DIY Oscilloscope—Really? Yes! DSO138 Oscilloscope Kit

I recently reviewed another kit, and today I'm continuing my short series of reviews of various DIY projects for beginner radio enthusiasts.
I'll say right away: this is certainly no Tektronix, or even a DS203, but it's an interesting little gadget in its own way, even if it is basically a toy.
Usually, before testing something, I take it apart first. Here, I have to put it together :)

In my opinion, an oscilloscope is a radio enthusiast's “eyes.” This instrument rarely offers high accuracy, unlike a multimeter, but it lets you see processes as they unfold—in “motion.”
Sometimes a quick “glance” like that can be more helpful than spending a whole day poking around with a tester.

Oscilloscopes used to be built with vacuum tubes, then transistor models took over, but the results were still displayed on a CRT. Eventually, digital models replaced them: smaller and lighter, and the logical next step was a kit for building one yourself.
A few years ago, I came across attempts on some forums to develop a DIY oscilloscope—some of them successful. A kit is certainly simpler and has lower specs, but I can say with confidence that even a school student could assemble it.
This kit was developed by jyetech. Here's the page for this instrument on the manufacturer's website.

Specialists may find this review overly detailed, but my experience talking with beginner radio enthusiasts has shown that they find information easier to understand this way.

Anyway, I'll get into all that a little later. First, the usual introduction: unboxing.

The kit arrived in a regular zip-lock bag, though a fairly sturdy one.
I think a kit like this would really benefit from some nice packaging. Not to protect it from damage, but simply to make it look good. After all, even unboxing should be a pleasant experience—it's a kit.


The bag contained:
Instructions
Circuit board
Cable for connecting to the circuits under test
Two bags of components
Display.


The device's specs are very modest. To me, this is more of an educational kit than a measuring instrument, although you can still take measurements with it, even if they're simple ones.


The kit also includes a detailed, full-color instruction sheet printed on both sides.
It covers the assembly sequence, calibration, and a brief user guide.
The only downside is that it's all in English, but the illustrations are clear, so most of it should still be understandable.
The instructions even show where each component goes and include checkboxes to tick off when you finish each step. Very well thought out.


A separate sheet lists the SMD components.
It's worth noting that there are at least two versions of the device. On the first, only the microcontroller is pre-soldered; on the second, all the SMD components are already soldered in place.
The first version is intended for slightly more experienced users.
My review covers this version; I only learned about the other one later.


The circuit board is double-sided, just like the one in my previous review—and even the color is the same.
The component markings are printed on the solder mask. Some components are labeled in full, while others are identified only by their reference designators from the schematic.


There are no markings on the back, just the jumper labels and the device model name.
The board is coated with solder mask, and it’s very tough (I had to find that out the hard way). In my opinion, that’s just what beginners need, since it’s hard to damage anything during assembly.


As I mentioned above, the board is labeled with the locations of the components to be installed. The markings are clear, so I have no complaints on that front.


All the pads are tinned, and the board is very easy to solder—well, almost easy. I’ll explain that little wrinkle in the assembly section :)


As I mentioned above, the board comes with an STM32F103C8 microcontroller preinstalled.
It’s a 32-bit microcontroller based on an ARM 32-bit Cortex™-M3 core.
It runs at up to 72 MHz and also has two 12-bit, 1 µs ADCs.


The model number, DSO138, is printed on both sides of the board.


Let’s get back to the list of components.
The small electronic components, connectors, and so on are packed in little zip-lock bags.


We empty the contents of the large bag onto the table. Inside are connectors, standoffs, and electrolytic capacitors. There are also two more little bags in there :)


Once we open all the bags, we find quite a lot of components. Though, considering this is a digital oscilloscope, I expected more.
One nice touch is that the SMD resistors are labeled. I do think it would have been helpful to label the regular resistors too, or include a small color-code reference card.


The display is wrapped in a soft material. As it turns out, it doesn’t slide around, so it won’t rattle around in the bag, and the circuit board protects it from damage during shipping.
Still, I think proper packaging wouldn’t have gone amiss.


The device uses a 2.4-inch TFT LCD with LED backlighting.
The screen resolution is 320 × 240 pixels.


A short cable is also included. It has a standard BNC connector for connecting to the oscilloscope, with a pair of alligator clips at the other end.
The cable is moderately flexible, and the alligator clips are fairly large.


And here’s the whole kit laid out.


Now we can move on to assembling the kit itself, and see just how difficult it is.

