August 20, 2016:
Revised: v1.0

Taming a Chinese Nighthawk 40m CW Transceiver Kit

This Chinese CW transceiver kit had a host of problems. With some effort, I got all of this kit’s many design faults resolved and the transceiver working nicely. A very satisfying result. Here are the details to quickly and easily fix your kit. 

Figure 1 : The completed Nighthawk 40m CW transceiver, at present without a top cover


Some months ago, I ordered a transceiver kit from one of those Chinese suppliers on the internet. Since we cannot get mail delivered reliably here in the desert, I get items like this kit sent to an address in London. When I next visit the city, I collect my mail. As you can imagine, a few months can easily pass between an order, the delivery from China, and my collecting it.

Despite the delays,this system seems to work reasonably well. However, when I collected this transceiver kit, it became quickly obvious that the kit I had received was completely different to the one I had ordered!
The kit delivered was a Nighthawk 40m CW transceiver (夜鹰), but I had actually ordered a “Kangaroo” ( 袋鼠) SSB transceiver kit. (How can anyone mistake a kangaroo for a nighthawk?  ;-)  ) 

This was to be just the first of a lengthy saga of problems with this kit.

Too much time had passed since my order to address the problem with the supplier in any sensible way, and my verbal Cantonese and Mandarin skills are non-existent, so I decided to just build the kit I’d received.

The kit instructions explained - once I’d translated them from Chinese into English - that the original design had been “improved in a number of areas”. The claimed improvements included:

And at first glance, this kit should have been a really easy to construct and get going – The design is 90% based on the well-known and widely used SW-40+ CW transceiver designed by Dave Benson K1SWL. What could possibly go wrong? Building it, I decided, would be a nice, relaxing construction experience...

Kit Components

The first step in kit building is to sort out the parts and check them against the parts list provided in the kit instructions. Instructions….? As usual with these kits, there were no instructions supplied. However, after little searching, I managed to obtain a set of PDF-format instructions via the supplier’s website. Great! In Chinese…. Ah, not so great.

Figure 2 : The kit of parts contained everything required to build the kit except the
instructions required to build the kit. These were obtained from the supplier. In Chinese

Fortunately, I can handle that problem. Once I had translated the instructions, I found them to be quite detailed and accurate, at least as far as the information required to build the kit and operate the transceiver was concerned. My translation of these kit instructions can be downloaded below. Feel free to criticise my Chinese translation skills.

Most of the parts provided were of adequate quality, and everything on the parts list was in the kit. The PCB was excellent. The layout made some component labels printed on the PCB overlay a little hard to read, but there were no significant problems.

All of the NE602 ICs supplied were quite clearly recycled parts. Corrosion was evident on almost all of their pins, and the parts had obviously been removed from the original PCBs with minimal care. Despite that, and after some cleaning and tidying up with the soldering iron, all of these chips were found to be working fine.

Little did I know at this stage that the biggest component problem in the kit was going to be the variable shielded inductors. The 7 x 7mm “7 MHz” inductors supplied in the kit were completely different from the now impossible to find original “42IF123” 10mm x 10mm shielded inductors used in the SW-40+ design. Internally, the “7 MHz” coils supplied proved to have a different primary : secondary turns ratio, a critical design parameter.

Nighthawk Kit Design Issues

I think it’s fair to say that all of the problems I encountered with the kit can be traced to the kit’s “design improvements”, i.e. the modifications made to the original SW-40+ design. These changes, IMHO, meant that this kit, as supplied, could never ever work properly.

The problems affect both transmitter and receiver. For example, the transmitter mixer circuit around U5 (NE602) did not work. The design changes in the kit resulted in an excessive VXO input level to this mixer. It could not then generate the required 7MHz transmit signal. The incorrect inductors supplied for use in the subsequent transmit bandpass filter also meant it could never work correctly either. In addition, the filter’s high impedance output was excessively loaded by the low input impedance input of the subsequent Q6 amplifier stage. This prevented correct alignment of the filter, a problem reported on several websites. All told, quite a mess.

In the receiver, the newly added diode variable attenuator did not work correctly either. It did provide variable attenuation, but it ranged from a minimum attenuation of about 20dB to a maximum of 40 dB. Basically, that circuit simply reduced all received signals by at least 20dB – The receiver was completely deaf! Like the transmitter, this kit’s receiver as supplied would never work correctly.

Fortunately, these problems can all be easily fixed. With some changes to the Nighthawk design, based on a number of hours spent with my circuit simulation software and some bench testing, some extra parts and a bit of circuit rearrangement got everything working properly. For example, with a couple of minor changes to the variable diode attenuator, it now has an adjustment range of about -3 to -27dB. That’s about 5dB more dynamic range than the original kit “design”, and it has proven to be really quite nice to use in practice.

 Figure 3 : The completed transceiver board. If you look really carefully (Right click for more detail),
you will see a few places where modifications have been done to get everything working properly.

The Nighthawk kit also featured changes to tx/rx switching, with the addition of an antenna changeover relay and a delay circuit to reduce relay chatter during CW transmissions. I found that a minor change to the capacitor value controlling the changeover time was desirable. It doesn’t quite match the full break-in keying capability of the original SW-40+, but the minor circuit change in this area seem to work fine for me.

The changes made to the transmitter power amplifier stages operate well in practice, I’m pleased to say. They appear to be quite stable and trouble-free. However, another related design error creates a significant problem. Amateur radio operators often use QRP transceivers like the Nighthawk in locations requiring battery power. That demands a low receiver current drain.

