A kit for building a simple DDS signal generator

Continuing the topic of electronic kits, this time I want to tell you about a device that can add to a beginner radio enthusiast’s collection of test equipment.
Strictly speaking, you wouldn’t call it a measuring instrument, but it definitely helps with measurements.

Radio enthusiasts—and not just them—often need to test various electronic devices. This can be during debugging or repair.
Sometimes you need to trace a signal through different circuits in a device, but the device itself doesn’t always let you do that without an external signal source.
For example, when setting up or testing a multistage audio power amplifier.

First, a little explanation of what this review is about.
I want to tell you about a kit for building a signal generator.

There are all kinds of generators. For example, these are generators too :)


But we’ll be building a signal generator. I’ve been using an old analog generator for many years. It’s very good at generating sine waves, with a frequency range of 10–100000 Hz, but it’s bulky and can’t generate other waveforms.
This time, we’ll be building a DDS signal generator.
DDS stands for Direct Digital Synthesizer.
This device can generate signals of arbitrary waveforms and frequencies, using an internal oscillator with a single frequency as its reference.
The advantage of this type of generator is that it offers a wide tuning range with very fine steps, and can also generate complex waveforms when needed.

As always, let’s start with a little about the packaging.
In addition to the standard packaging, the kit came in a thick white envelope.
All the components were in an antistatic resealable bag—a pretty useful thing to have around in a radio enthusiast’s workshop :)


The components were loose inside the package, and looked roughly like this when unpacked.


The display was wrapped in bubble wrap. About a year ago, I already reviewed this display and put it to use, so I won’t go into it here. I’ll just say it arrived without incident.
The kit also included two BNC connectors, but they’re a simpler design than the ones in my oscilloscope review.


The ICs and their sockets were packed separately on a small piece of polyethylene foam.
The device uses an ATmega16 microcontroller from Atmel.
People sometimes mix up the terms and call a microcontroller a processor. They’re actually different things.
A processor is essentially just a computing unit. A microcontroller, on the other hand, includes RAM and ROM in addition to the processor, and may also have various peripherals such as DACs, ADCs, PWM controllers, comparators, and so on.

The second IC is an LM358 dual operational amplifier. It’s a very common, mass-produced op amp.


First, let’s lay out the whole kit and see what we’ve got.
Printed circuit board
1602 display
Two BNC connectors
Two potentiometers and one trimmer
Crystal resonator
Resistors and capacitors
ICs
Six buttons
Various connectors and mounting hardware


The PCB is double-sided, with component markings printed on the top side.
Since no schematic is included, the board is marked with component values rather than reference designators. In other words, you can assemble everything without a schematic.


The plating is well done, and I have no complaints. The pads have an excellent finish and are easy to solder.


The vias between the two sides are doubled.
I don't know why they did it this way instead of the usual way, but it only makes the board more reliable.


I started drawing the schematic from the PCB layout. But while I was working on it, I thought that the kit must have been based on an existing design.
And sure enough, an internet search led me to the original version of this device.
The link has the schematic, PCB layout, and source files with the firmware.
I still decided to finish drawing the schematic exactly as implemented, and I can say that it matches the original version 100%. The kit's designers simply made their own PCB layout. This means that if alternative firmware exists for this device, it should work here too.
One note about the circuit design: the HS output is taken directly from a processor pin, with no protection at all, so there's a chance of accidentally burning out that output :(