Last time, I started assembly with the resistors, since they were the lowest-profile components on the board.
When there are SMD components, it’s best to start with those.
I laid out all the SMD components on the included sheet, labeling each with its value and reference designator from the schematic.


Just as I was about to start soldering, I thought that components in a package this small for a beginner could easily have been 1206 resistors instead of 0805. They take up only a little more space, but are easier to solder.
My second thought was: what if I lose a resistor and can’t find it? I can open the drawer and get another one, but not everyone has that option. In this case, the manufacturer has thought of that.
They included one extra of every resistor (too bad they didn’t do the same for the ICs), as a spare. Very thoughtful—kudos.


Next, I’ll explain how I solder components like these and what I recommend to others. This is just my opinion, of course—everyone can do it their own way.
SMD components are sometimes soldered using special paste, but beginners (and plenty of experienced hobbyists, too) don’t often have any on hand, so I’ll show you an easy way to work without it.
Pick up a component with tweezers and place it where it goes.


I often apply flux to the component’s pads first. It makes soldering easier, but cleaning the board harder—sometimes it’s difficult to wash the flux out from under a component.
So in this case I just used 1 mm flux-core solder.
Hold the component in place with tweezers, pick up a small drop of solder on the iron tip, and solder one side of the component.
Don’t worry if the joint looks ugly or isn’t very strong; at this stage, it’s enough for the component to stay in place on its own.
Then repeat the process with the remaining components.
Once you’ve tacked down all the components this way (or all the components of one value), you can solder them properly. Turn the board so the side you’ve already soldered is on the left. Holding the soldering iron in your right hand (if you’re right-handed) and the solder in your left, solder all the remaining joints. If you’re not happy with the joints on the second side, turn the board 180 degrees and solder the other side of the component in the same way.
This is easier and faster than soldering each component individually.


Here you can see several resistors in place, but soldered on only one side so far.


SMD ICs are marked just like through-hole ones: the first pin is to the left of the mark (usually at the lower left when looking at the marking), and the remaining pins are numbered counterclockwise.
The photo shows the IC footprint and an example of how the chip should be installed.


We handle ICs exactly the same way as the resistors.
Position the IC on the pads, solder any one pin (preferably an outer one), adjust the IC's position a little if needed, then solder the remaining pins.
There are different ways to solder the regulator IC, but I recommend soldering the tab first and then the pads, so the IC will sit flat against the board.
But there's nothing stopping you from soldering an outer pin first and then all the others.


All the SMD components are installed and soldered. There are just a few resistors left, one of each value; I'll put them in a little bag in case they come in handy someday.


Now let's install the through-hole resistors.
In my previous review, I talked a little about color codes. This time, I'd recommend simply measuring the resistance with a multimeter.
The thing is, these resistors are tiny, and at this size the color bands are very hard to read (the smaller the colored area, the harder it is to identify the color).
At first, I looked in the instructions for a list of values and reference designators, but couldn't find one because I was looking for a table. Only after assembly did I realize they were shown in the illustrations, complete with checkboxes for marking the installed parts.
Because I hadn't looked carefully enough, I had to make my own table and lay out the components to be installed next to it.
The resistor visible separately on the left was left over when I made the table, so I saved it for last.


As in the previous review, shape the resistor leads with tweezers (or a forming tool) so the resistor fits easily into place.
Be careful: the component reference designators on the board may be printed not only above the parts, but also underneath them, which can trip you up—especially when there are lots of components in a row.


Here's a small drawback of the circuit board.
The holes for the resistors are very large, and since the layout is fairly dense, I decided to bend the leads—but only slightly. As a result, they don't stay in place very well in holes this large.


Since the resistors didn't stay in place very well, I recommend not loading all the values at once. Install half or a third, solder them, then install the rest.
Don't be afraid to trim the leads short. A double-sided plated-through board is forgiving—you can always solder a resistor from the top if needed, which you can't do with a single-sided board.


All the resistors are soldered in, so on to the capacitors.
I handled them the same way as the resistors, sorting them according to the table.
By the way, I did end up with one spare resistor; apparently they included it by mistake.


A few words about the markings.
These capacitors are marked the same way as resistors.
The first two digits are the value, and the third digit is the number of zeros to add.
The result is the capacitance in picofarads.
But this board also has capacitors that don't follow this marking scheme: the 1, 3, and 22 pF values.
They're marked with the capacitance value directly, since it's less than 100 pF—that is, less than a three-digit number.


First I solder in the small capacitors according to their reference designators (quite a puzzle).


