Small, simple oscilloscope

Some time ago, I published a review of a kit for building a simple oscilloscope. It would have been fine, except its bandwidth was very limited. I wouldn't say the little device I'm reviewing here is much better, but its specified bandwidth goes up to 4 MHz.
Let's see what it is and how it works :)



This oscilloscope is marketed as a learning tool right from the seller's product page. In other words, it's intended for beginners with little experience, who are unfamiliar with more "advanced" models and could easily get confused by them.

First, the specifications. Apologies in advance for the occasionally awkward translation.
Sampling rate: 20MSa/s
Analog bandwidth: 4 MHz
Sampling resolution: 8 bits
Buffer size: 650 bytes
Vertical sensitivity: 10 mV/div ~ 5 V/div (1-2-5 sequence)
Horizontal sweep speed: 1.5us/div ~ 6ms/div (1-2-5 sequence)
Display: 2.4-inch TFT, 320X240 (ILI9325 driver chip)
Input resistance: 1MΩ
Maximum input voltage: 40Vpp (1:1 probe), 400Vpp (10:1 probe)
Input signal: AC
Waveform hold function (HOLD)

The oscilloscope is sold in several configurations (prices are approximate, from the product page):
1. Oscilloscope + USB power cable - $17.40
2. Oscilloscope + power cable + probe - $20.09

You can also buy these separately:
3. 40 MHz probe - $2.69
4. BNC-BNC cable + BNC-to-alligator clips - $1.74

I ordered option number two. In the end, I received these two packages.
By the way, I ordered the oscilloscope from the same seller as the LCR meter, which brought the shipping cost down a little.


Since shipping cost depends on weight, I first weighed the option 1 kit, then the full option 2 package.


Let's take a look, starting with what's included.
The large package contained the probe, power cable, and a few odds and ends. The bag is sturdy and has a "flap," making it handy for storing all the extra "stuff" later on.


The power cable has a standard USB plug at one end and a 3.5 mm round plug at the other.
The probe is a standard one, with a flexible cable.


The kit included a manual. After reading it, I realized the kit wasn't quite complete after all: it was missing a special spring ground contact and four colored rings. The rings aren't a big deal; I can live without them, but I could really have used that extra contact :(


The probe has a built-in 1:10 attenuator with a corresponding switch. The ground lead is insulated, though the alligator clip is pretty stiff.
As I mentioned above, according to the manual, my probe is rated for frequencies up to 40 MHz and voltages up to 600 V. The oscilloscope itself has more modest limits, so there's plenty of headroom here.


The probe can be compensated, but since I tested it at frequencies well below its upper limit, the adjustment made no difference.
A small screwdriver was included for the adjustment, but I ended up using it a lot with the oscilloscope itself, rather than the probe. Note, though, that the screwdriver comes with the probe, not the oscilloscope. By the way, the probe is very inexpensive, in my opinion; they cost much more in local stores.



And here's the subject of this review.
It looks like a typical school electronics club project: a simple case, though it does have laser engraving instead of ordinary stickers. That said, it's not really relevant to the review.


The top of the case has a 2.4-inch color display with a resolution of 320x240. My DSO203 has a larger display, both physically and in resolution (400x240), though not by much.
The control buttons are on the right, and operation is extremely simple: there are no menus, settings, or anything like that. Just five buttons -
1, 2. Input voltage, from 0.01 to 5 V per division. 9 steps.
3, 4. Sweep speed, from 1.5 µs to 6 ms per division, 12 steps.
5. Hold button, which simply freezes the readings on the display. As it turned out, it's the button I use most in some situations.


The top edge of the case has a BNC input connector, as well as the power switch and power jack.
Incidentally, the device draws only about 150 mA, so it can be powered independently. However, since it's quite sensitive, a pair of lithium cells and a low-dropout linear voltage regulator are recommended. Search online using the prefix “Low Drop.”


There’s a hole on the bottom for accessing the trimmer that sets the zero level.


This unit is very easy to take apart. First, remove the four self-tapping screws on the bottom.


Then remove the four screws on top and take out the circuit board. There’s a hole drilled in the case for the connector, so the board has to come out through that opening.


Inside, you can see the oscilloscope board and a display that will look familiar to many electronics hobbyists. If I’m not mistaken, the same display is used in the DSO138.


