DDS238-2 wattmeter, energy meter, WiFi-controlled relay rated for up to 65 A

I like all kinds of interesting electrical gadgets, so I ordered a couple of different ones to try them out, compare them with others, and tell you about them. Today I’m reviewing the first one: a WiFi-controlled relay.

A reminder: to keep up with new posts on the blog, I recommend following me on Instagram. I’ll post notifications about every new article there, and might also write about something interesting from time to time — link to my account

I called this device an electricity meter in the title, but let me clarify right away: you can’t use it as an officially registered meter. It’s a meter in terms of its functionality, and the readings are only for your own use.
Also, after my review of the EM-129, people sometimes asked me why it was so expensive when there are plenty of cheap Chinese alternatives that measure things and have a relay too. So today we’ll compare them and see what the differences are.

As far as I can tell, there are at least four versions of this relay:
1. No remote control
2. RS485 control
3. Zigbee
4. WiFi

This review covers the fourth version. I also found that there are other differences: for example, the delays when using Zigbee are significantly longer than with the WiFi relay.

The relay comes in a small box, and it looks like it arrived without damage.


The package is as simple as it gets: the relay and its manual.


The manual is in English, but I was really only interested in one thing: the technical specifications. I might also have wanted a list of the readings it displays, but I figured that out myself.


The relay has the usual form factor for DIN-rail mounting and is two standard modules wide.


1. There’s a small display window on the front. In my opinion, it’s not very easy to read. Below it is a button with the following functions:
switching display modes
turning the output on/off
rebooting
putting the device into WiFi pairing mode.

Below that are three LEDs: output status, WiFi connection, and pulse output activity.
2. The screws are covered by caps, which, as expected, have slots for seals. You can still get limited access to the terminals, though to be fair, the EM-129 doesn’t even have caps.
3, 4. In addition to the power terminals, there’s a pulse output terminal block that outputs 1600 pulses per kWh.
5. There’s a wiring diagram on the side. I have to say, the wiring arrangement is very inconvenient. Usually, devices like this (not just meters, but devices in general) have the input at the top and the output at the bottom. Here, the input is on the left and the output on the right, with neutral at the top and live at the bottom. And terminals 1 and 2 are... at the bottom, while 3 and 4 are at the top. What a mess.
6. The DIN-rail clip holds securely.


All right, let’s move on to the insides. To get in, you need to remove a couple of self-tapping screws and separate the two halves of the case. By the way, there’s no protection against unauthorized access here, apart from a paper sticker.
The upper board is removable. I like the design; it seems quite well thought out to me, and a little better than Novatek’s.


Inside is a latching relay rated for 80 A, even though the device itself is specified for 65 A. The coil voltage is 9 V. I often see similar relays in domestic and imported overvoltage protection relays, though with a slightly different power-contact configuration. That’s probably customized to meet a particular specification.


This part is really interesting: one of the relay’s power terminals also acts as a current sensor—in effect, a current-measuring shunt. The metal is even stripped bare in this section.
And here I’d take my hat off to the Chinese designers (if I wore one). The EM-129 uses a somewhat similar approach, but its section is noticeably longer (so the voltage drop is probably higher), and it can only measure down to about 50 W. This manufacturer specifies a minimum of 20 mA, or around 4 W, but I actually measured power as low as 3 W.
I think the developers at Novatek have something to learn from this.


To remove the relay from the case, you need to take out another couple of self-tapping screws. I have no complaints about the terminal design, even though I tightened them with quite a bit of force during testing.
I also have nothing bad to say about the relay’s contact group. Subjectively, it’s a very ordinary relay. The difference from the ones I’ve seen before is that this one has a single coil, and the operating mode is switched by applying voltage with different polarities. So the mechanism is different, but the contact group looks the same.


Top view of the power board. The voltage regulators are here. The module isn’t powered through a capacitor; instead, it uses a DC-DC converter. The EM-129 also uses a DC-DC converter.