Since I'm going into detail, I should describe the functional blocks in this circuit and explain some of them more fully.
I made a color version of the schematic, highlighting the main blocks.
I'm not very good at naming colors, so I'll describe them as best I can :)
Purple on the left — the power-on reset and the manual reset button.
When power is applied, capacitor C1 is discharged, which holds the processor's Reset pin low. As the capacitor charges through resistor R14, the voltage at the Reset input rises and the processor starts running.
Green — mode selection buttons
Light purple? — 1602 display, backlight current-limiting resistor, and contrast adjustment trimmer.
Red — signal amplifier and offset adjustment relative to zero (I'll show what it does near the end of the review)
Blue — DAC. Digital-to-analog converter. The DAC uses an R2R ladder, one of the simplest DAC designs. This is an 8-bit DAC, since it uses all the pins of one microcontroller port. By changing the code on the processor pins, you can get 256 voltage levels (8 bits). The DAC consists of resistors with two values, one twice the other; that's where the name comes from, combining R and 2R.
The advantages of this approach are high speed and very low cost. It's best to use precision resistors. A friend and I used this principle for an ADC, but we had few precision resistor values to choose from, so we used a slightly different approach: all the resistors had the same value, and where 2R was needed, we used two resistors in series.
This digital-to-analog conversion method was used in one of the first “sound cards” — Covox. It also had an R2R ladder connected to the LPT port.
As I mentioned above, the DAC in this kit has 8-bit resolution, or 256 signal levels, which is more than enough for a simple device.


The author's page also had a block diagram of the device, in addition to the schematic, firmware, and so on.
It makes the connections between the blocks easier to understand.


That wraps up the main part of the description. The extended version follows later in the article, but for now let's move on to the assembly itself.
As in the previous examples, I decided to start with the resistors.
This kit contains a lot of resistors, but only a few different values.
Most of the resistors are either 20 kΩ or 10 kΩ, and almost all of them are used in the R2R ladder.
To make assembly a little easier, you don't even need to measure their resistance: there are 9 20 kΩ resistors and 8 10 kΩ resistors :)


This time I used a slightly different assembly technique. I like it less than the previous ones, but it has its place too. In some cases, this technique speeds up assembly, especially when there are lots of identical components.
Here, the resistor leads are formed as before. Then all the resistors of one value are installed on the board first, followed by the other value, forming two rows of components.


On the back, the leads are bent slightly—not too much. The main thing is to keep the components from falling out. Then place the board on the table with the leads facing up.


Next, take solder in one hand and the soldering iron in the other, and solder all the populated pads.
Don't get carried away and install too many components at once. If you fill the whole board in one go, you can get lost in this “forest” :)


Finally, trim the protruding component leads flush with the solder joints. You can grab several leads at once with the side cutters (4–5–6 at a time).
Personally, I'm not a big fan of this assembly method. I'm showing it simply to demonstrate different ways of putting the kit together.
The drawbacks of this method:
The trimmed ends are sharp and stick out.
If the components aren't in a row, the leads can easily turn into a tangled mess, which just slows things down.

The advantages:
Fast assembly of identical components arranged in one or two rows.
Since the leads aren't bent much, it's easier to remove a component.

You often see this assembly method in cheap computer power supplies. The leads aren't clipped there, though; they're cut with something like a cutting disc.


After installing most of the resistors, we'll have a few of different values left.
The pair is easy to identify: they're two 100 kΩ resistors.
The last three resistors are:
brown — red — black — red — brown — 12 kΩ
red — red — black — black — brown — 220 Ω.
brown — black — black — black — brown — 100 Ω.


Solder in the remaining resistors. The board should now look something like this.


Color-coded resistors are a good thing, but sometimes it can be confusing which end to start reading the bands from.
With resistors that have four bands, there’s usually no problem, since the last band is most often silver or gold. But five-band resistors can be trickier.
The last band may be the same color as the bands indicating the resistance value.

To make the color code easier to read, the last band should be spaced farther from the others—but that’s the ideal case. In real life, things don’t always work out as intended, and the bands may all be evenly spaced in a row.
Unfortunately, in that case, your options are a multimeter or simple logic (when assembling a device from a kit): rule out all the known values, and the remaining options should tell you what value you’re looking at.
Here are a couple of examples of resistor color codes from this kit.
1. Two adjacent resistors have “mirror-image” color codes, so it doesn’t matter which end you read from :)
2. These are 100 kΩ resistors. You can see that the last band is a little farther from the others (in both photos, the value reads from left to right).


All right, that’s enough about resistors and their tricky color codes. Let’s move on to simpler things.
There are only four capacitors in this kit, and they come in pairs, so there are just two values, with two of each.
The kit also includes a 16 MHz crystal resonator.


