In the YouTube comments, there were a few posts lamenting that this device doesn't have a lithium ion battery, instead using 2AAA cells. Making matters worse, 1.2V rechargeables provide a dim display and early low-battery warning. I decided to use my new device as a demonstration of a lithium ion conversion. If I don't end up liking it, it can be fully reversed (at this point).
Disclaimer: If you decide to do this, it's at your own risk. It's not meant to be a full walk-through, but might be enough to inspire a hobbyist to give it a try on a device of any kind they want to "upgrade". Be aware that devices can be sensitive to static or induced voltages. Be sure your soldering iron is fully isolated and you practice static-safe procedures. Otherwise, damage may result to the components.
First, let me show you what the device looks like inside before any changes were made. As you can see, there are some open spaces especially on the right in the image below.
I started by planning the circuit. Below is my napkin sketch, and here is my parts list:
A 3.0V Low Dropout Voltage regulator to provide the same voltage as 2 AAA cells.
A lithium ion battery which fits into open space (it has a protection board).
A integrated charger/protection-circuit, with a resistor change to reduce charge current to match the cell capacity. For full reversibility, this will be in the battery bay instead of being installed inside. Yes, I'm doubly "protected".
One extra capacitor (1uF) to match the recommended circuit of the regulator (probably not required, there is a 10uF already on the PCB)
A connector to allow easy removal of the rear case. Not in the schematic because it was added later. It's on the left 3 wires.
Here's a picture of the TP4056 charging module. You can see the new resistor R3: the datasheet formula for constant current portion of the charge curve is 1200/R3. I only needed 3 wires because the (B+) and (OUT+) are common connections. I could have used 4 wires but this makes the cable more flexible overall.
Below is a view of the installed voltage regulator XC6206 - I just happened to have this SOT-23 component laying around. Having a low dropout voltage, it can supply almost the same output voltage as is input voltage, especially at the low current that is needed for the meter. It can handle 6V input and 200mA current, way more than I need. It is even laser-trimmed for ±1% output voltage. The current consumption is extremely low 1μA, so don't worry about battery drain, it may be less than the self-drain of the cell.
UPDATE after picture was taken: I later preferred more contrast from the screen, so I changed the 3.0V regulator (65Z5) to 3.3V (662K). Same component, higher voltage value.
The suggested regulator usage circuit includes a 1uF capacitor on the input (the TP4056 already has such caps) and a 1uF capacitor on the output (for stability under changing loads). There was an open set of pads on the PCB beside the stock 10uF MLCC, so I easily added a 1uF MLCC. I figured it should help reduce any possible noise from the relay switching faster than the 10uF stabilizes. Probably overkill, but it was fun to fill the spot shown with the yellow arrow! Be sure to use ceramic: Electrolytics and Tantalums have a leakage current and will slowly drain your battery.
Next, I put in Litihum Polymer cell, which is two 300 mA*h cells in parallel and a protection board. Meanwhile, I realized the need for a 3P connector. I avoided interfering features of the case and left room on the one side for a future charger module spot, should I decide to install the charger module permanently (and nix the connector). For now, the module is stuffed into the battery bay, and the wires go through existing holes in the case. Gave it a bench test to confirm 3.00V at the regulator output whilst the LiPo was at 3.78. As I mentioned above, this was changed to 3.3V.
Finally, I put some short heat-shrink strips around the charger module to gather the wires and protect metal contact. I left open space for viewing the charge LEDs and for the TP4056 to cool itself. I also more tightly arranged the wires to avoid the molded features of the black clamshell. In the end, looks less messy inside. I didn't adhere the LiPo to the screen - I only used a very thin piece of double-stick foam between the red case and the top edge of the cell bundle in the following view:
You may also notice I installed a 33 ohm resistor across the switching resistor for "always on backlight" at 133 (100+33) ohm total series resistance. I took this resistor back out the board after a successful experiment with how the mod works.
It lights the display at a reduced brightness, which increases when the regular backlight is turned on because my C-E shunt is bypassed by the transistor being activated (lowers resistance to the stock resistance 100 ohm - which is the resistor directly in-line with the 8 MHz crystal).
This feature may be useful if you plan to always plug in the USB and want a lit clock... I don't plan to use my device this way. If used, it does give the 613 a backlight when in clock-only mode, which is not a standard feature of the unit.
