Ammeter-voltmeter with a VFD display and power measurement

Continuing the topic of devices with VFD displays, I want to tell you about a rather interesting product: a fairly accurate ammeter-voltmeter that could come in handy for a hobbyist.


I first came across this ammeter-voltmeter on Taobao, where it costs about $13.50. The price is quite reasonable, but you also have to pay for the services of a purchasing agent. The one thing in its favor was that shipping within China was free, though that's rarely the case.
But when I started looking for it on AliExpress, I found it for not much more—around $19–20. I was about to buy one to try it out when I noticed the seller offered a wholesale discount starting at two units, bringing the price to about $17 each, or $34 for two. The difference from buying on Taobao was minimal, so I decided to buy it on AliExpress.
The most curious thing was that a few days later, the remaining stock for this Taobao listing had dropped by exactly two units :)

Unfortunately, the seller has since raised the price, and two boards now cost $50 instead of $34. Maybe it's because this is a different board revision, or maybe for some other reason—I don't know. Either way, let's see what this thing is.


Shipping fragile goods through a whole chain of postal services is always stressful for both the buyer and the item, so I was worried too. As it turned out, there was no need to worry; the packaging was fine.


Package contents
1. Ammeter-voltmeter
2. A pair of terminal blocks
3. A few buttons
4. USB-to-mini-USB cable.


The cable is straightforward: a standard USB-to-mini-USB cable. The rest is a little more puzzling. For some reason, there are two terminal blocks per board. One is clearly for connecting the ammeter, but the purpose of the other is a mystery, since it only fits at the power connector.
They included 9 buttons for two boards, even though each board only needs 3, so that leaves 3 spares.


The board is fairly large, partly because of the display. Its dimensions are 100 × 60 × 17 mm.

Specifications:
Voltage measurement range - 30 V
Current measurement range - 5 A
Supply voltage - 5 V
Current consumption - up to 300 mA



Dimensions from the product page:


It even includes the dimensions for the mounting holes and buttons—a very useful drawing.


A closer look at the board. As you can see, the display takes up quite a lot of space.


There are two SMD LEDs above the display. They make it easier to tell which information is shown where; I'll demonstrate in a bit.


The display is a nine-digit, seven-segment type, exactly like the one used in the volume controller I reviewed some time ago. There must be a lot of these in China, so they send install them wherever they can.


At first I thought the display was used, but on closer inspection it turned out to be just dirty, probably with flux residue. That said, it could still actually be used.


Below the display are three control buttons, arranged in the “Chinese” style: plus on the left and minus on the right.
The plus/minus buttons adjust the display brightness from zero (display off) to maximum. The Function button switches between display modes. All settings are saved, so the next time you turn it on, it will use the same brightness and mode as before it was switched off. I think that's the right way to do it.


All the components are on the back of the board. The overall build quality gets a C, and that's being generous.


The display's power converters: the one at the bottom is for the filaments, and the step-up converter on the right supplies the anode voltage.
You can also see the driver that controls what appears on the display.


The microcontroller and USB-UART converter are located at the other end. Yes, there’s also a buzzer that emits a short but loud “beep” when power is applied. As I understand it, it’s also used for alarms.


This is what I mentioned right at the beginning, and why the seller may have raised the price. I have board version 1.3, while the product page shows version 1.2. The key difference is that this board has the USB-UART converter and miniUSB connector built in, which are used to connect it to a computer for data transfer and power.


The ADC section. The advantage of this ammeter-voltmeter is that it uses separate ADCs rather than the ones built into the microcontroller. The seller specifies an 18-bit ΔΣ ADC: 3.75 SPS (18 bits), 15 SPS (16 bits), 60 SPS (14 bits), and claims a 5-digit resolution.
But, friends, what beautiful soldering... I’m speechless—at least, I can’t think of anything polite to say. And it’s like this on both boards. The problem isn’t just the mess; flux is hygroscopic and can affect circuit parameters, especially as the humidity changes.
Ideally, it should be cleaned and coated with varnish, but after that, recalibration will almost certainly be necessary.


There are some unexplained contact pads on the board, JP1–JP4: a pair near the ADC and another pair near the shunt. One jumper is shorted; another is meant to connect the shunt’s positive terminal to the voltmeter’s negative input. I think it would be difficult to figure out what they’re for without a schematic.
The shunt resistance is 17.5 mΩ.