I covered the capacitors and crystal resonator in my previous review, so I’ll just show where they go.
Apparently, the original plan was to use the same type of capacitor throughout, but the 22 pF capacitors were replaced with small disc capacitors. The PCB footprint has 5 mm lead spacing, while the small disc capacitors have only 2.5 mm spacing, so their leads will need to be bent outward a little. Bend them close to the body (fortunately, the leads are soft), since the processor sits above them and they need to sit as low as possible above the board.


The kit includes a couple of IC sockets and several connectors.
We’ll need them for the next step. We’ll also take a long female connector and a four-pin male connector (not shown in the photo).


The IC sockets supplied are standard ones, though compared with the sockets from the Soviet era, they’re a luxury.
In fact, experience shows that sockets like these often outlast the device itself.
The sockets have a notch—a small cutout on one of the short sides. It doesn’t matter which way you install the socket itself; the notch just makes it easier to orient the IC when inserting it.


Install the sockets with their notches oriented the same way as the markings on the PCB.


Once the sockets are installed, the board starts to take shape.


The device is controlled using six buttons and two potentiometers.
The original device had five buttons; the kit designer added a sixth, which serves as a reset button. To be honest, I still don't quite understand what it's for in practice, since I never needed it during all my testing.


I mentioned above that the kit included two potentiometers. It also came with a trimmer potentiometer. Here's a little about these components.
Potentiometers are used to adjust resistance on the fly. In addition to their resistance value, they are also marked with a taper code.
The taper describes how the resistance changes as you turn the knob.
There are three main tapers:
A (B in the international system) — linear: resistance changes linearly with the angle of rotation. These potentiometers are convenient, for example, in power supply voltage adjustment circuits.
B (C in the international system) — logarithmic: resistance changes rapidly at first, then more gradually as you approach the midpoint.
C (A in the international system) — reverse logarithmic: resistance changes gradually at first, then more rapidly as you approach the midpoint. These potentiometers are usually used for volume controls.
There is also a W taper, available only in imported versions. It has an S-shaped response, combining logarithmic and reverse-logarithmic tapers. To be honest, I don't know where these are used.
If you're interested, you can read more here.
By the way, I've come across imported potentiometers whose taper letter matched ours. For example, here is a modern imported potentiometer with a linear taper and the letter A in its part number. If in doubt, it's best to look for more information on the manufacturer's website.
The kit came with two potentiometers, but only one of them was marked :(

There was also one trimmer potentiometer in the kit. Essentially, it's the same as a regular potentiometer, but it's not intended for frequent adjustment; it's more of a set-it-and-forget-it component.
These potentiometers usually have a screwdriver slot instead of a knob, and only a linear resistance taper (at least, I haven't come across any others).


Solder in the resistors and buttons, then move on to the BNC connectors.
If you plan to use the device in an enclosure, it might be worth buying buttons with longer shafts rather than extending the ones included in the kit. That would be more convenient.
I'd also mount the potentiometers on wires, since they're very close together and would be awkward to use as they are.


The BNC connectors are simpler than the ones in my oscilloscope review, but I like them better.
The main thing is that they're easier to solder, which matters a lot for beginners.
I did notice one issue, though: the designers placed the connectors so close together on the board that it's physically impossible to tighten both nuts side by side; one will always sit on top of the other.
In practice, you rarely need both connectors at once, but it would have been much better if the designers had spaced them just a couple of millimeters farther apart.


That's the main board soldered up. Now we can install the op-amp and microcontroller in their sockets.


Before installing the IC, I usually bend the leads slightly so they sit closer to the center of the chip. It's very simple: hold the IC by its short sides with both hands and press the side with the leads vertically against a flat surface, such as a table. You don't need to bend the leads very much; it's mostly a matter of habit, but it makes the IC much easier to fit into the socket later.
When installing it, make sure none of the leads have accidentally bent inward, underneath the IC, as they may break when you bend them back.


Install the ICs according to the notch on the socket, which is itself aligned with the marking on the circuit board.


That completes the assembly of the main board.
After all these steps, the board should look something like this.