But the EM-129 also has a supercapacitor to power its built-in clock. This one has neither a supercapacitor nor, for that matter, a clock.
The main PWM controller is a PN8016. Nothing unusual here: a typical PWM controller with a built-in 800 V transistor.


1. The underside of the board, showing a pair of 4.7 µF, 450 V input capacitors and one output capacitor.
2. For protection, they used a resistor as a fuse and a 680 V varistor.
3, 4. Control board

A few words about the protection: the input relay has 450 V capacitors, but they are connected in parallel, so the maximum input voltage is about 320 V. I applied up to 450 V to the EM-129.
At the same time, this unit has a 680 V varistor, meaning it will actually start working at an input voltage of 480 V and is really only useful for protecting against interference and very short voltage spikes.
The EM-129's varistor is selected so that it starts working before the PWM controller can die from overvoltage.


To access the controller, whose markings had been rubbed off, I had to desolder the WiFi module. By the way, a Zigbee module may be installed in its place. And as far as I can tell, the current sensor is connected directly to the microcontroller, without an op-amp.


Let's plug the device into the mains and see what it can do.
WiFi isn't configured initially, and the LED flashes, but what surprised me was the relay activity indicator.
There's an exclamation mark next to the output activity indicator, and the LED itself is red. But for crying out loud, what idiot came up with an indicator that shows not whether voltage is present after the relay, but whether it's absent?


In other words, when there's voltage at the output, the LED is off; when there's no voltage, it lights up red... brilliant.
The app used to control the relay will appear throughout the review. I'll tell you more about it toward the end.


The display can show a lot of different parameters. Here they are in the order they appear after power-up.
First, it displays the number of pulses per 1 kWh. Here it shows "C 1600", but I never figured out what this means, since the same information is printed on the relay's case. After displaying this parameter, the relay switches to normal operation, and pressing the button cycles through 9 display modes:
1. Total energy counter, range 000000.0–999999.9 kWh. Rumor has it you can reset it by holding the button down, but the most I managed to do was unlink the device from WiFi.
2. Energy consumed counter, range 000000–999999 kWh
3. Energy returned counter, range -000000 to -999999 kWh
4. Load current, display resolution 0.01 A
5. Input voltage, display resolution 0.1 V
6. Active power in the load circuit, display resolution 0.1 W
7. Reactive power in the load circuit, display resolution 0.1 var
8. Power factor, display resolution 0.01
9. Mains frequency, display resolution 0.01 Hz


A few tests.
First, I checked how the device behaves at low input voltage.
The specified range starts at 100 V, but in reality the device kept working down to about 50 V at the input. It tries to operate at 21–25 V, but there's a small "catch".
As you can see on the smartphone screen, when I gradually reduced the supply voltage to zero, the device shut down at an input voltage of about 50 V. After a while, the app stopped detecting it and reported a lost connection.
At around 35–36 V, the device managed to start up normally, display its status, and even measure the voltage correctly, but it wouldn't reconnect. Raising the voltage above 100 V didn't help either. I waited about a minute, then switched the power off and back on (115 V), and after startup the device connected to WiFi within a few seconds.

A gradual increase in voltage is a rare situation, but it's theoretically possible nonetheless.


Next, I tested it across the 50–260 V range. The voltage measurement accuracy was good; even in the worst case, the difference was about 0.3 V.
I didn't apply more than 260 V because I saw no point. The device should easily withstand 320 V, and even more for a short time, but that would be outside normal operating conditions. The manufacturer specifies a range of 100–270 V (elsewhere, it says 100–300 V).


Assessing current measurement accuracy.
I used a couple of types of incandescent bulbs (15 and 125 W) and a couple of irons (400 and 2000 W) as loads.
The readings on the smartphone lag by about a couple of seconds, so there are some small differences. The device's own display also has a slight delay, while the multimeter is noticeably faster, which adds a little measurement error. The error increased because as the “load” warmed up, the current drifted down slightly, so it had time to “run away” further on the multimeter display than on the power meter.
So, on the left is the current shown in the app on the smartphone screen, in the center is the current on the device display, where the delay is shorter, and on the right is a value closer to the actual one.