For some strange reason, in this Chinese version of the SW-40+ transceiver, the gate bias voltage for the final power amplifier FET (Q5 – RD15HVF1) was designed to use the unswitched 8V regulated supply rail (VCC_8V). This added an extra and unnecessary current of up to 130mA on receive(!) By making a minor change to the design, rerouting R29 (4k7) so it connects to the transmit 8V rail (VCC_8VT), I reduced the receiver current from 160mA to just 30mA, making QRP battery operation of the transceiver practical.

So, what changes are required to get everything working?

Modification Details

Modifications (a), (b) and (c) are critical for correct operation. Modifications (d) and (e) are highly recommended, but not essential.

a.    Receiver variable attenuator
Figure 4 : Receiver RF attenuator modifications (Right-click for more detail) 

This modification reduces the minimum signal attenuation to acceptable levels while also increasing the range of adjustment.

b.    Transmit mixer oscillator input
Figure 5 : Transmit mixer circuit modifications (Right-click for more detail)

This modification reduces the relay hold time and speeds up the tx/rx changeover.

c.    Transmitter bandpass filter

 Figure 6 : Transmitter BPF modifications (Right-click for more detail)

This modification ensures the BPF operates correctly and can be aligned.

Note: I opted for top-coupling (with the extra 3p3 capacitor) rather than modify the value of C27. Chances are, your kit could have a different brand/type of Chinese inductor supplied for T2 and T3. In that case, the turns ratio could be different, and then we would be back to square one again with a modificationinvolving C27. Using top-coupling avoids that problem.

Here’s a picture showing the capacitors added under the PCB.

Figure 7 : Changes are required under the PCB (See text)

d.    Tx/Rx Timing

This non-critical modification reduces the relay hold time and speeds up the tx/rx changeover.

e.    Final Power Amplifier (Q5) Bias

This non-critical modification reduces receive current from 160mA to 30mA. Transmitter keying characteristics are unaffected.

This modification also requires a change to the original (and slightly obscure) bias setting instructions. The gate bias voltage on Q5 (RD15HVF1) should be about 2V, the gate threshold voltage. This ensures the PA stage operates in Class-C, suitable for CW transmission. (I was tempted to modify the PA for Class-E operation, to improve efficiency still further, but I had to draw the line somewhere in this project!)

The correct PA gate bias voltage for your transceiver can vary from 1.8 to 2.3V, according to the datasheet for Q5. To set this value correctly, we must monitor the drain current of Q5 during transmit, without any RF input feeding that stage. After the R29 bias modification, this adjustment is a little more difficult to do because there is no provision in the transceiver to measure the PA stage current under no-RF transmit conditions.

But here’s how we can do it AFTER this modification has been made:

Other Changes

There are a couple of areas in the transceiver remaining for me to improve. One is the keyer. The dit/dah timing ratio doesn’t quite sound quite right to me when I’m transmitting, although I’m hardly an expert! I also found the various keyer adjustment steps (e.g. keyer speed, keyer sidetone) to be too coarse.

To fix this, I’d ordinarily just drop in a ATtiny and some new software but the PIC12F629 pinout doesn’t allow for a direct plug-in chip replacement. I’ll have to wait for a very long rainy weekend before I write new assembly code software for a PIC chip. (And since I live in the middle of a very large hot desert, that may not be for some time…) More than likely, I’ll take the easy way out and “shoehorn in” an ATtiny YACK keyer. It's got loads of features and a ton of adjustments.

The other section of the transceiver I’d like to modify is the transmitter power amplifier. I think it is possible to make some really useful improvements there by migrating the design to Class-E, especially if the transceiver is to be used on battery power. That would require some fairly extensive changes, but it certainly looks possible.

Putting it in a Box

Now that i had it all operating correctly, to complete my transceiver, it needed to be in a suitable box. I did have a perfectly sized aluminium box, but the BNC antenna connector on the kit PCB made it impossible to fit the completed PCB in place. Also, I don’t have any metalwork facilities. Instead, I decided to build a box from blank PCB. That material is much easier for me to handle.

The result is shown in Figure 8. I ended up removing the BNC connector to allow the board to fit in the box, and I fitted my usual “phono” antenna connector in place of the BNC. I know this will upset some people, but I have found those phono connectors to be perfectly adequate for QRP use.

I printed the two knobs I needed on my 3D printer. The original design for these came from Thingiverse. They just press into place on the potentiometer shafts.

Figure 8 : An inside view of the completed (modified) Nighthawk transceiver. It’s a tight fit,
and measures about 100mm wide, 40mm high and 130mm long.


To sum up, while it was disappointing to discover so many fundamental design faults with this Chinese kit, and some poor quality parts, I have to admit to a great deal of satisfaction in hunting down all of the design problems, one by one, and fixing every single one of them.

This is not a difficult kit to build, especially now I know how to fix the problems. The high quality PCB and the board layout certainly made the build, and the various modifications, a great deal easier than some other equipment I’ve worked on in the past, and the finished transceiver looks very tidy, I think.

With all of the modifications added, and they are not hard to do, I think this transceiver now works quite well. I’m not sure it will ever be my favourite radio, but I do like the final result. It is compact, and the combination of close to 10W output power and its reasonable receiver sensitivity gives good results.

If you need an antenna tuner for the Nighthawk, I’ve designed a compact antenna tuner which works nicely. It’s described here on my website.

Finally, I hope the information I’ve provided here will help others who may be struggling with getting their Nighthawk transceiver kit operating correctly.

Figure 9 : The back panel of the transceiver records the basic specifications


Translated Kit Instructions: This zip file contains my translation of the ORIGINAL kit instructions. The modifications described above are recommended if you wish the kit to operate correctly.

Nighthawk Panel Artwork: My front panel artwork, in case someone wants it

Want to go back to the main page? Click here to return directly.