|
Hello all.
I've been having a great time operating my 7mhz QCX. Worked Russia, Switzerland, Denmark etc and had a great time.
This afternoon I hitched up the 'scope to a 50 ohm dummy load and checked the RMS voltage across it at 7.030 mhz. It was 11.6 volts RMS, equating to an RMS power output of 2.69 watts (I'm using a good 12 volt power supply, exactly 12v on both receive and transmit).
The waveform was a nice approximation to a sine wave. I'll upload a screenshot, in case any one is interested.
I then checked the output at frequencies above and below, at 100khz intervals. The highest output was around 7.9mhz, 14.1 volts RMS, i.e.3.97 watts.
I can type in the full range of results, again if anyone is interested.
Now I know that 50% more power doesn't make a huge difference, and the rig works very nicely as it is, but I am curious to know whether I could try tacking? some capacitors across the ones in the existing low-pass-filter and move the 3.97 watts into the 7mhz CW zone? Or might I risk messing up the filtering and causing mayhem? I seem to be okay at winding toroids, they have the specified number of turns and are tidy (my T1 winding went very well, first time!), so I don't think they are to blame.
Any advice or comments gratefully received.
Best wishes,
Julian
G4ETS
|
Oops, forgot to add the screenshots! Sorry
J
|
This is strange as the passband insertion loss should be fairly constant. At higher frequency the attenuation should increase, not decrease. Do you have any ripple in the passband?
Could it come from the resonant circuit at the input of the filter (L4)?
Jean-Claude
|
Have you got an FFT function on that scope ? It would be interesting to
look at the harmonics present.
PeterO
G0DZB
|
I'll type in the voltages at 100khz frequency intervals, from 5mhz to 8mhz, that should answer the ripple question.
And yes, have FFT on the scope, will play with both tomorrow and send in results.
Julian
|
One problem with using the QCX TX to measure the LPF is that the resonant L4+C30 will also be a part of the equation. Best would be to insert a known signal (SNA/VNA) at C29 (removed) and see how the LPF looks.
I tried simulating the 40m LPF in Elsie and increased the inductor values by 10% (~2 turns), the insertion loss went up from 0.22dB to 0.25dB @ 7030kHz, so nothing to worry about IMHO. Another test worth trying is to put the load before the LPF (disconnected) to see which frequency has the highest output, adjust L4. Nice to hear what results you get.
|
Here is the FFT image. I've also put an image with no signal, to give an idea of the noise floor.
The signal looks pretty clean to me. It's just that I'm not getting the power peak where I want it, i.e.7.030mhz.
Next step is to try what Daniel suggested? Another test worth trying is to put the load before the LPF (disconnected) to see which frequency has the highest output, adjust L4.
But I think I'll get on the air for an hour or two first, as it works fine and it's a lot of fun as it is!
73s
Julian
G4ETS
|
Here is a table of the RMS output voltage/power I'm getting.
I've tried tacking 20pf and 50pf capacitors across C30, but that lowers the output at 7.030mhz
Julian
G4ETS
|
Here is a table of the RMS output voltage/power I'm getting.
I've tried tacking 20pf and 50pf capacitors across C30, but that lowers the output at 7.030mhz
Julian, I think L4 is the key? I think Hans has said it's resonance is what makes the PA (almost) work in class E. And that is is one reason why the QCX is mono-band. 73 Alan G4ZFQ
|
OK Alan, when I'm feeling brave I'll try winding a new version of L4, with 19 turns instead of 16 and let you know how I get on.
J
|
OK Alan, when I'm feeling brave I'll try winding a new version of L4, with 19 turns instead of 16 and let you know how I get on. Julian, That's not what I meant. Use the 20m LPF, if it has a low cut-off then using it on 30m should show noticeably more power. If it does then you need to reduce the number of turns for it to work properly on 20m. 73 Alan G4ZFQ
|
I think that this table of outputs may have the assumption that the input power to the low pass filter is not varying with frequency. Since it is a Class E (or very nearly so) output, there may be a variance in output before the filter (there will be lots of harmonics interacting before it gets to the low pass filter).
In any case, you have only about 1.6 db difference from your peak output to your "desired" frequency output.
---------
Paul KE7HR
|
In the construction notes, Hans says "A class E Power Amplifier contains a resonant circuit at the frequency of operation, so it is only suitable for single band use".
All I am querying is whether my L4 toroid is tuned to somewhere in the 7mhz CW band, for which the QCX is designed, or perhaps it's resonant at 7.9mhz where I get the most output?
If so, will adding a few turns bring it down lower so it peaks "on the nose"?
Bear in mind that I grew up in the era of Class C, tune for maximum smoke!
Julian
G4ETS
?
