DarTZeel amplifier clone for those who like a little warmth
$42.5
And it also runs hot in the literal sense, but more on that in the review :)
First, a few words about Dartzeel itself. It’s a well-known company that mainly makes power amplifiers positioned as high-end equipment.
They do look pretty good, though somewhat unusual.

The construction is also quite well executed. For example, this unit has two powerful transformers, large-capacity capacitors connected not even by a busbar but by special, heavy-gauge links, and large heatsinks.

No, this isn’t the clone yet. As far as I can tell, it’s one of the early versions, although there aren’t as many differences from the version shown above as you might think. The key selling point is said to be years of fine-tuning the circuit design and component values to achieve the sound the designer wanted. But I’ll say right away that, in my opinion, this is mostly a way to justify the high price of these units; otherwise, you can’t break into this market segment.
Yes, yes, cable risers again, and warming up the air in the listening room with a special pink-noise disc. If the first version was aimed at the ears, the later ones seem to put more emphasis on visual appeal.

But the Chinese wouldn’t be the Chinese if they didn’t start cloning these amplifiers. An inexperienced person might not even be able to tell them apart from the outside: the same colors, the same two large red power indicators, and even the internals look somewhat similar.
Of course, they do a decent job, but they’re still clones, however you look at it. That doesn’t stop people from writing such beautiful articles about them, though. I recently read one, and it does indeed feature “the veiling of microdynamics with a certain velvety quality.”

The first difference you can spot from the outside is that the clone has ordinary “gold-plated” speaker terminals, while the original has proper ones. The original also has three inputs—Coaxial, Balanced, and Zeel input with a 50 Ω impedance—while the clone only has a standard coaxial input.

Since this is an amplifier review, that’s what we’ll focus on. First, here’s what the originals look like.
The amplifier is at the top, and the power supply circuit is below. Judging by the photos, the amplifier’s power supply is regulated, though I could be wrong.

A closer look at the amplifier. Yes, it’s a little thing. Essentially, this is what produces the sound. The transformers, regulators, and power-supply filter capacitors all play an important role, of course, but I still think the amplifier itself matters most.
The amplifiers shown above use a “dual mono” design: two independent channels in one enclosure, sharing only the power switch.

And here’s the Chinese clone of the amplifier board shown above. Look closely and you’ll see that they’re practically identical; the main differences are the types of components used. In effect, they took the original and simply copied it, replacing the parts with cheaper ones.

But apparently the copyists got carried away and started making a higher-power version with four output transistors instead of two, connected in pairs. Even here, you can see that it’s essentially the same original board, with just a few minor modifications.

After that, things became more of an improvisation. The updated board has obvious differences in appearance and construction. For example, the transistors are arranged in a single row so the board can be mounted perpendicular to the heatsink; previously, it could only be mounted parallel to it. They also added a speaker protection circuit to the board, and mounted two of the transistors on their own small heatsinks. The circuit design, however, remained exactly the same as in the earlier board versions.

Today I’ll be looking at an expanded version of the updated board, with a new PCB, pairs of output transistors, protection circuitry, and additional power-supply filter capacitors.
I ordered the kit from Taobao; the link is above. It’s sold as a set, and the listed price is for a pair.
Each board was individually wrapped in soft foam, and according to the forwarding agent, the whole kit weighs 561 grams.

Each board is also wrapped in heat-shrink film. And yes, this is a finished product, not a kit, although I think a kit would have been more interesting.

Here’s the complete set: two amplifier boards and two bags of hardware.

At first I thought, “Wow, they’ve included plenty of spares,” but then I found that the second bag had fewer parts. Still, there are a few extras overall: five insulators per board instead of four, and 19 screws instead of 18.

The boards are laid out as mirror images. Fortunately, this is a bit easier with transistor amplifiers than with designs that use ICs.
Specifications from the product page:
Board dimensions: 145 × 92 mm
Power supply: ±35–55 V DC for the power stage and 12–15 V AC for the auxiliary supply.
Output power: 2 × 200 W
Load impedance: 4–8 Ω.

They can be installed as shown above, in a row, or side by side if you’re using an enclosure with heatsinks on the sides.

This approach isn’t new and has long been used when designing systems like these. Personally, though, I think it’s unnecessary when building separate monoblocks, like the one shown below, because the more complicated board layout will perform worse than two identical boards.

The boards are marked left and right, though of course that’s just a convention; there’s nothing stopping you from swapping them.

Right-channel board

And the left one.

Overall build quality is average; some components are crooked. If you assemble it yourself, you can do a neater job—which is why I like kits.