If I went this path, I would probably use a 50-100 ohm resistor for even lower output than the 33 provides. I could also mount a series switch on the shunting resistor and poke it through the case somewhere.
I removed the one battery terminal to make space for the charger module. the plate just slides up without a struggle. The other didn't get in the way, in fact it helps hold in the module.
When charging and using, I found that I achieved everything I was looking for and no downsides, except the battery bay door doesn't fit because of my wires.
I later found a way to get the entire micro plug to fit inside the bay by converting it to "right angle" entry. See third image below.
Charging...,
...Charging...,
...Charged!
Oh, and you can see the bumpers I put in the corners to keep the meter from sliding around both when flat on the table and when using the stand.
One day later, I decided to ensure I can close the battery door on the charging module. That way, any modification to the case is hidden. I removed the other AAA battery terminals for the purpose of completeness. This simply required straightening the solder tabs and pushing them out from the inside using tweezers.
Next, I needed to enlarge the hole. Below is the "before" picture where the wires are just big enough for the pass-thru.
Below you can see after whittle-ing away some material at the base. Now there is room for the wires plus the retention tab of the door.
The door fits nicely now, hiding the charging module which is easily accessed anytime I want to top off the charge. I guess you can call it a charging hatch?
I expect the lithium ion protection circuit to kick in around 2.8V, so I may never see the low battery warning symbol which appears around 2.5V. Perhaps I could put one of those thin stick-on battery 5-LED fuel gage devices in the charging bay as well, which would show the battery status at the touch of a button. There is a little more space.
Good result! With constant 3.3 V always supplied, the display always has a good contrast, as if new alkalines were installed. The backlight and beeper always have good output with the consistent voltage. And the clock/calendar won't lose it's settings (like it would with AAA battery changes) - as long as I don't let the charge go down to zero. Not likely to happen with the capacity cells I fit in there.
I forgot to mention another "mod" (done earlier in the week). I decided the continuity and especially the alarm beeper were not loud enough. I checked the current draw with the continuity constantly activated - around 12 mA. The dropper resistor was originally "101" or 100 ohms. With some calculating followed by experimenting, I found I could raise that to around 18 mA using a "560" or 56 ohm resistor, which was louder - but still a little bit shy of the 20 mA (or more) commonly used in ANENG's other meters. Now the alarm might actually wake me up!
Another bonus mod. I decided to add a kick-stand activated backlight. That is, when the kickstand is opened, the display is lit, no matter whether the meter is activated. The plan is 175 ohms to the LED normally (reduced brightness), 100 ohms when the backlight button is activated (full brightness).
I did this using a normally closed magnetic reed switch inside the unit, and a rectangular magnet foam taped to the inside of the stand. When the kickstand is closed, the reed switch is opened. When the kickstand is opened, the magnet moves away, allowing the reed switch to go to it's normally-closed state.
See the small magnet? Double-sided foam tape is underneath.
This is the way the reed switch was pre-tested to work properly. A normally-closed reed switch is actually a normally-open reed switch with a small magnet strapped on the side with heat-shrink tubing. This closes the reed switch inside. When another magnet is brought near, and oriented correctly, it will interfere with the little magnet and allow the reed switch to open back up. So it's important to pre-test the arrangement before affixing the switch, maintaining the original direction and rotation. Note the white paint at the end of the tube, it's not at the other end.
Hot glue is your friend. Doesn't stick to heat shrink glue as well as it does the case, so I put a second little blob near the top to help lock in the glue blob at the bottom. Once again, I used a connector to make it easy to separate the back case.
Here's the wiring. When the kickstand is opened, some current flows past the switching transistor, then through the reed switch, limited by an extra 75 ohms I added in (total 175 ohms). The transistor still works when activated by the DMM CPU, then the resistance goes down to 100 ohms for 1 minute before reverting back to the reduced brightness.
If the kickstand is closed, the backlight goes out and the unit works as it did before. Once again, this change is fully reversible.
Why do this? First off, it detects your intention to use the meter and turns on the backlight for you. Second, it's a reduced brightness, so it doesn't take away too much battery life. Finally, if you want to keep the time on display day and night, you can plug the charger module into a constant USB power source and enjoy a constantly-lit clock.
My goal here is to inspire you to follow through on your ideas for improving things. It's a great way to gain some satisfaction from your hobby. Just don't mess up any expensive equipment, OK?
Fantastic work Paul! ANENG take note. Lithium batteries are a good thing!