Hi Enno, Fred, all

There are two things which I think can increase power output. Talking here, on the assumption the transceiver is already performing properly i.e. correct LPF etc. I expect Fred is referring to the first of these two:

1) QCX uses a PNP transistor Q6 to shape the RF envelope on key down/key up to remove key clicks. There is a voltage drop across this transistor as usual. If it is replaced by a P-type MOSFET with appropriate other component value changes, then this eliminates most of the voltage drop. Since the Class-E PA then runs at a little higher supply voltage, it produces more power output.

2. Use of a MOSFET circuit instead of the schottky diode for power supply reverse polarity protection increases the component count slightly, but also has a much lower voltage drop, again providing a little higher voltage to the PA.

I don't know of any disadvantage to doing these modifications, they should work fine.

73 Hans G0UPL
http://qrp-labs.com

At power-up you are asked to select the band. All this does, is default the power-up VFO A, VFO B, Preset frequencies, alignment freqeuncies, and beacon frequency to a frequency 20kHz above the lower band edge of the chosen band. It does not do anything else to the configuration. If you want to change the band you can use menu item 3.9 to reset these parameters to the new chosen band; or, you can just use the rotary encoder to tune there, or manually set presets to the new band and load them into the active VFO. Lots of possibilities. But a new firmware chip is NOT required.

Three circuits have band-specific components:

1) Output low pass filter, for transmitter harmonic suppression. If you don't change that then you will either have too much fundamental attenuation (if your transmit frequency is above the filter frequency; or too little harmonic suppression (if your transmit frequency is too far below the filter frequency).

2) Input band pass filter, that protects the receiver mixer (to some extent) to strong out of band signals. If you bypass it, everything will still work but at the expense of impaired intermodulation performance since everything across the frequency spectrum will hit the mixer.

3) Class E PA resonant circuit: if you don't change it, then the transmitter PA will operate at lower efficiency; which will mean it will dissipate more heat. It may eventually lead to frying the BS170 transistors.

73 Hans G0UPL
http://qrp-labs.com

The QRX has no AGC. VK3YE designed a small and simple agc. By replacing the led and ldr with a small optocoupler a verry small unit could be created for the QCX.

Usefull ?

73 John PD7MAA

It is not likely that you'll break the MPS2907 but if you do, ZTX551 should be Ok too.

For the next batch I found the MPS751 which has easier availability, better price, and 2A capability. So we'll use MPS751.

73 Hans G0UPL
http://qrp-labs.com

I did not have any trouble myself, through all my development and testing. And I have heard now of several other people who built/operated it successfully with the supplied MPS2907 with no trouble. However, since I now have THREE reports of failure of this MPS2907, I will upgrade it to the ZTX550 for the 2nd batch (approx serial number 500 and up).

73 Hans G0UPL
http://qrp-labs.com

I have observed similar things on my 17m QCX, but only when running from a PSU. On a 12V SLA the output is 2.5W, but when I tried it on a PSU to see if increased voltage would increase the output I saw the output fall and the current increase like you describe.

I blew Q6 as well I replaced mine with a (pin compatible) ZTX550 which has a slightly higher rating (1A).

Mine has 1nF for C33 and is working OK on 18Mhz (2.5W output). When I get a moment I'll look at the voltages on Q5 during transmit.

Hans has worked on these sidetone and keyer problems. I have tested a beta version of the firmware and things are much improved. Hopefully Hans will release it soon after he is back from his well earned holiday.

CW playback: Also been fixed.

QSK: This has also been improved.

I've been having fun with mine (on 17m which has not been in very good shape).

PeterO
G0DZB

Hi Maik

Thanks for the feedback.

Do you have an AVR programmer? Or an Arduino which can do the trick of behaving like an AVR programmer? I can send you a new firmware version which fixes all the minor issues you mention (sidetone volume issues, CW stored messages only sending Message 1, some keying improvements, and various minor other things).

The issue with power output is interesting, and the blown PNP transistor. 0.9A is of course way too much and indicates another issue. I did not see anything lile that in all my testing or ever break one of the PNP transistors or BS170s. The MPS2907 should never see anything near 0.6A. And of course its power dissipation is very low, since it is either on or off most of the time except when doing the shaping at the key-down or key-up.

73 Hans G0UPL
http://qrp-labs.com

Hi,
last week I received my kit #156 that I pre-ordered at Hamradio fair in June. Band = 20 m.
Assembly was very easy with the detailed instruction manual by Hans. My compliments !

The Radio initially worked and the alignment process of the BPF , I/Q Balance and Phase High/Low was easy.
I connected it to my antenna and power meter and could see "some" Output. So I worked the Initial QSO with a ukranian station.