I messed up a little with the 100 nF capacitors by not adding them to the table right away, so I had to write them in by hand later.


I bent the capacitor leads only partway, to about 45 degrees. That's enough to keep the component from falling out.
By the way, you can see in this photo that the pads connected to the board's ground plane are designed correctly: there's a thermal relief gap to reduce heat transfer, which makes these joints easier to solder.


I got a little too relaxed with this board and only remembered the inductors and diodes after soldering in the ceramic capacitors, even though it would have been better to install them first.
But that didn't really change much, so let's move on to them.
The kit included three inductors and two diodes (1N4007 and 1N5815).


The diodes are straightforward: the locations are labeled, and the cathode is marked with a white stripe on both the diode and the board, so it's hard to get them backwards.
Inductors can be a little trickier; they sometimes have color codes too. Luckily, all three inductors in this case have the same value :)


On the board, inductors are marked with the letter L and a wavy line.
Here's the section of the board with the inductors and diodes soldered in.


The oscilloscope uses two transistors of different polarities and two voltage regulator ICs, one for each polarity. Be careful when installing them: the markings 78L05 and 79L05 look very similar, but if you put one in the wrong way around, you’ll probably be off to buy replacements.
The transistors are a little easier. The board only indicates their polarity, not the transistor type, but you can easily check the type and reference designator in the schematic or component placement diagram.
Forming the leads is noticeably harder here, since all three leads need to be bent. Take your time so you don’t break them.


The leads are bent the same way, which makes things easier.
The transistor and regulator positions are marked on the board, but just to be safe, I took a photo showing how they should be installed.


The kit included a fairly powerful inductor (relatively speaking), used in the converter to generate the negative voltage, and a crystal resonator.
Their leads don’t need bending.


Now, about the crystal resonator: it’s rated for 8 MHz and has no polarity, but it’s best to put a small piece of tape underneath it because its metal case rests on the traces. The board has a protective solder mask, but I’m used to putting some kind of insulating layer underneath in cases like this, just to be safe.
Don’t be surprised that I said earlier the processor’s maximum frequency is 72 MHz, while the crystal is only 8 MHz. The processor has frequency dividers and sometimes multipliers inside, so the core can easily run, for example, at 8 × 8 = 64 MHz.
For some reason, the inductor pads on the board are square and round, even though the inductor itself is non-polarized, so just solder it in place. It’s best not to bend the leads.


The kit came with quite a few electrolytic capacitors. They’re all rated at 100 µF and 16 V.
They must be soldered in with the correct polarity, or you might get some pyrotechnics :)
The capacitor’s long lead is the positive terminal. The board marks the polarity both by the corresponding lead and next to the circle showing the capacitor’s position, which is quite handy.
The positive lead is marked. Sometimes it’s the negative lead that’s marked, in which case roughly half the circle is shaded. And then there’s Asus, a computer hardware manufacturer that shades the positive side, so always pay attention.


We’re gradually getting to a fairly uncommon component: the trimmer capacitor.
It’s a capacitor whose capacitance can be adjusted over a small range, say 10–30 pF. These capacitors usually have a low capacitance, up to 40–50 pF.
It’s a non-polarized component, so formally it doesn’t matter which way you solder it in—but sometimes it does.
The capacitor has a screwdriver slot (like the head of a tiny screw), which is electrically connected to one of its leads. In this circuit, one lead is connected to the board’s ground plane, while the other is connected to the rest of the components.
To minimize the screwdriver’s effect on the circuit, solder it in so that the lead connected to the slot is connected to the board’s ground.
The board shows which way to install it, and later in the review there’ll be a photo showing this.


Buttons and switches.
It's hard to get anything wrong here, since it's pretty difficult to install them the wrong way around :)
Just make sure to solder the switch body terminals to the board.
For the switch, this not only adds strength, but also connects the switch body to the board's common ground, so the body acts as a shield against interference.


Connectors.
The trickiest part to solder. Not because the components are small or require precision; on the contrary, the soldering point can be hard to heat up, so it's best to use a more powerful soldering iron for the BNC connector.


In the photo you can see—
the BNC connector, the additional power connector (the only connector here that can be installed backwards), and the USB connector being soldered.


There was a small snag with the display—or rather, with its connectors.
The kit was missing a pair of double contacts (pins), which are used here to secure the side of the display opposite the signal connector.


But after looking at the signal connector pinout, I realized I could easily snip off a few contacts and use them in place of the missing ones.
I could have opened my desk drawer and taken out a connector like that, but it wouldn't have been interesting and, in a way, wouldn't have been fair.