The connectors for the input and the speaker protection circuit’s power supply, along with a terminal block for the amplifier’s own power supply, are located at one end of the board.
Warning! Unlike the LM3886-based amplifier I showed earlier, this one runs on DC. There’s no bridge rectifier on the board, although it does have small filter capacitors, so be careful.

The speaker outputs are pads on the board. The idea is that the board will be mounted inside an enclosure and the wires from the terminals soldered directly to it—a perfectly sensible solution.

Each supply rail has a pair of 390 µF, 100 V capacitors. The board labels them as 470 µF. It’s not a huge difference, but it’s there.
The input has a 2.2 µF, 160 V capacitor made by Wima—or at least that’s what the listing says. Given the price of the kit, though, why not? There are only two of them in the whole set, after all.

The amplifier’s input stage. This is also where the only adjustment component is located; it’s used to set the DC offset at the amplifier output to zero.
The circuit design is a little unusual. As I understand it, the entire amplifier is a DC amplifier, and the transistors operate at fairly high quiescent currents. As a result, they’re already quite powerful starting from the third stage.

The fourth-stage transistors are mounted on small heatsinks; the TO-220 package alone can no longer dissipate enough power.

The pre-driver stage is already cooled by the main heatsink. All the transistors also need to be electrically isolated, so the kit includes eight thermal pads.
>The output stage uses pairs of 2SC5200 and 2SA1943 transistors. I can’t say whether they’re genuine. I wanted to compare their internals with samples bought from local stores, but couldn’t find any that were guaranteed to be genuine.
Each output transistor has a 0.22 Ω resistor in its circuit.

Well, it turns out there are ICs after all: the speaker protection circuit uses a C1237HA. On the left is a 12 V regulator that powers the IC and relay. The relay is the most ordinary 12 V type.

The boards are covered with an opaque black solder mask, so you can only make out the traces approximately. Still, the layouts aren't quite mirror images: one board has a clearly high-current trace on the underside that the other board doesn't have.

The components mounted on the heatsink are soldered along the full length of their leads, which is very convenient. The soldering is perfectly decent, with no obvious defects. You can even see that they tried to clean the boards, but there are fingerprints, so I'd still wash them again, if only to make them look nicer.


I couldn't find the circuit for the amplifier section of the exact version under review, but I did find a couple of very similar ones. I think any differences, if there are any, are limited to the types of transistors used.

Actually, no—there's at least one difference in component values: the zero-adjustment resistor. It's 50 kΩ in the schematics, but 10 kΩ in the unit under review. I can say that this may actually be better.

A few words about the heatsinks I planned to use for these amplifiers. Last time I wrote that my test heatsink wasn't enough for the amplifier, and I also had the idea of using the case from a Romantika amplifier (I'll show later what I had in mind). So I ordered a pair of heatsinks that were a bit more substantial than the previous Chinese “little guy.”
I ordered the heatsinks here; they cost around $3–3.5 each.

I was intrigued by the somewhat unusual shape of the fins, which are supposed to improve cooling efficiency.
The heatsink measures 190 mm long and 60 mm wide. The inner fins are 28 mm high, the outer ones 30 mm, and the base is 5 mm thick.
They normally come in widths in 50 mm increments, but my case could accommodate a width of 60 mm, so I asked them to make it 60 mm wide.

To keep the review from getting too long, I'll put the rest of the description of the heatsinks under a spoiler.
Accordingly, I marked out the heatsinks, drilled them, and tapped the holes. Although the power components on the board look identical, there are actually slight differences, so I had to mark out each heatsink separately and then label them L and R.

I didn't use the original pads for several reasons:
1. I don't like them very much.
2. The TO-220 transistors are too close together, so the pads overlap.
3. Mica seems like a more suitable insulating material to me. Ceramic would probably be better, but this is what I had.
I had quite a lot of mica strips left over from one of my projects, so I decided to use a single piece for groups of components rather than one pad per component. Also, after marking and drilling the mica, I split it into thinner layers, which is easy enough to do. The pieces at the bottom right have already been split. Then it was just KPT-8 thermal paste, a screwdriver, and the original screws.

And here's the handsome beast I ended up with for testing.