The display is held in place by its connector, pressed down from above by the case, with two plastic posts melted onto the board underneath.


The circuit board is quite well made. Each component is marked not only with its reference designator but also its value, which is extremely rare. A repair technician’s dream :) I once made a video about how to identify the value of a burnt resistor; you wouldn’t need that here.


The power supply and input op-amp section. The soldering is pretty good, but it looks like some components were replaced after assembly; you can see flux residue.


The input circuitry and signal dividers. Unfortunately, the oscilloscope only works with AC, though that’s more than enough for most tasks.
There’s a 330 nF, 250 V capacitor at the input.


The input divider. The board has reed relays, and the divider provides 9 input voltage ranges. The first three relays are in the first op-amp’s circuit, then another pair is in the second op-amp’s circuit, giving 3 × 3 = 9 ranges.


The zero adjustment resistor. The oscilloscope arrived with the zero level “drifting.” I adjusted it, but experience has shown that the zero still likes to “wander” from time to time, so I need a small screwdriver fairly often.


Oscilloscope components:
1. Input dual op-amp LM6172, with a maximum frequency of 100 MHz.
2. ADC: ADS830E, with a maximum frequency of 60 MHz.
3. Asynchronous FIFO buffer with an access time of no more than 12 ns.
4. Atmega16A microcontroller; the 20 MHz crystal resonator is on the left.
5. A simple logic IC.
6. 7660 voltage converter, which generates a negative 5 V rail.
There’s also a 3.3 V linear voltage regulator on the board, visible above in the photo.


The soldering on the underside is reasonably good, but there’s a lot of flux residue.


Besides the component values marked on the board, there’s also a schematic. It’s for a version with a different power supply, though. Here, the power comes not from a 7660 but from a mains power supply with split rails.
Unfortunately, the schematic quality is a bit disappointing, but it is what it is.
You can see the input attenuator, ADC, buffer, and microcontroller with the display. The circuit is as simple as can be, but quite decent for what is essentially a toy.


Let’s take a closer look at what the oscilloscope is built around.
After the first attenuator, the signal goes straight to an amplifier.
It uses a pretty good op-amp with a bandwidth of up to 100 MHz, which is more than enough for the claimed 4 MHz.


Next is a decent 8-bit ADC made by Burr-Brown, with a maximum frequency of 60 MHz—again, plenty of headroom.
Interestingly, the DS203 I use has a dual ADC, but its sampling rate is only 40 MSa/s.


The FIFO buffer, as I understand it, has a maximum operating frequency of around 80 MHz. It uses an IDT7205. This series appears to be available in a military-grade version.


The Atmega16A handles displaying the signal and graticule, as well as measuring frequency.


The first thing I decided to do was check the noise level. The input wasn't shorted; if you short it, the screen just shows a straight line.
On the left, the oscilloscope is just sitting on the table; on the right, I've put my hand on the case near the input attenuator.


Don't worry: in reality, the oscilloscope screen looks much better, with a clear, high-contrast image.


The oscilloscope can't take screenshots, so I had to resort to the “old-fashioned way.”


To start with, I used the built-in signal generator in my trusty DS203.
A 20 kHz sawtooth and triangle wave, respectively. The unit under review and my oscilloscope both display them quite well.


A 20 kHz sine wave and square wave.
The sine wave matches, but the square wave's rising edge is heavily rounded.


I assume the generator above was operating in DDS mode, so I raised the frequency above 20 kHz, since the square-wave generator definitely operates in this mode.
At 200 and 500 kHz, I'd say the results are pretty good, except that one waveform is distorted in one way and the other in a different way. It looks as if the image is mirrored. In both cases, I used the cable from the DS203, connecting it to each oscilloscope in turn.


Then I happened to notice an interesting quirk. It may be a software bug, or it may be intentional, but the oscilloscope lets you reduce the sweep time well below the specified 1.5 µs per division.
I started cycling through the sweep modes and, by increasing the time, was able to stretch out the signal.
Of course, the frequency counter started showing random values.


All right, now I'm curious. Let's feed it 1 MHz.


1.5 µs on the left; the “incorrect” stretched mode on the right.


Let's try 2 MHz.