There’s an OC output near the USB connector, marked ОС, presumably meaning Over Current. Since the board lets you set protection thresholds, it may well be a protection output. I haven’t checked yet, for reasons I’ll describe below.


Although the boards are identical, they do have a few minor differences, such as different types of capacitors and visibly different shunts, despite having the same resistance.


The overall build quality is disappointing in places: there are piles of solder, and judging by the appearance, it was done with a cold soldering iron. By the way, there’s an additional transistor array near the power connector, presumably for reverse-polarity protection on the power input.
The connectors on one of the boards are dirty, as if they’d been used before.


Connection. It’s exactly the same as with ordinary ammeter-voltmeters, and that’s where the problem lies. Small, inexpensive meters with low resolution won’t show the effects of this connection, but this one has noticeably higher voltmeter resolution, so the voltage drop across the shunt becomes an issue. In other words, the voltmeter measures the voltage before the shunt, not after it, so the measured voltage falls as the load increases. And because the voltmeter has higher resolution here, the resulting error becomes more apparent.
As I mentioned above, the shunt resistance is 17.5 mΩ and the rated current is 5 A. That means the voltage drop across the shunt alone will be 87 mV at that current. Add the drop across everything else and it comes to almost 0.1 V, which is a lot. Ideally, this drop could be compensated for “on the fly,” since it can be calculated mathematically, but nobody has done that.
Also, as always, the voltmeter’s negative connection point can complicate things. In some tests, I didn’t connect it at all.


The connection diagram on the product page cheered me up a little :)


The brightness control can turn the display off completely, including the filament, as you can see from the drop in current consumption (left). Overall, power consumption isn’t very high: up to 1 W at maximum brightness.


I mentioned above that there are two LEDs above the display. They light up depending on the display mode: the red one is above the current indicator, and the blue one is above the voltage indicator. There are 6 display modes in total:
1. Ammeter + voltmeter, displaying 4 digits
2. Ammeter + voltmeter, with the voltmeter resolution increased to 5 digits and the ammeter resolution reduced to 3 digits
3. Same as mode 2, but “mirrored”
4. Voltmeter with 7-digit resolution
5. Ammeter with 7-digit resolution.
6. Wattmeter.


There’s a similar description on the product page, but it’s somewhat inaccurate. The mode sequence is different, as is the backlight color, and in Wattmeter mode both LEDs actually light up.
The basic idea behind this kind of indicator is simple and convenient, and I like it. The only problem is that the LEDs are rather dim, and with the light filter in place it’ll be even worse. I’ll have to think of something.


The first test was carried out using a reference voltage source that I reviewed some time ago.
The actual values measured for this review are shown in the Fluke 8508A column.


I tested it in the maximum-resolution mode and found that the least significant digit is always 0, while the two digits before it change in large steps.
The photo shows the readings changing; there are no intermediate values.


That’s why I took 8 photos instead of 4, to show what the measurement looks like. Since the measurement range is floating, once the voltage reached 10 V, the last insignificant zero disappeared and the last two digits changed. But again, the step size was about 280 µV, or 0.28 mV, or 0.00028 V, whichever you prefer.

In principle, this makes sense: if the ADC’s full range covers 0 to 30 V, then with 18-bit resolution the step size is 1/262144, or 114 µV. In other words, the developer simply displayed everything coming from the ADC without worrying too much about which digits were meaningful. In practice, you can count on a maximum of 5 digits, which is what the specifications stated.


Now for some pictures to show what these boards are actually like. The product page says they’re calibrated using precision measuring instruments. My instrument isn’t in that class, of course, so the results won’t be quite as clear-cut.
Let’s start with a voltage of about 1 mV; I doubt anyone would measure anything lower.
I’d say the results are pretty good. By the way, the photo at 5 V shows a resolution of 280 µV, meaning the reading changes in steps of about that size.


I tested current measurement starting at about 0.5 mA. The reading starts at 0.22 mA and then mostly increases in 0.11 mA steps: 0.22–0.33–0.44 mA, and so on.
Most of the photos are shown in pairs to demonstrate the measurement’s minimum-step fluctuation while the current remains unchanged.


