Simple Compact 5v Voltage Regulator With Low Battery Alarm




Hey Guys

In this instructable I will show you how to make a very nice 5v voltage regulator which will also warns you, in a very annoying manner, that your battery is running out. This is based on MicroScream circuit, published back in 2004. I came up with the compact layout and the addition of an actual regulator. See the attached video for the final product.

A bit of history: I used to fly DLG's, which stands for Discuss Launched Gliders. This is a very addictive branch of the RC modelling sport. The idea is to throw the glider by the wingtip as high as you can and then stay up for a designated period of time. You can find bunch of movies on YouTube, highly recommended. These glider are pure hi tech - composite structure, advanced airfoils and costly equipment inside. They are made as light as possible. There was a time that we used 2s LiPo batteries regulated down to 5v as an on board power. This is when I came with this regulator. The device itself is a voltage regulator, with smoothing capacitors on input and output. It is paired with small, two transistors + 3 resistors + buzzer, circuit that warns you if the main battery, 2s LiPo in this case, drains below a certain value.

The unit weights only 5gr ready to use. It is a linear regulator - no RF interference, and can handle up to 1A of current and it is an analog circuit. I made and sold around 200 of these, and I would like to think that somewhere there are several of these still in the air :).

I don't make them anymore, but I though that someone could benefit from the simplicity and the compact size which can be used not just for airborne purposes.

I tried to make this guide as clear as possible, hope you like it.

Enjoy, Dani

Check what G is doing today :)

Step 1: Schematics, Preliminarities and Parts List

Please note, I got the capacitors mixed up in the schematics: 22uF should be on the output, 0.1uF on the input. Just like in the description.

Thank you Josehf Murchison!

This circuit regulates 2s LiPo (7.2v-8.4v) down to 5v constant voltage. The alarm sounds once the input voltage drops below 7.2v=3.6v per cell. Note that it monitors input voltage only, so in any case you will always get regulated 5v as output.

First thing first, decide on the alarm voltage. This is an analog device, so everything should be set in front. The way to adjust the alarm voltage is to change R1. In this case: R1+R2=4.7+2.49=7.19 - slightly below the desired 7.2v. Go ahead and change R1 if you want something else, keep R2 as it is. Keep it slightly below the desired value to allow for hysteresis.

Second, part list:

1. Voltage regulator, AMS1117 5.0V. There is also 3.3V version of this IC, so choose the right one.
Data sheet:

Where to buy:

2. 2N7000 mosfet

Data sheet:

Where to buy:

3. TL431 shunt regulator

Data sheet:

Where to buy:

4. Buzzer, any 5v active buzzer will work

Try this one:

5. Proto board:

Or any leftovers that you have at home.

Resistors and capacitors you can choose out of many affordable sets.

Step 2: Prep Your Board, Solder Regulating Circuit

First thing is to cut your board parallel to the short side, as to have strip with 5 holes width.

This strip is enough to make two regulators, because this is so simple and convenient, I'm showing the full procedure for two units. The first picture shows the components for two circuits.

Second, AMS1117 is the only SMD component in this circuit, it needs prep with small drops of tin on the board, note the 3x2 tinned points.

Third, remove one copper dot from each side (one for each regulator).

Next, put the regulator on top of the dots and head with soldering iron, it will melt the underlying tin and will connect the regulator.

Next, insert the two capacitors and the R1 through holes and bend towards the AMS1117. 0.1uf goes on the input, 22uf goes on the output terminals. Pay attention here, electrolytic capacitors (22uf in our case) have polarity.

Solder together with the AMS terminals, don't be shy on the tin here.

The regulating circuit is ready.

Step 3: Solder on the Resistors

Add the remaining resistors. Note the shift in holes.

The R1 and R3 are both connected to the + terminal (that is why we have the shift with R2)

Step 4: Add the Semiconductors

First, bend the legs of the transistors like shown in the picture.

Insert the legs through the holes, note that they share two of the three legs.

Connect the legs like shown in the pictures, follow the diagram for clarity.

Step 5: Cut the Board, Connect the Buzzers, Connect the +

Now it is time to cut the board in half, remember that we are making two units.

Bend the buzzer's legs and solder them as shown.

Note that the drain leg from 2N7000 goes directly to the negative of the buzzer.

Finally, connect a short wire from the battery + to buzzer + (green wire in picture)

The unit is ready.

The video shows it's functionality. Note that the output voltage is not affected by the signal and it is always stable. In order to stop the alarm you need to recharge the batteries

Step 6: Add Leads, Shrink, Use

Now, add suitable leads with your choice of connectors. Note that the negative leg is common.

In my case: female JST on the input and JR servo on the output.

Add a stress releaf near the solder points of the leads - dob of silicone or hot glue.

Shrink the whole unit, allow the hole of the buzzer exposed in order to buzz.

You are all set!

Enjoy, Dani



  • Classroom Science Contest

    Classroom Science Contest
  • Colors of the Rainbow Contest

    Colors of the Rainbow Contest
  • Beauty Tips Contest

    Beauty Tips Contest

11 Discussions


2 years ago

Great Indestructible, and thanks for sharing!