Well, what can I say? In my opinion, it’s pretty good—and at low currents, even excellent. And that’s with a very low-resistance shunt. At 2–10 A, the little meter read slightly high, but by just 20–30 mA, which at that current is 0.5–1%.


While I was at it, I measured the voltage drop across the device.
The measurements were taken directly at the terminals, but given the relay’s design, you could say I was essentially measuring across it.
At 1.83 A, the drop was 2.345 mV, giving a calculated resistance of 1.29 mΩ.
At 9.14 A, the drop was 12.694 mV, giving a calculated resistance of 1.39 mΩ.

We can take the higher resistance as the actual value—say, 1.4 mΩ—and use it to estimate the power dissipated at different currents.
10 A = 0.14 W
20 A = 0.56 W
30 A = 1.26 W
40 A = 2.24 W
50 A = 3.5 W
60 A = 5.04 W

In my opinion, it’s best to limit the continuous current to 40 A, though it’s perfectly reasonable to raise it to the claimed 60–65 A, as long as you limit the duration.


For the next test, I had to use a clamp meter because the multimeter can’t measure such a high current. Naturally, the measurement error is higher this way, but setting up an alternative would have been more complicated.


You can see that the readings differ more when measured with the clamp meter. For example, at 6.38 A, the power meter read 6.26 A, whereas the difference was clearly smaller when the multimeter was connected directly.
The difference is about the same at 20, 40, and 64 A, where the power meter showed:
Clamp meter - power meter
20.67 A - 20.24 A
40.18 A - 39.05 A
64.5 A - 62.99 A

Also, keep in mind that the app displays readings with a slight delay, and the current isn’t stabilized. In this setup, the difference reached 2.5%.


Next, I decided not to stop at 64 A and raised the current a little—first to 72 A, then, about five minutes later, to over 80 A.


As it turned out, even after five minutes at 72 A, the temperature rise was relatively modest. The wires got hottest.
1. Top view
2. View of the right side of the relay housing.
3. Overall view of the “test setup.”


As claimed, the relay has two energy counters: one for energy received and one for energy delivered.
1–3. Initially, it was connected in the normal configuration, and the total counter and the received-energy counter showed the same readings, while the delivered-energy counter showed zero. A reminder: the total counter has a minimum increment of 0.1 kWh; the others, 1 kWh.
4–6. After reversing the phase, the total reading continued to increase, the received-energy counter stayed the same, and the delivered-energy counter started counting.
7–9. I ran it for a while longer. Everything works correctly: the total counter tracks the total, and the other two each track their own energy.


This is what the display looks like under conditions close to real-world use: current, power, and power factor.


I noticed this particular quirk completely by chance while sorting through photos for the review.
These measurements were taken with the phase reversed. Only afterward did I notice that the current is displayed much more accurately here. Two consecutive photos show the power meter reading fluctuating by just ±0.2 A.


About the app.
You’ve already seen a bit of it during the tests above; here’s a closer look.

There are two apps available, Tuya Smart and Smart Life, but I got the impression they’re essentially the same app. At least, the interface and the power meter’s parameter screen are identical.

Install either app and connect the power meter. It seemed complicated at first, but it turned out to be very straightforward. Power up the meter, then press and hold the button until the LED starts flashing rapidly.
In the app, choose to add a new device,
then select a Wi-Fi outlet in the window that appears,
and the name of your access point will appear (your smartphone must already be connected to it; 2.4 GHz only).
Enter the access point password, tap “Next” a few times,
then select “Flashing rapidly.” In about ten seconds, the power meter will connect to Wi-Fi and appear in the app.


The relay interface is quite simple. There’s a main screen with an on/off button, plus several additional buttons at the bottom:
1. Duplicates the large button, and if you're in another menu, returns to the main screen. Response is fast, about 0.5 sec.
2. Timer: set the time to turn the relay on or off.
3. Countdown: set how long until the relay turns the load off/on or switches the output to the opposite state.
4. Menu for displaying counters and measurements.


