Hi, 

I have just finished building a QCX Mini for 40 Meters, and everything aligned exactly as per the manual, fabulous!

Its connected to a laboratory grade power supply, and the output is hooked it up to my Siglent Spectrum Analyser via a 40db attenuator.

Having built a QCX for 60m previously, I was amazed that the output power is much more than I expected, and produces 6.32 watts from a 13.8 volt supply, so I took some more readings -:

7 volts    1.50 w     370mA
8 volts    2.05 w     420mA
9 volts    2.64 w     470mA
10 volts   3.30 w    510mA
11 volts   4.00 w    550mA
12 volts   4.75 w    600mA
13 volts   5.67 w   640mA
13.8 volts  6.32 w  680mA

My own feeling is this that the PA transistors will not tolerate the output power very long before they overheat, and fail, so has anyone else had this issue, and if so what is the best solution to reduce the output power. I am very hesitant to change the low pass filter, which is built exactly as per instruction, and also I do have the protection 1N5819 diode in circuit.

Any advice / help would be much appreciated 

73 Tim G4ARI

On 28/03/2023 15:53, Tim Raven via groups.io wrote:

My own feeling is this that the PA transistors will not tolerate the output power very long before they overheat, and fail, so has anyone else had this issue,

Tim,

Reduce the voltage!
It says 12 volts use that as a maximum.
If you look back there are many tales of QDX PA failures, keep power down.
Even half that amount will make much difference in results.

73 Alan G4ZFQ

On 28/03/2023 16:13, Alan G4ZFQ via groups.io wrote:

If you look back there are many tales of QDX PA failures,

For that you could read QCX, they are better than the QDX but do not run too hard.
I still say reduce voltage, any other way of reducing power can make a PA less efficient, make it run hot..

73 Alan G4ZFQ

Hi Tim,

If this is a QDX, I agree with Alan, and you need to reduce the voltage to the QDX.  

As far as I know, there is not the same power issue with the QCX.  If you are genuinely concerned, add one or two diodes to drop the voltage when used on the shack supply, and do not use them when portable.

The other option is to adjust the turn spacing on the lowpass filter to reduce the power out.  Usually, you are trying to adjust the windings to increase the power.  The opposite should work as well.  What is happening is the impedance match is adjusted to either a better or worse match to the output amplifier.

73
Evan
AC9TU

When I build mini's, I target 5w at 12v input. I run a lot with Bioenno batteries (nominally 12.9v for most of the discharge curve), with nary a problem. Go forth and communicate.

-mike/w1mt

On Tue, Mar 28, 2023 at 05:26 PM, Evan Hand wrote:

If you are genuinely concerned, add one or two diodes to drop the voltage

I agree, diodes are a good way but there's one little caveat when choosing a diode.

Choose one that is NOT overrated for the job.

Diodes tend to achieve the circa 0.6v drop when they are being run near their max current
rating.

Shove a 10 amp diode in line when drawing 500ma will usually only drop about 0.3v etc.

Or just include a few more diodes than necessary in the circuit to get the right voltage ;-)

Well that's been my experience, others might have found different results, so it's worth
investigating.

--
- 73 de Andy -

On Tue, Mar 28, 2023 at 06:43 PM, Ronald Taylor wrote:

Tim it’s not at all uncommon to get that kind of power out of a QCX Mini on 40 meters. I see it a lot  and it’s not going to cause any issues in normal CW operation. However, if you decide to run a 100% duty cycle operation like WSPR, please please lower the voltage to get it down to at least half that power

+100

 
--
- 73 de Andy -

On Tue, Mar 28, 2023 at 12:26 PM, Evan Hand wrote:

Usually, you are trying to adjust the windings to increase the power.  The opposite should work as well.  What is happening is the impedance match is adjusted to either a better or worse match to the output amplifier.

Use care with this, as moving the match to the "wrong" impedances may reduce the power by reducing the efficiency, which is the wrong way to go because it can increase the power dissipation.  Keep an eye on your voltage/current/output power to avoid this, and only use changes which decrease the dissipated power. 
Reducing the output amplifier voltage is a proven method of power reduction, which also minimizes dissipation.  If you power the final devices separately from the main supply you can go very low in voltage and hence power while the rest of the circuitry still has the correct voltages.
73, Don N2VGU

Just checking, the low-pass filter is working OK? Sometimes an inoperative LPF will produce extra power at the output…. Since you have a spectrum analyzer that should be easy to determine.