For now, I can't power both channels at once, so I experimented with just one. The power came from a TPP270 transformer, with the output windings connected in series with a center tap. The primary windings were configured for 214 V (100+100+7+7). The power supply filter was 2×10,000 µF, and with no signal at the amplifier input, the voltage across each rail's capacitors was about 33 V.
Auxiliary power came from a TPP 224, but it has no effect on the sound; it powers the protection circuit.
The specified supply voltage range for this amplifier model is ±35–55 V, but it was originally designed for an output power of up to 200 W. The original version, with one pair of transistors, is rated at:
100 watts RMS @ 8 (Hi) and 2 (Lo) Ohms
160 watts RMS @ 4 (Hi) and 1 (Lo) Ohms
I didn't need that much power with my little 20–30 W speakers, so I figured a supply voltage of around 35–36 V would be plenty. For the test, I chose the closest available voltage.
Also, my speakers have an impedance of 4 (3.5) ohms, so 45–55 V would just be overkill and would only make the amplifier run hotter. However, another user has reported that similar amplifiers behave unstably with a reduced supply voltage of 25–27 V.

A few tests. First, why the amplifier really does run warm—no quotation marks needed.
This is a thermal image of the amplifier after about 20–30 minutes of operation with no load. The hottest components are the transistors mounted on separate heatsinks, at around 90–95 degrees. Then I played audio at a relatively low power level (compared with the amplifier's output power). The output transistors were at about 60 degrees, while the driver stage was close to 90. The heatsink was hot as an oven—you couldn't hold your hand on it—which shows that heat is being transferred effectively.

I completely forgot to mention that the amplifier needs its output offset adjusted to zero. Initially, the voltage was about -0.25 V. It keeps drifting as the amplifier warms up, so it's best to adjust it once the temperature has stabilized.
Yes, this is the very same warm-up that audiophiles confuse with cables “burning in” to improve their sound. In fact, warm-up is standard practice with measurement equipment, but here it's needed because of the circuit's shortcomings.
Even the original documentation states:
DC voltage output drift:
Version A: < ±590 mV max.
Version B : < ±10 mV max.
Version A: < ±590 mV max.
Version B : < ±10 mV max.
Yikes, the output voltage on version A can drift by almost 0.6 V....
What's more, it all depends on temperature, supply voltage, and so on. Here's an excerpt from the manual (thanks to Wooferman):
If your AC mains voltage is stable, and your listening room is at standard temperature, we recommend removing the compensation by putting the jumper to off. In doing so, you will fully benefit from the very philosophy of the NHB-108 model one, which is to refuse any compensation or global negative feedback which could impair the sound. Most users have reported to us that this setting has the best sound.
Google translation
If the AC mains voltage is stable and your listening room is at a standard temperature, we recommend disabling the compensation by setting the jumper to off. This lets you fully benefit from the philosophy behind the NHB-108 model one: rejecting any compensation or global negative feedback that could impair the sound. Most users have told us that this setting sounds best.
But I got curious about how much current the amplifier draws at idle, so I measured 412/357 mA on the unbalanced amplifier. If you disconnect the load or set the output to exactly zero, you can find out how much the amplifier itself draws: 382 mA per rail, almost 400. Remember, the tests were run at the minimum supply voltage, and even then the dissipation came to about 26 W per channel. Even if the current doesn't increase as the voltage rises, at the nominal 50–55 V we'll be looking at around 40–45 W per channel, or close to 100 W for the whole amplifier!
For now, I plan to reduce the idle current a little by amputating one pair of output transistors, since I don't need that much power. Also, the author of the original amplifier once answered a question about paralleled output transistors, saying that because their characteristics differ, they turn on at different times and smear the sound.
In theory, he's right in a way, but I think:
1. It's very difficult to get this effect in a linear circuit.
2. It's even harder to hear the result of this effect.

It's clear that describing an amplifier's sound in words is very difficult, so next up are some basic tests.
First, a 1 kHz sine wave. The supply voltage at idle was about 32–33 V.
1. 3.3 Ω load, maximum swing before clipping, output voltage 12 V, input (DC) voltage 23 V, calculated output power — 43 W
2. 3.3 Ω load, onset of clipping, output voltage still 12 V, input (DC) voltage 23 V, calculated output power — 43 W. The upper half-cycle clips first; as the input voltage is increased further, both half-cycles clip.
3. 5 Ω load, maximum swing before clipping, output voltage 15 V, input (DC) voltage 25 V, calculated output power — 45 W
4. 10 Ω load, maximum swing before clipping, output voltage 17.5 V, input (DC) voltage 28 V, calculated output power — 30 W
You can see the input voltage sag under load. The transformer used was rated at about 60 W, so if you want decent output power, you'll have to shell out for a transformer rated at no less than 150–200 W per channel. And you'll need heatsinks, roughly like the ones shown in the first photos.