Nice project. Really compact.

Not to be a party pooper, but...

8.4V - 5V = 3.4V.

3.4V * 1A = 3.4W

I don't think there is any possibility that your ASM117 sot223 (with the heat sink pin touching, but not soldered to only one, single sided copper pad), would ever be capable of dissipating that kind of heat...

(The datasheet says 45-90 degC/W depending on ground plane, of which you have virtually none).

Have you tested it at 8.4V/1A for any significant time? Did the AMS117 shutdown/melt (seriously).

If nothing else, consider soldering a piece of the largest gauge solid copper wire that will fit through the he holes in your perfboard and "weaving" it up/down through the four vacant holes in front of the heat sink. That'll give you something.

Now try to draw an amp. Might work. The wire will get warm!

3 replies

Well, 1A is the property of the IC from the data sheet, so I consider this as my ceiling. As I mentioned, I did more than 200 of these, not a single failure - for the application they were intended. It means being in a VERY tiny place stuffed with foam/no heat dissipation and driving 4 servos + receiver + occasional altimeter continuously :).

I did a bench test once when I connected 4 servos which ran continuously for about 4 hours until the battery drained. I snugged the regulator in a foam and under several rugs to see how it handles. In addition, I stalled the servos here and there. It survived with no problem, and then I used it with no issues.

AMS1117 is very nice in this manner. Once I reversed the polarity on one, intentionally, it made a small hole in the middle, which was very funny. Otherwise no failures at all.

Enjoy, Dani

No offense, but - from experience, good engineering is more than, "it worked that one time I tested it." Especially if you're going to make claims and sell it.

You need to make sure your design is valid under all conditions, against worst case vendor specs.

5V, 1A in the specs has caveats. Caveats you must understand before making promises to customers!

My guess is that you got by in your test case, because you didn't have a full 8.2V lipo, and never drew a full amp more than briefly.
And you weren't monitoring the OUTPUT voltage. What size lack was it? That'll give us an idea of average load.

If you really want to test your spec get five 25ohm resistors (at least 1W each) in parallel on the output. That'll make your 1A load. Monitor the voltage. I'll bet you 99.44% somethings going to fail - quickly.

First, remember that linear regulators are basically variable resistors. They decrease the voltage - as waste heat. Lots of heat. 3W max, in this case. That doesn't sound like much, but it is when it's coming from a tiny part. I overloaded a linear regulator once enough that the solder melted and it fell off the board!

If you look at the specs, you see some interesting things.

1 - thermally regulated
If it overheats, it will "limit" in some way.
Maybe decrease output voltage in an attempt to decrease output current.
You're not monitoring the output voltage, so we don't know if it's stable at 1A.

2 - max junction (silicon) temperature 125dC.

3 - thermal impedance of sot23 is 45/90dC/W -> if heat sink tab attached to ground plane.
That's the kicker.

Thermal impedance is rated in different ways. Junction to case, case to ambient etc.

You don't have the heat sink attached to anything. They don't specify junction to ambient w/ no ground plane, but it's certainly MUCH higher than 90dC/W.

But for the sake of my explanation, let's assume it's only 100dC/W.

Let's say we're in free air, 25dC, no foam etc. We've got over 3W of heat going on. Work backwards. 25 + (3 * 100) is 325dC...

The case is going to be REALLY hot, but the silicon inside is going to be VERY ANGRILY hot.

So, how would you make this work?

1 - today, you can design it as a " quiet" switcher.
2 - Use a TO-3 package. They have much better case/ambient impedance.
3 - use an etched board with a large, unbroken copper area that the heat sink tab is soldered to. The copper will radiate the heat. Note this still may not be enough!
4 - here's what I would consider.
You know you're only going to let the batteries go down to ~7V.
Dropout voltage of ams117 is 1.3V, so you only need ~6.3V going into it.
Between you're battery sense point, and the regulator input add a 1A, 1W (or 2W) diode (regular, not shot toy). That adds a 0.7V drop and wastes almost 1W of your 3W outside the ams117.


2 years ago

WannaDuino!!! Likes iT!!!

Amazing, i wil build it also. I like your IBLE.



2 years ago

Awesome job. I always have a tough time getting perf board projects to be compact and look nice - you succeeded in both accounts!

2 replies

Reply 2 years ago

I too had the same problem. I think it has to do with trying to do the layout while looking at the circuit diagram. If you reconstruct the circuit as a ladder diagram you will see the common connections in a more convenient configuration.


Reply 2 years ago

Thanks man. That's highly appreciated!


2 years ago

Nice circuit. If you were to used SMD devices you could get it even smaller and lighter. I used to build circuits when I flew R/C back in the day. I used TL431 adjustable regulators for battery cycling circuits for cycling NiCads. Thumbs Up.

1 reply

Reply 2 years ago

Thank you, you are right of course.

I tried to optimize between SMD and through hole components in order to maintain my sanity :) after all it is made by hand.