With the board done, let's move on to the display.
The kit includes the pin header, which needs to be soldered on.
After installing the header, I first solder one of the end pins. It doesn't matter if the joint looks neat; the important thing is to get the header sitting firmly and perpendicular to the board. If needed, reheat the joint and straighten the header.
Once the header is aligned, solder the remaining pins.


That's it, we can clean the board now. This time I decided to do it before testing, although I usually recommend cleaning after the first power-up, since sometimes there's still something left to solder.
But in my experience, kits are much simpler, and there's rarely any soldering left to do after assembly.


There are various ways and products you can use to clean a board. Some people use alcohol, others a mixture of alcohol and gasoline. I use acetone, at least while I can still buy it.
After cleaning the board, I remembered the tip about using a brush from the previous review, since I use cotton wool. Never mind, I'll have to try it next time.


I've gotten into the habit of coating the board with protective lacquer after cleaning it, usually on the underside, since getting lacquer on the connectors is a no-go.
I use Plastik 70 lacquer.
This lacquer is very “light,” meaning it can be removed with acetone and soldered through with an iron if needed. There's also a good urethane lacquer, but it's noticeably harder to work with: it's tougher, and soldering through it with an iron is much more difficult. This type of lacquer is used in harsh operating conditions, and when you're sure the board won't need any more soldering for quite a while, at least.


After the lacquer is applied, the board looks glossier and feels nicer to the touch. It gives you a sense that the job is finished :)
It's a shame the photo doesn't capture the overall effect.
I used to find comments like “this tape recorder/TV/radio has been repaired—you can see the soldering marks” pretty funny :)
With good, proper soldering, there are no visible signs of repair. Only a specialist can tell whether a device has been repaired.


It’s time to install the display. The kit includes four M3 screws and two mounting standoffs.
The display is secured only on the side opposite the connector, since on the connector side it’s held in place by the connector itself.


Install the standoffs on the main board, then mount the display, and finally secure the whole assembly with the two remaining screws.
I liked that even the holes lined up with impressive accuracy, without any adjustment—just insert it and screw it in :).


All right, let’s give it a try.
I apply 5 V to the corresponding connector pins and…
Nothing happens, except that the backlight turns on.
Don’t panic and start searching forums for a solution—it’s all fine, that’s exactly how it should work.
Remember, there’s a trimmer on the board, and it’s there for a reason :)
Use this trimmer to adjust the display contrast. Since it was initially set to the middle position, it’s perfectly understandable that we couldn’t see anything.
Grab a screwdriver and turn the trimmer until the image on the screen looks right.
If you turn it too far, the contrast will be too high: all the character positions will be visible at once, while the active segments will be barely visible. In that case, just turn the trimmer the other way until the inactive elements almost disappear.
You can adjust it so the inactive elements aren’t visible at all, but I usually leave them just barely visible.


Next, I should move on to testing—but not so fast.
When I got the board, the first thing I noticed was that, besides 5 V, it also needs +12 V and -12 V—in other words, three supply voltages in total. It immediately reminded me of the RK86, which needed +5, +12, and -5 V, and had to be powered up in a specific sequence.

I had no trouble with 5 V or +12 V, but -12 V turned out to be a bit of a problem. I had to put together a small temporary power supply.
The process was the usual: rummaging through my parts stash to find something to build it from, laying out the board, and making it.


Since I only had a transformer with a single winding and didn’t want to build a switching power supply, I decided to make a power supply using a voltage doubler circuit.
To be honest, this is far from the best option, since this circuit has quite a high ripple level, and I had barely enough voltage headroom for the regulators to filter it properly.
The circuit at the top is the better way to do it; the one at the bottom is what I built.
The difference is an extra transformer winding and two diodes.


Transformer I used one with almost no headroom, too. But it’s sufficient as long as the mains voltage is normal.
I’d recommend using a transformer rated for at least 2 VA, or preferably 3–4 VA, with two 15 V windings.
By the way, the board’s power consumption is low: at 5 V, including the backlight, it draws just 35–38 mA. The current draw at 12 V is even lower, but depends on the load.


In the end, I got a small board, a little larger than a matchbox, mostly in height.