Solder the female connector parts to the board.


The board has an output from the built-in 1 kHz generator, which we'll need later. Although these two contacts are connected to each other, we still solder in a jumper; it'll make it easier to connect the signal cable's alligator clip.
A clipped-off lead from an electrolytic capacitor works well for the jumper; they're long and fairly stiff.
The jumper is to the left of the power connector.


There are also a couple of important jumpers on the board.
One of them, labeled JP3, needs to be shorted right away. Just bridge it with a drop of solder.


The second jumper is a little more complicated.
First, set the multimeter to measure voltage and connect it to the test point above the regulator IC's tab. Connect the other probe to any point connected to the board's common ground, such as the USB connector.
Power up the board and check the voltage at the test point. If everything is OK, it should be around 3.3 V.


After that, jumper JP4, located slightly to the left and below the regulator, also needs to be bridged with a drop of solder.


There are four more jumpers on the back of the board. Leave them alone: they are for board diagnostics and putting the processor into programming mode.


Back to the display. As I mentioned above, I had to snip off several pairs of contacts to use in place of the missing ones.
When assembling it, though, I decided to remove pairs from somewhere near the middle rather than at the ends, and solder the end pair in place. That way, it will be harder to mix anything up during installation.


Although the display has a protective film on it, I recommend covering the screen with a piece of paper while soldering the connector. That way, any drops of boiling flux will land on the paper instead of the screen.


That's it, we can power it up and test it :)
By the way, one of the diodes we soldered earlier protects the electronics against reverse polarity. That's a useful touch from the designer, since it only takes a second to fry the board by connecting the power supply the wrong way around.
The board is marked for a 9 V supply, but the specified range goes up to 12 V.
For my tests, I powered the board from a 12 V power supply, but I also tried two lithium batteries connected in series. The only difference was slightly dimmer screen backlighting. I think that with a low-dropout 5 V regulator and the protection diode removed (or connected in parallel with the power supply, with a fuse installed), the board can be powered quite safely from two lithium batteries.
Another option is to use a 3.7–5 V power converter.


Since the board powered up successfully, it's best to clean it before calibration.
I use acetone. It's banned from sale, but I have a small supply. Toluene has also been used; medical alcohol is another option, as a last resort.
But the board definitely needs cleaning. There's no need to dunk the whole thing; just wipe the underside with a cotton swab.


Pay special attention to the operating-mode switches and the input connector.
The frequencies aren't very high, but the parasitic resistance caused by flux can still cause problems.


Finally, put the board up on its “legs” using the included standoffs. They're a little shorter than they should be and wobble a bit, but this is still more convenient than just laying the board on the table. Besides, the component leads could scratch the tabletop, and this way nothing can get under the board and short something out.


First, let's test the built-in signal generator. Connect the red-insulated alligator clip to the jumper near the power connector; the black wire doesn't need to be connected anywhere.


I almost forgot: a few words about what the switches and buttons do.
There are three three-position switches on the left.
The top switch selects the input mode.
Grounded
AC mode, which blocks the DC component, also known as AC coupling or a closed-input mode. It's well suited to measuring AC signals.
DC mode, which allows the DC component to be measured, also known as an open-input mode. It lets you measure voltage including its DC component.

The second and third switches select the voltage scale.
If 1 volt is selected, one vertical division on the screen represents 1 V.
The middle switch selects the voltage, and the bottom switch selects the multiplier. Together, the three switches provide nine fixed voltage ranges, from 10 mV to 5 V per division.

The buttons on the right control the sweep and operating modes.
Button descriptions, from top to bottom:
1. A short press turns on HOLD mode, freezing the readings on the display. A long press (more than 3 seconds) turns the digital signal-parameter readout on or off: frequency, period, and voltage.
2. Increases the selected parameter.
3. Decreases the selected parameter.
4. Cycles through the operating modes.
Sets the sweep time, from 10 µs to 500 s.
Selects the trigger mode: Auto, Normal, or Single.
Selects the trigger edge: rising or falling.
Sets the trigger level.
Scrolls the waveform horizontally, letting you view the signal beyond the edge of the screen.
Sets the waveform's vertical position. This helps when measuring signal voltage and when the waveform doesn't fit on the screen…
The reset button simply restarts the oscilloscope. As it turns out, that can come in handy sometimes.
There's a green LED next to the button. It flashes when the oscilloscope has synchronized to the signal.

All settings are saved when the instrument is switched off, and it starts up in the same mode it was in when switched off.