Now for the square wave, the test the 3886-based amplifier failed. Here, everything is excellent, partly thanks to the design of the amplifier stages themselves, and partly because there are no additional RLC networks at the amplifier output.
By the way, there are reports about this amplifier and its output saying that the popular darTZeel NHB-108 version (the original) had three revisions, and the first two were prone to failure when powered up without a load....
8.1.2. Do not power ON the darTZeel NHB-108 model one when no loudspeaker is hooked to the Speaker Terminals (6). In such a case the internal fuses could be blown by the Crowbar circuit, especially if the DC compensation offset has been disabled.
Google translation
8.1.2. Do not power on the darTZeel NHB-108 model if no loudspeaker is connected to the speaker terminals (6). In this case, the internal fuses may be blown by the Crowbar circuit, especially if the DC compensation offset has been disabled.
I tested this mode here; everything works, although I only checked it by chance, since I had no idea about this bizarre glitch.
1. 3.3 Ω load
2. 3.3 Ω load just before the protection kicked in.
3. 5 Ω load
4. 10 Ω

Now for RMAA.
Last time I wrote that it would be better to use my Creative Live sound card for the tests instead of the built-in one. I have one of those too, and remembering that, I finally dragged an old computer out of the balcony, where it had been sitting for a couple of years, and fired it up. Fortunately, everything was already installed, so all I had to do was install RMAA itself.
It has this sound card installed.

It served me for at least 15 years. I originally bought it at Nix, then it was shipped over here and resold by another company.

But the tests left me completely stunned.
I disabled all the EAX effects, set the tone controls and balance to zero, and even tried ASIO. But after enabling it, there were no input/output configuration options, and RMAA crashed when I tried to launch it.
First, the numbers.

But look at the frequency response—how is this even possible?

After spending probably half a day on various experiments, I decided to run the tests using the computer I use now, and hauled the old one back to the balcony.
But I ran into another problem. I don't know how, but a direct input-to-output connection produced this result.
Apart from the distortion at the end, the graph starts with a severe low-frequency roll-off.

Apparently, during my earlier experiments I slightly fried the sound card's input. I had to bring the old computer back in from the balcony, install drivers for its built-in sound card, and test with that. Attempts to use anything from the mini PCs came to nothing: they only have one jack, and trying to use an HDMI-to-VGA converter as a source was a dead end too.
Interestingly, the wavy fluctuations in the high-frequency part of the response are about the same as on my sound card, which is also made by Realtek, but the low-frequency range is noticeably better.

RMAA results obtained with the Realtek card and the amplifier under review. By the way, can RMAA perform relative measurements to compensate for the limitations of sound cards?

The frequency response graph matches the direct input-to-output connection, but because it's shifted slightly, it may look flatter in the low-frequency range here. It isn't; the two are identical.

Before the conclusions, a few words about what I was planning to build and why.
At one point I considered building something along the lines of the good old Odyssey 010, but the high heat makes that a nonstarter. I don't want to install a fan.

The second option is to arrange the amplifiers as shown in the photo, leaving a little space between them. I also have the idea of building a full dual-mono version, i.e. the channels in the amplifier would be completely separate, with their own transformers for the main and auxiliary power supplies, their own capacitor banks (lots of 470 µF, 40 V capacitors—I have plenty), and the whole design would be fully symmetrical. But I'd also like to put a conventional “analog” volume control in the center of the case, with just one button—for power. The amplifier grounds would be joined at the source. Ideally, I'd separate the DAC too, but that's more complicated, though still doable.
That way, I want to solve the problem of finding the right point to connect the channel grounds and, in turn, eliminate hum and interference caused by connecting them at the wrong point.
But life has thrown a wrench in the works for now. If removing one pair of transistors reduces the power dissipation, I think I can keep going in this direction. For now, though, it's running too hot.