Well, I thought, “that's it, it's dead.” The screen shows nonsense: in the first mode you can't make anything out, and in the second it looks almost like a triangle.


But I don't give up, so I'll feed it 4 MHz. Even on my oscilloscope, the waveform is barely recognizable as a square wave.


On the unit under review, though, the screen is a complete mess, but...
1. The original signal at the shortest sweep setting. The frequency counter works normally and displays the 4 MHz input, but the waveform is painful to look at.
2. Increase the sweep time, as I did above. Well, now it's a triangle.
3. Let's change the input attenuator from 1 V per division to 0.5 V. Oh, that's already noticeably better.
4. And now let's stretch the sweep even more. It even looks like a square wave :)

On the other hand, nobody actually promised performance above 4 MHz.


I ran the next test in this mode. By the way, at 6 MHz the built-in frequency counter starts showing nonsense. But as it turned out, the frequency displayed on the screen is still a multiple of the actual input frequency.
1. At 6 MHz, the screen shows 2 kHz, i.e. 3,000 times lower.
2. At 8 MHz, the screen shows 2.8 kHz, also about 3,000 times lower than 8 MHz.

But it does work. I got the impression that somehow only part of the actual signal is being displayed, heavily divided down—in other words, most of it has been “taken out,” leaving a waveform that looks reasonable.
Unfortunately, I don't have much equipment for testing at high frequencies.


To be fair, I initially tried running the tests with a different signal generator.


I wouldn't have included those results in the review if I hadn't noticed a few things along the way.
First, here's the 8 MHz signal as-is, and then stretched out as I did above.


But if you stretch it even further, it looks like this. Perhaps this will give someone something to think about.


And here's what the triangle and sawtooth waves from this generator look like on the oscilloscope under review and on my main one.
Frequency: 65 kHz.


Since I'm testing it, I'll also check how this oscilloscope handles more realistic signals. For example, here's a waveform from one of my power supply reviews. though here I used a capacitor in parallel with the probe, as I did in my recent power supply reviews.


The same power supply, roughly the same load, but with different signal display settings.
Looks similar? I think so.


The waveform I showed above may not be very clear to some, so I picked one of the power supplies with more familiar-looking ripple.
The same power supply: 50% load on the left, 100% on the right. The waveforms match in both cases, and on the oscilloscope under review you can also zoom in by a factor of 2 or 4.
However, my oscilloscope is already set to its minimum of 50 mV per division, while the one under review can be made five times more sensitive, down to 10 mV per division. One small "quirk" did turn up, though: one oscilloscope showed a greater ripple amplitude than the other. By the way, the oscilloscope under review displays the peak-to-peak value quite accurately.


Group photo: DSO138, the oscilloscope under review, and DS203.


To sum up, this oscilloscope pleasantly surprised me, first and foremost with its pretty decent components and simple circuit design. In terms of features, it will of course lose even to the DSO138, let alone the DS203. But in terms of specifications, it is a cut above the DSO138, and I would say it is not much worse than mine in some respects. Keep in mind that the DS203 uses an ADC with a maximum sampling rate of 40 MHz, while the one under review has a 60 MHz ADC.
The input attenuator uses no fancy switching circuitry, just the simplest relays, but this solution works.
One downside is that the input is AC-only, unlike the AC/DC input on the DSO138 and DS203.
On the plus side, the controls are extremely simple. Unfortunately, this still comes with a catch: the built-in trigger, which keeps the signal on screen, can be tricky to use. This is what I was talking about above when I mentioned the Hold button. In some situations, the oscilloscope cannot keep the signal steady on screen, and it starts to "jump around". Pressing Hold usually fixes it; you just have to get used to this.
The strangest thing is the square wave at 20 kHz.

Otherwise, this is a very interesting option for beginners in electronics: it is easy to use and practical, for example when working with power supplies.
Also, this oscilloscope is sold in an enclosure (both an advantage and a drawback), and runs on 5 V. I tried powering it from a power bank, and it works great.

I bought it through the forwarding service yoybuy.com; the kit cost about 22 dollars. Shipping costs depend on the country; the review lists the weight of the individual parts.
If you sign up with the forwarding service using this link, you should get a coupon for 10 off 50 dollars. It is a referral link, but I don't get anything from it.


That's all from me. As always, I look forward to your questions, and I hope you found the review useful.