73, Willie N1JBJ

On Tue, Mar 28, 2023 at 08:53 AM, Tim Raven wrote:

My own feeling is this that the PA transistors will not tolerate the output power very long

I perfectly agree with your statement! Reducing the voltage like suggested from our friends above is a straight forward path to follow, but not the only way to tame the tune and by far not the best.

How about increasing the efficiency of the Class E amp like old master Nathan O. Sokal, WA1HQC instructs us to do in his original ClassE Article? 
Connect a scope to the transistors and adjust the LPF as such as to achieve a "soft landing". This way not only will you keep the power level within common sense, but you will produce less heat for the same output power AND you might even improve your harmonic and spurious emission. You have a nice spectrum analyser: take a look at the harmonics produced now and after tweaking for optimal efficiency and see how different this is compared to the brute-force adjustment intended to maximize the output power by all means. 

When adjusting keep in mind that the impedance of the paralleled Transistors is a function of the supply voltage and it will be hardly possible to achieve a perfect tune and match for a broad voltage range.

With PA's of this kind and for the lower bands (<20m) I usually aim for 2-3W @9V and not more than 3-4W @ 12V.
This might seem  little compared to the power figures circulated in this group, but it is an optimum that perfectly fits my personal needs.

Thinking that one such poor plastic transistor might bravely withstand 200mA, I like to avoid going higher than 600mA for the three of them. Looking to your figures, your tuning is not even so far from my liking!!!
You say 12V 4,75W 600mA? If I would be in a hurry, I would inspect the harmonic content, take a quick look at the shape of the waveform at the meeting point of the transistors before the LPF and be happy!

And I do one thing more, thing that is not agreed upon in this group, the FET transistors being known as NOT suffering from thermal runaway like their bipolar cousins do. I do not intend to fix something not needing fixing, but I like to distribute the load evenly over the three transistors, otherwise the bravest from them will take the most beating (like bad coachmen do with their poor horses). I pair the transitors according to their static parameters, i.e. same DS current for the same G voltage. This makes me at least feel better, even if it is not necessary.

Needless to say, with this tuning of the LPF and pairing of the transistors, I have never succeeded in burning the transistors regardless of the duty cycle or the load mismatch at the output. 

I have uploaded following picture showing a "soft landing" several times  in this group, but  it keeps on disappearing or at least, I can not find it myself. So please excuse my repetition, here it is again (it is for 30m, I should have one for 40m too...): What you see below is the gate drive, the DS voltage in the middle and down below the sine wave after the LPF. Please excuse the "ringing" of the gate signal due to the fact that the probe was badly screened and little mismatched.

Happy tuning!
73 de dl2arl

Tim, I use 11 volts from my MFJ 4230MV to power all of my 12 volt build QDX's. With careful attention to vswr on my antenna/tuner before keying.  Mark K9TR

All, 

Many thanks to everyone for their help with this. I should mention that this is a QCX and not a QDX as some have alluded to, which I intend to use it primarily on CW, but in the event that I do use WSPR, I will now ensure the supply voltage is reduced by making up a power lead with a couple of diodes in series.
 
I have replaced the 1N5819 with a 1N4001 which has a higher forward voltage drop, so that the output, when running  from my 13.8 volts PSU  is now less than 6 watts. 
I have measured the 2nd harmonic at 54dB down on the fundamental, which is slightly better than the figure measured by the ARRL in their review, so that has reassured me that the LPF is working as intended. Lastly I found a post about a QCX-mini 40 which was built by Hans and which mentioned in his QA report to the recipient, an output of 7 watts, so the power out I am getting doesn't now appear to be as out of the ordinary as I first imagined. 

I have secured L1, L2, L3, and L4 with a small amount of hot melt glue, and  has made no difference whatsoever to the output power.

The QCX-mini assembly instructions have a number of very good pictures of a 40m version, which show in detail the construction of all the toroids, so I copied these as closely as possible, and maybe this has helped produce this level of output power. The photograph below shows my completed QCX-mini.

73 Tim G4ARI