The interface for displaying meter data is rather poorly designed. The large digits at the top show the daily kWh counter, which resets every day. The smaller digits below show:
1. Current in the load circuit, in mA
2. Power consumed by the load, in 0.1 W increments
3. Input voltage, displayed in 0.1 V increments
4. Total consumption counter, displayed in 0.01 kWh increments

And here are a few immediate observations:
1. The values are displayed in whole units everywhere—that is, volts, watts, kWh—while current is measured in mA. That's confusing; it would be more consistent to display current in 0.001 A increments as well.
2. There's tons of space on the screen, so why are the digits so tiny? Do I need a microscope to read them?
3. The wattmeter reports a bunch of other measurements, as well as two additional counters. Where are they?
Overall, there's still a lot of work to do here.

At least I got a screenshot while passing almost 90 A through the relay. The pulse-output LED was blinking like crazy :)

You can display a consumption graph, which is handy.

From the description on the store page -

IMHO, this is kind of ridiculous. Are there really meters that won't count without WiFi? :)


There are very few settings for the device. By the way, as far as I found out, the Zigbee version has a separate "Settings" tab, but for some reason this one doesn't. I wanted to update the firmware, but it says the installed version is already the latest.


That's basically all there is to the device itself; everything else is a software feature of the app.
For example, you can create a scenario that links an event to a change in the device's status, such as:
1. Switching the output on or off
2. When the timer expires
3. When the current rises above or falls below a set threshold
4. The same for voltage
5. The same for power
6. The same for the kWh meter, though this is probably only useful for exceeding a limit—for example, setting a limit on how much electricity something or someone can consume.

One review of a similar Zigbee device said its response time was about 30 seconds. The one reviewed here is noticeably faster, at around 5–6 seconds. Even so, the delay is too long for it to serve as proper overvoltage protection, since everything is handled by the app rather than the wattmeter.

As a power limiter, though, it works quite well. That kind of delay actually helps prevent false trips. That's the mode I was experimenting with, which is how I found out the response time.
I suspect the delay is artificial, because even allowing about 2 seconds for data transmission and roughly 0.5 seconds for the control command, the total should still be under 5 seconds.


The app also lets you trigger an action when your phone enters a certain area—for example, turn something on when you (or rather, your phone) get home and turn it off when you leave. It also supports voice control and plenty of other features, but again, those are app features and don't have much to do with the wattmeter itself.



Conclusions.
Overall, I liked the relay, although it does have a few shortcomings. Most of them are related to the app rather than the relay itself. The issues specific to the relay are fewer and include:
1. The output voltage indicator works in reverse
2. Little headroom on the input voltage: rated for up to 270/300 V, but in reality it can withstand up to 320 V.
3. The display is tiny, and even the lens-like window doesn't help.
4. WiFi may freeze if the input voltage is raised gradually from zero.

Otherwise, it's quite good. I especially liked the fairly accurate measurements and its ability to work properly with very low-power loads. The relay also switches the rated 65 A without issue and can handle 80–90 A briefly, but I wouldn't recommend running it continuously above 40 A.

But compared with the much more expensive domestic EM-129, things aren't quite so rosy. The wattmeter reviewed here can basically "boast" only about its ability to measure low-power loads and its fairly well-thought-out design. In other respects:
1. No thermal protection
2. No built-in logger, which is why there's no battery or supercapacitor either
3. No software or hardware protection against overvoltage or undervoltage
4. No built-in protection at all
5. It works through a Chinese cloud service, as far as I can tell, so delays may be longer. The EM-129 can work without internet access, within a local network.
6. Essentially no warranty
7. Not even a brightness adjustment for the indicators :(

Overall, next to the EM-129 it looks pretty unimpressive. What partly redeems it is its reasonably good circuit design and build quality, support for Tuya and so on, and of course the price—the 3–4× difference is significant.

The store provided coupon BGc9950b, which brings the price down to $28.99 through February 28, although a dollar off is pretty stingy :(
I've seen this relay on AliExpress for $25.

That's about it. I could have gone into more detail about integrating it with other systems, of course, but I was much more interested in looking at the hardware.