Modifying the RTC power circuit in the UT181A multimeter

The Unit UT181A multimeter has a useful feature of its own: a built-in real-time clock. But it turns out there’s a glitch—or, to be more precise, a glitch with my particular meter. Someone else may run into the same problem, so today I’ll tell you how I dealt with it.


I wasn’t sure what to call what I had to do. It didn’t really count as a modification, since the circuit remained unchanged, and it wasn’t exactly a repair either, because the multimeter basically still worked.

A few years ago, I bought myself a new multimeter and even wrote a review with tests. In the review, I mentioned that it arrived with a dead battery. I remember writing to the seller, and in the end they even refunded me a small amount. I replaced the battery and thought nothing more of it.


But I was wrong. It turned out that after a relatively short time—maybe six months, maybe less or more—the battery had gone flat again. As a result, you may have occasionally seen some wildly incorrect dates in my reviews, and they weren’t always the same.


At some point I got thoroughly fed up with it and decided to find out what was wrong. The RTC circuit is built around an SD2068 chip.


I looked online and found its datasheet, which says that the chip was designed to be powered through two inputs: pin 3 from the backup battery and pin 8 from the main power supply.
After checking the chip’s pins, I found that pin 3 is actually connected to ground, while pin 8 is connected through two diodes to both the battery and the main power supply, which is present when the multimeter is switched on.


I couldn’t find the full datasheet, but I did discover that this chip has an identical twin called the BQ32000. Its datasheet says it draws a whopping 100 µA from the main power supply.
100 µA really is a lot. For example, a typical 2032 cell has a capacity of 210–240 mAh, so at that current it would run flat in 100 days.


The solution seemed obvious: disconnect the battery’s positive terminal, lift pin 3, and connect the backup power to it. I did that, but it didn’t work out: I couldn’t even set the time. I’d go into the settings, set it, press OK, and the time wouldn’t change. I had to put everything back the way it was.


Next, I had to “phone a friend.” I asked two owners of this meter about the clock and asked them both the same question: how well does your clock keep time?
Both replied very quickly and said theirs worked perfectly.

So it seemed the problem was specific to my unit. During our conversation, one of them measured the current draw on his meter and shared links to the BQ3200 and SD2068 datasheets. The other provided information about the SD2068 that I hadn’t been able to find.

It turned out that the complete SD2068 datasheet shows a much more interesting circuit. It includes a configuration like the one in the multimeter, with power supplied only through the main input and isolated by diodes. That circuit wasn’t in the datasheet I found.


It also turned out that the current draw from the main input should be in the range of 1.2–1.5 µA when only the clock is running, and 40–120 µA during data transfer. Again, the datasheet I found didn’t include this table.



At first, I thought I’d got a chip with excessive current draw and even considered buying a replacement. I found this RTC on AliExpress. It’s relatively cheap, but only sold in lots of 5 or 10. I was even about to order one, but then I remembered that one of the other users had measured a battery current of 1.33 µA on his unit.
I measured mine:
Meter off: 25 µA
Meter on: −15 µA

Yes, you read that right: current is flowing toward the battery. But it could just as easily be flowing toward the multimeter through the second diode.

1, 2. I desolder the chip and still get a current of about 30–32 µA, even with the chip removed. It gradually drops.
3. Just out of curiosity, I warm the area where the chip sits slightly with a soldering iron, without touching the circuit board itself. The current rises again.
4. I wait for the board to cool down. The current is about 26 µA with the chip removed; earlier, it was 25 µA with the chip installed. Allowing for the different temperature or something else, it seems that most of the current draw comes from leakage through the diode isolation circuit.


The diodes are on the other side of the board, under the display. The one closer to the edge handles power from the main supply, while the one farther in handles power from the battery.
First, I replaced the diode that handles power from the main supply. This can be done without removing the display.


In the end, my 61E doesn’t register any current at all. With the chip installed, the current draw is 1.22 µA!
As you can see in the photo above, the original diodes were marked SD. The only replacements I could find were Vishay S1D diodes, but they’re larger, rated for 200 V and 1 A, and clearly not the right type.
Checking them with a multimeter showed a forward voltage drop of about 0.12 V, which suggests they’re Schottky diodes. And this may be where the problem lies. Schottky diodes are very good, but they often have one drawback: high reverse current, which makes them unsuitable for this application.



There is also a second problem. I could have left the old diode in the battery circuit, but then the chip would always be powered by the battery, since the voltage drop across the diode from the main supply would clearly be greater than from the backup supply.
So the second diode also needs to be replaced. Both diodes should be 1N4148, as specified in the datasheet.

The chip gets 2.717 V, and the battery is at 2.991 V. It's quite old.


With power from the main supply, the voltage before the diode is 3.283 V. While replacing the diodes, I also spotted a convenient measurement point on the board, to the left and below the battery.
The RTC chip is getting 2.988 V, and the current in the backup battery circuit is zero.


But my curiosity got the better of me, so I bought a new 2032 battery at a nearby store. Its voltage turned out to be almost the same as the main supply: 3.225 V, a difference of just 60 mV.
That means that under a heavier load, the voltage drop across the main supply diode could increase, causing the chip to draw power from the backup battery too. That's what happened here, but the current was only 0.3 µA. I don't think that's critical, though it's cutting it close.


That wraps up the modification. The meter works, and the current drawn from the battery has dropped from 25 µA to 1.2 µA—a 20-fold reduction.


Conclusions.
I think the manufacturer originally used Schottky diodes to reduce the voltage drop in this circuit, even though ordinary 1N4148 diodes are more than adequate. Then complaints started coming in, they switched to 1N4148s, and everything was fine. I was just unlucky enough to buy a meter from that particular batch. Others may have the same problem, though, so I decided to post a description in case it helps someone.

Next, I want to replace the built-in battery with something newer, perhaps an 18650.
By the way, the link in the title goes to the cheapest UT181A listing on AliExpress.

And of course, a huge thank-you to infino and electronus—without their help, fixing the meter would have been much harder.