Finishing the LCR Meter + Switching to Battery Power

The LCR meter board has been sitting around for probably two years now. Everything would be fine, but it’s not very convenient to use in its half-finished state, so I decided to finish the case and rework the power supply while I was at it.


I already had a case, display, batteries, and a bunch of odds and ends for it, but I was so busy all the time that I never got around to it. Now I’m gradually finishing projects I started, and generally getting around to all the things I put off over the past year or year and a half.
As I go, I’ll link to reviews of the parts I used, as well as to a product that I haven’t reviewed—in fact, it’s not really needed; I’ll just show it along the way.




The LCR meter and case: I reviewed the former, and it became quite popular. I made a video about the latter. I have three versions of the case: a black one from AliExpress, a light-colored one from Banggood, and another black one that I bought locally. I decided to use the one from AliExpress.

At first I tried to fit the board in with the connectors, but eventually gave up for two reasons:
1. It got in the way of the indicator
2. I just don’t like BNC connectors. I don’t know, I’ve never liked them because they tend to lose contact over time.

By the way, one of the internal posts near the screw that holds the bottom part of the case was broken off. I find it strange not only that it’s there, but also that it could be broken off at all. The plastic is quite rigid, though, and reminds me of the polystyrene they used to make all sorts of equipment cases from.

I desoldered the connectors and the switch, but experience showed that desoldering the power connector was a mistake—I had to solder it back in later.


I sketched out the front panel, trying to fit everything into the available space. The seller had a version of the instrument in exactly the same case, but it used a different measurement board and a different keypad board. In my case, I used the newer version, which is the one sold everywhere now. If anyone needs it, I can upload the Front Designer file, but be sure to check your printer calibration.


Using a draft printout, I drilled the round holes and cut out the rectangular ones, then test-fitted everything. I used plastic standoffs here, but had to put washers under the display standoffs because I wanted it flush with the front panel.
People often cut an opening in the panel just for the screen, but I don’t like that approach because it leaves the display recessed too far. I usually bring the whole display out instead.


Next, I tried to find a place for the board. I considered two positions: the first seemed awkward, while the second shortened the wires to the keypad and especially the display, but required trimming the board itself and using long wires to the switch.


In the end, I settled on what is probably the classic layout. In this position, the probe connector sits right around the middle of the board, which makes soldering the wires very convenient.


The display and keypad were connected with a shielded eight-conductor cable. Shielding doesn’t matter for the keypad, but it helps with the display wires because this is a measuring instrument, and I really don’t need interference picked up by the cable.
I did have to squeeze things in and break one of the rules: I used the shield as both signal and power ground, since nine lines are the minimum needed to connect the display:
1. power
2. ground
3. 4 data lines
4. 3 control lines

I had to move the contrast adjustment resistor to the display board, so that made 8 wires plus the shield, which carries ground.


The plan was to power the instrument from batteries, not because I intend to take it anywhere, but for two reasons:
1. I don’t use it often, and the outlets are often occupied, so battery power will come in handy
2. The instrument is quite sensitive to its power supply, especially to noise, so batteries are the best choice for clean power.

But the batteries need to be charged somehow, and I definitely wanted the instrument to retain its 12 V input because:
1. A dedicated charger is always getting lost somewhere, and I saw no point at all in building one in.
2. You have to hunt around for power supplies with other voltages, but 12 V is always close at hand.

The tricky part was that the battery pack has a voltage range of 9–12.6 V, so using either a step-up or a step-down converter would be inconvenient. The former would cause problems when the batteries were discharged, while the latter wouldn’t be able to charge them fully.

The solution: use a SEPIC. Of the ones I had lying around at home, there were two: a small one based on the 34063 and a larger, factory-made one.
The first one worked, but not quite right, and the current was only about 200–240 mA, so I used the second one instead. It can charge at a higher current, and it also makes it possible to use an input voltage range of around 5–15 V, which greatly expands the range of power supplies that can be used.



A little about the modifications. I initially wanted to use a converter based on the 34063 controller because it can operate in CC mode on its own; all you need to do is choose the current-limiting resistors. But its low efficiency and low power output made me use a different converter, which I’ve already written about.
The modification involved adding a simple current-limiting circuit that provides CC mode.
I added all the components I actually used to the converter schematic (it was in the review); they’re highlighted in blue. I also cut one trace on the board between the output capacitor and the output pad.

1. The 1 Ω resistor sets the current. The voltage across it turns on the transistor, whose collector voltage is fed to the FB input of the board’s PWM controller. With the values shown, the charging current is about 600 mA.
2. The adjustable TL431 shunt regulator indicates when the battery is discharged. It’s set to 10 V, or 3.3 V per cell, so once the indicator comes on, there’s still enough time to finish working and then put the device on charge.
3. A common-cathode LED. I also had to use two 3.6 V zener diodes (anything from 3.3 to 4.7 V will work). Without them, the CC LED (green) won’t light because the voltage at the transistor’s collector will be only about 0.5 V. With the zener, it will be 0.5 V plus the zener voltage. One downside of this solution is that the current will only be limited when the output is around 4–5 V, but I have a 3S pack, so it’s not critical.
The zener diode is needed in the discharge indicator circuit because even when the TL431 is on (voltage is normal), it has 2.5 V across it, which makes the LED light when it shouldn’t. One or two diodes in series with the LED would have been enough, but the zener was closer at hand.