At this point I thought: why am I testing just one board? Testing two would be much more informative, and I could also check how well they’re calibrated against each other. It’s not much, but it’s still a bit of statistics.
I took two boards, connected their voltage-measurement inputs in parallel, and powered them from a power bank so I wouldn’t be tied to the mains.
That’s when an interesting quirk showed up, as seen in the first and second photos. The board powered up last displayed nonsense instead of zeros.
I had to use two power banks to isolate the boards’ power supplies completely. After that, the readings on the screens still weren’t always zero, but they matched: for example, both showed zero, then one board showed 280 in the three least significant digits, then the same thing happened on the other board or on both. In effect, this was one digit fluctuating in the last place, but here one digit corresponds to 280 µV.


Here I can say that both the 5-digit accuracy and the calibration between the boards are very good—at least, not much worse than my multimeter.


Unfortunately, current measurement isn’t as smooth: one board reads noticeably high.


But then came a real surprise. When I placed the order, I wondered why the manufacturer had limited the specs so much—30 V and 5 A—when I’d have liked a little more.
When it came time to test, I found that:
1. The maximum measurable voltage is 37.5 V. In the second photo, the voltmeter hits its limit and the reading simply stays there. The limit is different for each board.
2. I increased the current to 8.5 A and it was still measuring. I stopped because I didn’t want to push the shunt any further: it was getting noticeably hot, dissipating about 1.2 W, and I’d happened to pick a board with a thin shunt. In principle, though, there’s nothing stopping you from using a higher-power shunt. By the way, the reading drift as it warmed up was very small, which is good to see.



At the very beginning I mentioned that the board supports a connection to a computer, and the seller provided the software without any trouble. But it turned out not to be quite so simple.
The software is designed to display voltage, current, and power, which is convenient enough.


There’s also a dedicated button to open calibration mode, where you enter the actual voltage or current value. As far as I understand, the board is supposed to use that as the correct value. There is also a button to restore the factory settings.


As I mentioned above, the board can signal when thresholds are exceeded. The current and voltage thresholds are set in the software, and you can set both upper and lower limits.


And everything would be almost perfect—apart, of course, from the software being in Chinese—if only it worked.
To start with, my antivirus flags the program, but I’m used to that with Chinese software. It also only works when run as administrator. The most I’ve managed to get out of it is for the received-packet counter to tick up and the display precision to change: when launched normally, it shows 0.0000 and receives nothing; when run as administrator, the display changes to 00.000 and data starts coming in, but nothing else happens...
I wrote to the seller, but haven’t heard back yet.

Further testing showed that the software can set the alarm threshold without any problems, and that works. But the current voltage and current still don’t show up on the software’s screen.



There’s a lot of text above and plenty of pictures, but it’s probably easier to read the conclusions to get a quick idea of what this ammeter-voltmeter is like.

What I liked: the VFD, which was partly why I bought it; separate ADCs; decent accuracy (the voltmeter is comparable to my “unit”); several display modes and extra indicators; adjustable brightness; no flicker when photographed (sometimes that matters); the ability to connect it to a computer; and configurable alarm thresholds. Its actual measurement range is significantly wider than specified. The readings update about 4 times per second.

What I didn’t like: current measurement accuracy is still a bit off; the software is in Chinese, and I have a feeling I’ll have to struggle to get it working properly; the build quality isn’t great; and the voltage drop across the shunt affects the voltmeter reading.

Bottom line: now I’m sitting here wondering where to stick put it. Its measurement accuracy and resolution seem like overkill for a power supply or electronic load, while using it as a standalone measuring instrument feels awkward, and it isn’t very functional either. I bought it simply because I wanted one; I had an idea to build a two-channel linear power supply. There’s a noticeable error in current measurements, at least on one unit. The board also struggles to measure small currents—but come on, at 1 mA the voltage drop across the shunt is just 17 µV! And it measures it!
It’s unclear why they put 7 digits on the display when one doesn’t work at all and two have a very coarse step size. The meter only delivers its advertised 5-digit resolution at 10 V and above; below that, it only gives 4 digits.

So, it’s a mixed bag. Personally, though, I liked the boards more than I didn’t, and I think I’ll find some use for them, if only for their specs and appearance :)