Later I checked Output power and current consumption and found out that something may be wrong…
Here are my findings and recommendations what to Change in the circuit:

The power was less than 1 Watt with 13.8V supply voltage and a current of approx. 0.9A was drawn.
Also I found that the the PA Transistors BS170 got very hot, as well as the keying Transistor Q6 (MPS2907).
While trying to optimize the bandpass filter coils and L4 suddenly Output power went fully down.
I checked the components and found out that Q6 was blown…. Looking at the datasheet of MPS2907 it revealed that this Transistor is specified for a continous collector current of 600 mA only. I replaced it with a BD236 that I had available and the Transmitter worked again. Unfortunately the BD236 is not pin compatible, so it was soldered on the backside of the PCB , bending the Pins into place. I would recommend to Change Q6 for a higher current rated PNP.
Now the Output power was still low and current too high. Further Investigation indicated that the Class E amplifier is somehow loaded by the path to the RX/TX Switch via capacitor C33. My Feeling was that at 14 MHz the value of 1nF is a Little too high. Probably too much RF is coupled to the drain of Q5 and causes some nasty effects here. So I changed C33 to 150 pF and immediately the current dropped to about 0.6 A and I had much more power. Optimization of the bandpassfilter inductors and the L4 choke resulted in 30-32 V p-p at 50 Ohm, so about 2.2 to 2.5 Watt. That was also measured with the internal power meter (really nice function!).

Further minor issues I found during the first QSOs:
- if the side tone Level is increased above a value of 50, the tone sounds very strange - there is some Kind of beat frequency involved and it appears that the volume of the tone goes up and down making it very difficult to use it in conjunction with the paddle.

- There seems to be a Problem with the playback of the CW Messages. I programmed the first three Messages but when I want to Play them , no matter what message I select, the Radio always Plays the message 1 only.

-QSK function is very cracking and clicking - anyhow I prefer semiBK anyway…

Ok, still a great Little kit for an unbeatable low Price.

73
Maik DJ2QV

I've managed to convince a few VK stations to build 30m versions of this kit!

Please keep in mind the following additional useful information:

Excellent construction and technical information manual is here (PDF).

Reduce current drain on RX from 140mA or so:

The backlight is connected to the + supply (nominally 12V, recommended range 7 to 16V) via a 270-ohm resistor. There would be nothing to stop you from cutting one of the wires of that resistor and taking a couple of wires to a switch, to control the backlight. Note that the BLUE backlight LEDs are a little harder to read with no backlight, than the yellow/green colour ones. Of course the module used in the QCX is the same as the Ultimate3S one http://qrp-labs.com/ultimate3/u3s - although it is a high quality module custom-manufactured to order for QRP Labs (Not the cheap eBay kind which have only a single backlight LED), it does have the same "industry standard" 80 x 36mm PCB and 16-pin connection pads. And the same HD44780-compatible controller chip. So you can substitute any other 16 x 02 LCD module if you wish, too.

"Switching" the backlight should be easy with an external switch. Options might be to open the backlight lead completely, add resistance to lower light intensity and reduce current draw, or add a push-button (momentary ON) to view the LCD with backlight only when desired

Modern LEDs as used in today's LCD displays will provide fairly bright illumination at currents of 3 to 10 ma. Try using a larger resistance value in series with the backlight LED. Usually 1K in series with a 5V or 2.2K with 12 volt source is adequate for both day and night viewing. Experiment to see what is best for your particular situation

Dual Band Operation

You won't be able to use the band pass filter T1 on either band. It is single band because it is quite sharp (see spectrum analyser traces in the manual). You could probably find a number of turns for the tuned winding of T1, which allowed the trimmer capacitor to tune for either 20m or 30m. Measured tuning ranges are also published in the manual. BUT, it would mean that when you want to change band, you would need to re-peak the BPF each time using the self-alignment menu item "8.7 Peak BPF".

You could build the LPF for 20m and operate 30m into it but you may not achieve adequate harmonic suppression on 30m band.

The Class-E resonant circuit is quite broad and you could find a compromise-value, it would probably result in a little loss of efficiency and power, but may be acceptable.

Another option would be to have a separate board with switched circuits for these three blocks (LPF, BPF, Class-E).

If you want to widen the filter for reception of SSB (on other bands):

The 200Hz filter is quite sharp (see the measurements in the back of the manual). You would need to bypass it. The receiver has quite a lot of available gain so the loss of 18dB may not worry you too much. But you might want to consider building a separate LPF with say 3kHz cut-off and a gain stage, and making them switchable.