The circuit is simple, but it works reliably and without any problems. The green LED lights while charging and goes out when charging is complete. When the battery is discharged, the red LED lights up.
The circuit lets you charge the device while it’s running, effectively acting as a basic UPS. But since ripple could affect the device’s operation, it’s better not to do that. Of course, you can also simply power the device from an external power supply.


Yes, the assembly is a bit rough, especially the voltage-monitoring section, but I didn’t really feel like etching a board. I figured this would do, and in any case, it’s all reliable :)


For power, I used some almost no-name Chinese batteries that had been lying around for five years. I also reviewed them a while back.
Sure, you could say the batteries are junk: their actual capacity is about 2000 mAh. But I was pretty surprised that after sitting around for five years, they took only about 1 Ah during charging and delivered 2 Ah. In other words, their self-discharge over five years was only 50%, which is pretty great for no-name batteries. Their capacity was enough for me. I usually have only two or four branded ones, and there was no need for high-capacity or high-current batteries anyway. I did consider installing these with a capacity of around 3500 mAh, but why bother when even these will last the device through a full workday...

Two test runs showed that batteries 1, 3, and 4 had the highest capacity, but I used 1, 2, and 3 because battery 4 seemed to have high self-discharge at the beginning, that is, after charging was disconnected.


So, you could say the device is now assembled. The VATN display worked really well: crisp white characters on a black background. As I wrote in my review, where I showed how to modify it as a replacement for a standard 2004 display, it’s not OLED, of course, but it’s much better than a regular one. I recommend it.


The front panel is made from ordinary self-adhesive paper, with a strip of regular wide tape stuck over it. I’ve made a lot of panels for my devices this way, and I can say they look perfectly fine. The power supply I made more than five years ago still looks good; back then, I used ordinary paper and secured the edges with the same tape I used to protect the front. Self-adhesive paper is much more convenient, but be careful when printing. I accidentally printed the panels on the back and only noticed when I cut them out :(


A little about the probe connector.
I didn't like having four connectors, so I decided to replace them with a single multipin connector. Be careful here.
I used an 8-pin connector: 4 signal pins + 4 grounds. You need at least 6 pins—4 signals and 2 grounds.
The ground is shared only by the first three connectors. The fourth is electrically connected to the first, but follows a different trace and connects at a separate point.
Also, the pads closer to the components (highlighted in blue) carry only the ground for the fourth connector. The first three have a pad there, but it isn't connected to anything—apparently they simply forgot to connect it to ground. I connected to the contacts highlighted in red.


Since there's only one connector, I also had to think a bit about which pins to use. As you can see, the connections aren't just made at random: the input contact (Hc) is surrounded by the ground contacts of the other signals, which reduces interference from the other signal contacts, at least a little. In the second picture, red is signal and black is ground.
Drawing viewed from the front panel, i.e. from the outside contacts.


By the way, I used a standard 16 mm, 8-pin “aviation” connector. I bought it on AliExpress, but the overall quality of the connector is so-so. The contacts on the panel-mount part don't solder very well, and the cable-side part could lock in a little more firmly. Overall, I'd give it a C-minus.
They’re sold in sets, so now I have one spare panel-mount part. I could even build another device to use with these probes.


In my opinion, using a single multipin connector is much easier and more convenient than using four BNC connectors. And as I wrote, I really don't like how unreliable the BNC contacts are.


The four original cables absolutely refused to fit into the connector at first. They kept catching, and the insulation was getting cut, so I had to open it up a little with a needle file—or rather, just round off the sharp edges.
Everything else worked out fine. Just to be safe, I wrapped the signal cable with tape so its braid wouldn't touch the braids of the other cables.
I’m also wondering whether I should connect the connector body to ground. In theory, the ground for the fourth connector is connected there; it has a ground symbol on the circuit board.


Now I have two convenient cables and eight BNC plugs that I have no use for, but can't bring myself to throw away...


Demonstration in operation, with the probes shorted and a 1 mΩ resistor.


Can I say that's it? No, because I have an idea to add a compartment inside the case for storing the probes. The case is very tall for this board, and roughly half the height is just sitting unused. I want to cut an opening in the back and make a sliding-shutter door, then form a compartment inside so the wires can't get where they shouldn't and damage the internals.
That way, the wires will always be close at hand and won't take up space in the cupboard—and they'll be harder to lose, too.

As for the device itself, that's pretty much it: it works. I should probably do a full calibration just to be safe, although it already works as it is.