?
|
In the construction notes, Hans says "A class E Power Amplifier contains a resonant circuit at the frequency of operation, so it is only suitable for single band use".
All I am querying is whether my L4 toroid is tuned to somewhere in the 7mhz CW band, for which the QCX is designed, or perhaps it's resonant at 7.9mhz where I get the most output?
If so, will adding a few turns bring it down lower so it peaks "on the nose"? Julian Bear in mind that I grew up in the era of Class C, tune for maximum smoke! My understanding is that that is what class E is, just more smoke for less heat! I failed to notice you had shunted C30 to try to get resonance. I do not know if the 20pF minimum you used was too much and the resonance went too low. Did you try? Or, maybe there was some other effect. From your results it seems that L1/C30+BS170 capacitance is resonant at 7.9MHz because there should be a peak. If I were doing it then yes, I'd try a few more turns on L1 but must admit I really do not have a clue. 73 Alan G4ZFQ
|
For an HF linear I was advised many years ago by G2DAF to tune for a peak in screen current.
Class E is not linear and you tune for the correct load impedance - not for maximum output.
John F5VLF
toggle quoted message
Show quoted text
On 11/09/2017 09:47 AM, Alan G4ZFQ wrote: In the construction notes, Hans says "A class E Power Amplifier contains a resonant circuit at the frequency of operation, so it is only suitable for single band use".
All I am querying is whether my L4 toroid is tuned to somewhere in the 7mhz CW band, for which the QCX is designed, or perhaps it's resonant at 7.9mhz where I get the most output?
If so, will adding a few turns bring it down lower so it peaks "on the nose"? Julian
Bear in mind that I grew up in the era of Class C, tune for maximum smoke!
My understanding is that that is what class E is, just more smoke for less heat!
I failed to notice you had shunted C30 to try to get resonance. I do not know if the 20pF minimum you used was too much and the resonance went too low. Did you try? Or, maybe there was some other effect.
From your results it seems that L1/C30+BS170 capacitance is resonant at 7.9MHz because there should be a peak.
If I were doing it then yes, I'd try a few more turns on L1 but must admit I really do not have a clue.
73 Alan G4ZFQ
|
For an HF linear I was advised many years ago by G2DAF to tune for a peak in screen current.
Class E is not linear and you tune for the correct load impedance - not for maximum output. John Ah! So should Julian just leave it as it is or is there a simple way to check? I assume therefore that should Julian monitor the PSU current it will show lower efficiency at 7.9MHz? Julian, When you made your power output chart was that after the LPF? ((You did say "Next step is to try what Daniel suggested Another test worth trying is to put the load before the LPF (disconnected) to see which frequency has the highest output, adjust L4". Which could give misleading results unless you ONLY measure the 7MHz component, not with all the possible harmonics.)) 73 Alan G4ZFQ
|
I meant maximum amplitude (Vpp) in the unfiltered output of the class E connected to a 50ohm load, not RMS voltage. You should still use a 100nF as DC blocking as well. The waveform shown in the manual in section 7.4, at the positive peak when the transistors are not conducting, should be maximised at the resonant frequency ?! Use BW limit in the scope if it looks too messy.
Maybe replace C30 with a trimmer, such as a JML06 ? I'm rebuilding a 17m version to 20m soon, and will probably experiment a bit with this as well.
|
Hello Alan,
When you made your power output chart was that after the LPF?
Yes, I'm just putting a 50 ohm dummy load across the antenna output and looking at the output with my 'scope.
Julian G4ETS
|
Alan,
Most of my experiments with Class E have been with the Scopematch designed by M0BMU. As with other kits produced by Hans it is best to build it as per instructions and then investigate if it turns out not to do what is expected.
Regarding the specific matter of estimating efficiency at various frequencies by observing the PSU current, I do not have enough experience to comment.
73 John F5VLF PS sorry for the delay in replying, but we have had substantial building works in the house and this has slowed everything down
toggle quoted message
Show quoted text
-------- Forwarded Message --------
> For an HF linear I was advised many years ago by G2DAF to tune for a
> peak in screen current.
> Class E is not linear and you tune for the correct load impedance - not
> for maximum output.
John
Ah!
So should Julian just leave it as it is or is there a simple way to check?
I assume therefore that should Julian monitor the PSU current it will
show lower efficiency at 7.9MHz?
Julian,
When you made your power output chart was that after the LPF?
((You did say "Next step is to try what Daniel suggested Another test
worth trying is to put the load before the LPF (disconnected) to see
which frequency has the highest output, adjust L4". Which could give
misleading results unless you ONLY measure the 7MHz component, not with
all the possible harmonics.))
73 Alan G4ZFQ
|
Daniel Ekman SA2KNG
Paul Jorgenson KE7HR