There's also a USB connector on the board, but as far as I can tell, it isn't used in this version. When connected to a computer, it shows up as an unknown device.
There are also contacts for reflashing the device.

All modes selected with the buttons or switches are also shown on the oscilloscope display.


I haven't updated the firmware, since the latest version available at the moment, 113-13801-042, is already installed.


Calibrating the instrument is very simple, thanks to the built-in signal generator.
When you connect it to the built-in square-wave generator, you'll most likely see something like this: instead of clean square waves, the top or bottom corners will be rounded off, either upward or downward.


Adjust this by turning the trimmer capacitors.
There are two capacitors: adjust C4 in the 0.1 V mode and C6 in the 1 V mode. No adjustment is needed in the 10 mV mode.


Adjust until the square waves on the screen look clean, as shown in the photo.


I looked at this signal on another oscilloscope; in my opinion, it’s stable enough to calibrate this oscilloscope.


Although the capacitors are installed correctly, even in this configuration the metal screwdriver has a slight effect. As long as the tip is touching the adjustment component, the reading stays the same; as soon as I lift it, the reading changes slightly.
In this case, you can either make tiny adjustments or use a plastic (insulated) screwdriver.
I got this screwdriver with some Hikvision camera.


One end has a Phillips tip, cut down specifically for these capacitors; the other end is flat.


Since this oscilloscope is more of a learning tool for studying how things work than a fully fledged instrument, I don’t see the point in doing a full-blown test. Still, I’ll show and check the basics.
1. I completely forgot: sometimes the manufacturer’s ad pops up at the bottom of the screen while it’s running :)
2. Digital signal parameter readings. A signal from the built-in square-wave generator is applied.
3. This is the oscilloscope’s own input noise. I’ve seen mentions of it online, as well as claims that the newer version has a lower noise level.
4. To check that this really is noise from the analog section and not interference, I switched the oscilloscope to shorted-input mode.


1. I switched the timebase to 500 s per division. In my opinion, that’s really only for extreme cases.
2. The input signal level can be adjusted from 10 mV per division
3. Up to 5 V per division.
4. A 10 kHz square wave from the DS203 oscilloscope’s generator.


1. A 50 kHz square wave from the DS203 oscilloscope’s generator. You can see that the signal is already badly distorted at this frequency. There’s not much point in trying 100 kHz.
2. A 20 kHz sine wave from the DS203 oscilloscope’s generator.
3. A 20 kHz triangle wave from the DS203 oscilloscope’s generator.
4. A 20 kHz sawtooth wave from the DS203 oscilloscope’s generator.


Next, I decided to take a closer look at how the instrument handles a sine wave from an analog generator and compare it with my DS203.
1. Frequency: 1 kHz
2. Frequency: 10 kHz


1. Frequency: 100 kHz. The kit doesn’t let you set the timebase below 10 ms, so this is the best I can do :(
2. This is what a 20 kHz sine wave from the DS203 can look like in a different input attenuator setting. Above is a screenshot of the same signal with the attenuator set to 1 V × 1; here it’s set to 0.1 V × 5.
Below you can see what this signal looks like when displayed on the DS203.


A 20 kHz signal from an analog signal generator.


A side-by-side photo of the two oscilloscopes, the DSO138 and DS203. Both are connected to an analog sine-wave generator set to 20 kHz, and both oscilloscopes are set to the same operating mode.


Summary.
Pros
An interesting educational kit
A well-made circuit board with a durable protective coating.
Even a beginner radio enthusiast can assemble this kit.
Thoughtful contents; I was pleased to find spare resistors included.
The instructions explain the assembly process well.

Cons
Limited input signal frequency.
A couple of contacts for mounting the display were missing from the kit.
Basic packaging.

My take: to put it briefly, if I'd had a kit like this as a kid, I probably would have been very happy, despite its shortcomings.
To elaborate, I was pleasantly surprised by this kit. I think it's a good way to gain experience assembling and setting up an electronic device, as well as learning to use a very important tool for radio enthusiasts—the oscilloscope. It may be basic, with no memory and a low frequency range, but it's still much better than fiddling around with sound cards.
Of course, you can't consider it a serious instrument, but it's not marketed as one. As a kit, it's more than adequate.
Why did I order this kit? I was simply curious—we all love toys :)

I hope you found this review interesting and useful. I'm looking forward to your suggestions for things to test :)
And, as always, you can download additional materials, firmware, instructions, source files, the schematic, and description.
The schematic is also available separately.

The product was provided by the store for review.