At first glance, the board layout may seem a little odd. I could have rotated the transformer 180 degrees for a neater layout, and that’s what I did at first.
But with that arrangement, the mains-voltage traces ended up dangerously close to the instrument’s main board, so I decided to change the layout a little. I wouldn’t say it’s perfect now, but at least it’s a bit safer.
The fuse holder could be removed, since it isn’t really needed with the transformer used here. That would make the layout even better.


Here’s the complete kit. To connect the power supply to the instrument board, I soldered together a small rigid 4×4-pin connector.


The power supply board connects to the main board via the connector, and now we can move on to how the instrument works and to testing. Assembly is complete at this point.
Of course, I could have put it all in an enclosure, but for me this instrument is more of a handy auxiliary tool. I’m already looking at more advanced DDS generators, but they’re not always affordable for beginners, so I decided to leave it as is.


Before testing, I’ll describe the controls and features of the device.
The board has five control buttons and a reset button.
The reset button is self-explanatory, so I’ll describe the others in more detail.
There is a slight “bounce” when pressing the left/right buttons. The software debounce may be set too short. It mainly shows up when selecting the output frequency in HS mode and the frequency tuning step; no problems have been noticed in the other modes.
The up and down buttons switch between the instrument’s operating modes.
1. Sine
2. Square
3. Sawtooth
4. Reverse sawtooth


1. Triangle
2. High-frequency output (separate HS connector; the other waveforms are available on the DDS output)
3. Noise-like (generated by randomly cycling through combinations at the DAC output)
4. ECG signal emulation (an example of how any waveform can be generated)


1–2. The DDS output frequency can be set from 1 to 65,535 Hz in 1 Hz steps.
3–4. There is also a separate menu item for selecting the tuning step; it defaults to 100 Hz.
You can change the frequency and operating mode only when generation is off. Changes are made with the left/right buttons.
Press START to begin generating.


There are also two potentiometers on the board.
One adjusts the signal amplitude, and the other adjusts the offset.
I tried to show what this looks like in the oscilloscope traces.
The top two show changes in the output signal level; the bottom two show offset adjustment.


Next are the test results.
All signals except the noise-like and HF signals were tested at four frequencies:
1. 1,000 Hz
2. 5,000 Hz
3. 10,000 Hz
4. 20,000 Hz.
There was significant roll-off at higher frequencies, so there’s not much point in showing those oscilloscope traces.
First, the sine wave.


Sawtooth


Reverse sawtooth


Triangle


Square wave from the DDS output


ECG waveform


Square wave from the RF output
Only four frequencies are available here, so those are the ones I tested:
1. 1 MHz
2. 2 MHz
3. 4 MHz
4. 8 MHz


Noise-like waveform in two oscilloscope sweep modes, to give a better idea of what it looks like.


Testing showed that the signals become quite distorted starting at around 10 kHz. At first I blamed the simplified DAC and the basic synthesis implementation, but I wanted to investigate more thoroughly.
To check, I connected the oscilloscope directly to the DAC output and set the synthesizer to its maximum frequency, 65535 Hz.
The waveform looks better here, especially considering that the generator was running at its maximum frequency. I suspect the simple amplifier circuit is to blame, since the signal looks noticeably “cleaner” before the op-amp.


And here’s a group photo of a beginner radio enthusiast’s little “workbench” :)


Summary.
Pros
Well-made circuit board.
All components were included.
Assembly was straightforward.
A wide range of functions.

Cons
The BNC connectors are too close together.
No protection on the HS output.

My opinion: You could say that the device’s specifications are pretty poor, but keep in mind that this is an entry-level DDS generator, so it wouldn’t be fair to expect much more from it. I was pleased with the quality of the circuit board; assembly was a pleasure, and there wasn’t a single spot that needed “reworking.” Since the device is based on a fairly well-known design, there’s hope that alternative firmware will become available and expand its functionality. Taking all the pros and cons into account, I can certainly recommend this kit as a starter project for beginner radio enthusiasts.

Phew, I think that’s everything. If I’ve messed anything up, let me know and I’ll fix it or add more :)

The product was provided by the store for review.