Video version of the review
And now for the conclusions.
The amplifier works, and it sounds good. I can't say anything bad about it. Of course, it's no High End, but it sounds pretty good.
But it does have some drawbacks:
1. It gets very hot
2. The output DC offset drifts
3. The filter capacitors need to have a large capacitance.
I'll say a little more about that last point. In my review of the 3886 amplifier, I wrote that it had a little noise. I didn't hear any noise here at all. But there is a slight 100 Hz hum. If you use only the capacitors supplied with the boards (I tested that too), the hum is noticeable. With 10000 µF capacitors, it almost disappears, but it's still there: if you put your ear right up to the woofer, you can hear it.
These are all drawbacks of simplified circuit design (in my opinion), and in the end you need huuuge capacitors or a regulator on the power rail.
By the way, I completely forgot another possible cause of the hum: near the input connector there's a power connector for the protection relay. The traces from it run all the way around the board (along the edge) and could easily be picking up interference. I'll have to try rerouting them.
There's another strange effect, too. During tests with the old computer's built-in sound card, the amp started self-oscillating even with no signal. As I turned the volume control past a certain point, the input voltage suddenly dropped (current draw increased), and one of the resistors on the board started getting extremely hot. This didn't happen in any of the other tests, even at maximum levels.
I think the culprit is poor layout on the computer's sound card and connectors, but it did happen.
And a few words about what I think of all this.
First, the circuit design and all that talk of velvety highs, transparent mids, and punchy bass. It's all subjective impressions layered over the shortcomings of the circuit design. The thing is, a properly designed device is hardly sensitive to its components; many things can be compensated for through circuit design. Simple circuits, whether the one shown in this review or an even simpler John Linsley-Hood design, are highly sensitive to both the power supply and the quality and specifications of their components precisely because of their simple design. As a result, choosing different components affects the sound about as much as using different types of wood affects the sound of acoustic instruments. This is often considered part of the high-end experience. And the claims about hand-matched components and other such perks are mostly marketing; the fact that the amp really does perform worse without them is once again due to the design's shortcomings, not its merits.
Take the DarTZeel under review, for example. There's nothing special about it, really—just an ordinary amplifier. In my opinion, Sukhov's power amplifier would leave it miles behind, but for some reason people keep building Hood amplifiers and similar designs.
I think it's simply a hobby: searching for that perfect sound, the sound of different components. Though I'd look toward better designs, with automatic offset adjustment, short-circuit protection (this one doesn't have it), and so on.
In short: the amplifier works and sounds pretty good. It has built-in speaker protection. Its main drawbacks are excessive heat and output offset drift.
This review was sponsored by the purchasing agent yoybuy.com, which covered the shipping costs.
The kit cost $46.88 including shipping to the agent and weighed 561 grams. Shipping from the agent depends on various factors, such as the quantity and whether there are other items in the order.
$42.5
This page was found when searching for:
клон усилитель звука из китая, усилитель tierra схема, усилок на 5200 и 1493 с защитой по выходу, 13, дартзила, усилитель dartzeel электросхема, 8, 41, dartzeel nhb-108, 5, "TiERRA" интегрированный усилитель класса А схема, китайские усилители hi-fi сборные конструкторы, тесты китайских плат усилителей на транзисторах, 4, 2, усилитель dartzeel упрощённая версия, burmester 933 клон купить, 1, 9, умзч хитачи, умзч dartzeel, клон усилителя дартзеел китайский фильтр, лучшие схемы усилителей, усилитель - клон агеева, клон усилители для наушников а класса day kit, набор для сборки усилителя, сборка усилителя своими руками, собрать усилитель звука, комплект для сборки усилителя
клон усилитель звука из китая, усилитель tierra схема, усилок на 5200 и 1493 с защитой по выходу, 13, дартзила, усилитель dartzeel электросхема, 8, 41, dartzeel nhb-108, 5, "TiERRA" интегрированный усилитель класса А схема, китайские усилители hi-fi сборные конструкторы, тесты китайских плат усилителей на транзисторах, 4, 2, усилитель dartzeel упрощённая версия, burmester 933 клон купить, 1, 9, умзч хитачи, умзч dartzeel, клон усилителя дартзеел китайский фильтр, лучшие схемы усилителей, усилитель - клон агеева, клон усилители для наушников а класса day kit, набор для сборки усилителя, сборка усилителя своими руками, собрать усилитель звука, комплект для сборки усилителя
Questions and answers
What are the stated specifications of the DarTZeel clone boards?
Each board measures 145 × 92 mm; stated output is 2 × 200 W into 4–8 Ohm loads. The boards require ±35–55 V DC and an auxiliary 12–15 V AC supply.
How does the author assess the DarTZeel clone amplifier?
It works and sounds fairly good, with built-in speaker protection. The author lists substantial heat and output DC offset drift among its main drawbacks.
Does the DarTZeel clone have short-circuit protection?
No. The author says it lacks short-circuit protection, although speaker protection is built in.
The figures from this review are in the measurement database: Other devices
Calculate manuallyAll calculators
Discounted products
You may be interested in
Discounted products
Support the project
All the content on the site and the advice are free, but we’d be grateful for your support of the project and the channel!
Comments2
Discuss on the forumHave a question or some experience to share? Discuss the article with the author and readers: the forum is a convenient place to share schematics and photos.
Go to the forumOnly registered readers can comment. It takes a minute to sign up, and you can log in with the same account you use on the forum.
Log inRegister







Comments are published in their original language.