving choice for prototyping as the pin-out has a relatively manageable number of pins which host 5V and 3.3V, UART, I²C, SPI, GPIOs and Analog. It’s also breadboard friendly thanks to the use of 2.54mm standard headers.

Interestingly, the shield is Made in Serbia – this was not my expectation.

The packaging is very much cardboard-based which is environmentally friendly. They have turned the unboxing into its own “educational” experience, advertising some of the key benefits of the system.

The box folds out to reveal the board, in the open, and beneath it an anti-static shielding bag containing screws and standoffs. These parts didn’t really need the anti-static treatment, but it was probably what was on hand.

In the meantime, the Click Shield ideally should have been packaged in anti-static protection, given that it contains an MCP3204 12-bit SPI ADC and TXB106 Bidirectional Level Shifting Voltage Translator. The board itself allows hosting two Clicks and does not have a pass-through header unfortunately. It contains a pair of switches that allow changing the routing of the Analog lines from the two supported Click modules. The silkscreen on top makes it clear as to the functions hosted on each pin and the connections on the Pi GPIO header. There is even a test hook for ground, in case you’re probing about the board, which is a nice touch. The only slightly inconsistent thing is the soldering of the 40-way GPIO header which has a number of small solder spatter balls and flux remnants. This appears to be hand-soldered, so that is not entirely surprising.

The underside does not host any components, but does have the pinout silkscreen shown for convenience. The quality of the board manufacture is sufficient but imperfect, given the weak silkscreen on the left side. It appears to be a two-layer HASL finish board, marked as Revision 1.00 with a QC label from 2021.

Unpacking the Current 6 Click required going through some cardboard … a sleeve in a sleeve. At least it kept it secure, but without any ESD protection. I’ve generally had good luck with cardboard, but it’s never a guarantee. By the way - what's wrong with wires and soldering? It's not necessarily a waste ...

The Current 6 Click looks like this, with components all mounted on the top. This includes the MAX40080 Precision, Fast Sample-Rate Digital Current Sense Amplifier, LTC4301L 2-Wire Bus Buffer with Voltage Level Translation, ADP151ACBZ Ultra-Low Noise 1.8V CMOS Linear Regulator and a DMN2058U N-channel MOSFET. There is also an LED and jumper that allows for Vio selection. The silkscreen marks which pins are used for which functions. This board is Revision 1.01.

Looking closer at the board, can you believe that the part we are evaluating is U1 – not much bigger than the resistor R4! This is great news in terms of footprint-constrained implementations – small chips mean less board area consumed. The downside is that it is a potentially finnicky BGA component, so hand-mounting could be a little more challenging than average. A 0.01Ω shunt resistor (RS) is pre-installed. With the first range that expects an input of ±50mV, this equates to a current of ±5A. With the smaller range selection of ±10mV, this corresponds to a current of ±1A. Because of the MAX40080’s shunt input range of ±2V maximum, this means that the current through the board should not exceed ±200A or that will cause damage to the IC. However, at that current level, the traces and shunt resistor may well be vaporised …

The design of the click has the connections on a “neck” which rises out of the body of the Click turning it into a bit of a “lever” which has the potential to unseat the Click if the pins aren’t held tightly enough. There are four connections on the orange terminal blocks – unfortunately the blocks themselves seem a bit small for the sort of wire I’d want to use. However, this configuration has the benefit of allowing you to either configure high-side or low-side current sensing. In many applications, the ground must remain unbroken to ensure stable operation, thus high-side sensing (having the shunt resistance on the positive side) can be achieved by shorting Vin (Vs) to RS+. The MAX40080 is capable of voltage measurements and handling a high common-mode voltage up to 36V, making it suitable for such high-side measurements at voltages up to 36V. However, low-side sensing is perhaps easier to implement in other cases where positive-ground is used, which can be configured by shorting RS- to GND. In this case, if voltages above 36V are involved, I speculate (but cannot confirm yet) that the chip may still be used perhaps without connecting Vin (Vs) and sacrificing voltage measurement capability.

Apparently, the IC is pretty power-efficient too – if power is a concern, removing R1 to disconnect the LED will help. A more drastic approach would include running 1.8V directly including I/O, bypassing the need to use any regulators or level shifters, but that’s probably best left to a board redesign.

It all Clicks!

From here, it’s a simple matter of assembling the boards together, which literally “click” together. The exception is the standoffs which require a screwdriver to assemble – the holes in the board were a bit tight for the screws making the process a little more involved than desired.

In my case, I had a Raspberry Pi 3 board spare, along with a Mulicomp Hat case that allowed for the board to have the lid opened for the board to be exposed.

The stack has some height – just make sure that the wires don’t lever the Click out of the board, or else some unfortunate connections may occur – Vin to the Raspberry Pi GPIO may be quite a sizzle!

Conclusion

The Mikroelektronika Click system is designed for rapid prototyping and in this case, will allow us to evaluate the MAX40080 without much hardware design effort thanks to it being all pre-built into a set of blocks which “click” together. While there were no major problems, the manufacturing of the Click Shield could be a little better with better quality silkscreening and soldering of headers, plus screw-holes that more easily accommodate the screws. The Current 6 Click similarly seemed a bit constrained by the size of the terminal blocks and their positioning seems to invite levering the board out of the socket by stiffer wires. The choice not to use anti-static packaging, instead opting for cardboard, did seem a bit strange to me. Regardless, for prototyping and conformance with the “Click” standards, it seems to have done a good job.

Now that I’ve unpacked the equipment, given you a tour of it and my thoughts, it will be time to set up the OS, software and get some initial results. I suspect this process is probably going to take some work …

---

This is part of the Experimenting with Current Sense Amplifiers Design Challenge – A Current-Sensing PSU Upgrade Blog Series: