Topics

Low Power Output - S/N 2857 - 20M

Jim - AC0OW
 

As-built had power out of 1.44 watts (24vpp) at L1 out (with 50 ohm dummy load).  Checked resistance via L1,L2,L3 - All okay.  Power at BS170 drains was ~ 5 watts.  Following fault-finding guidance, I took 2 turns off each of L1,L2,L3 and again checked continuity.  Power went up to 2.25 watts (30 Vpp).  Not much improvement.  Checked power just prior to L1 and got 5 watts.  So, Where should I go from here.  Looks like something after L1 must be loading down the Xmit chain.  Any suggestions would be appreciated.

Otherwise rig is great.  Worked PJ4 for first contact. even with QRPp Hi.  

Richard Harris
 

Hi Jim,

A builder of a 20m QCX mentioned that C25 and C26 (390pF) in the output filter were getting hot when transmitting.  I had just completed a 20m QCX and so I checked these capacitors with an IR thermometer and they were also warming up on transmit.  I cut them out and replaced with 330pF + 56pF NPO 100v ceramic capacitors (soldered on the back of the PCB) and the output power rose significantly  It may be worth checking these capacitors, particularly if you have the 20m version. 

Richard G3OTK  


On Tuesday, 20 February 2018, 19:49, Jim - AC0OW <ars.ac0ow@...> wrote:


As-built had power out of 1.44 watts (24vpp) at L1 out (with 50 ohm dummy load).  Checked resistance via L1,L2,L3 - All okay.  Power at BS170 drains was ~ 5 watts.  Following fault-finding guidance, I took 2 turns off each of L1,L2,L3 and again checked continuity.  Power went up to 2.25 watts (30 Vpp).  Not much improvement.  Checked power just prior to L1 and got 5 watts.  So, Where should I go from here.  Looks like something after L1 must be loading down the Xmit chain.  Any suggestions would be appreciated.

Otherwise rig is great.  Worked PJ4 for first contact. even with QRPp Hi.  


SM7ETW Jan
 

Hi

You can NOT measure power before the lowpass filter, the waveform is very far from sinus, also if using the built in wattmeter note that it will also react for DC voltages. You must tell us the supply voltage and before or after the protecting
diode.
Mine gives 1 Watt at 7.5 V before the diode.

Jan

SM7ETW Jan
 

To check the low pass filter you could try and transmit on 13 and 12 MHz into a dummy load. If ok, then step upwards from the ok frequency to check the curve of the filter.

Jan

Jim - AC0OW
 

Thanks for all the suggestions today.  Hans, any ideas?

Right now I am thinking C28 must be bad.  Not likely, I suppose, but don't know what else it could be.  Power is okay at output of L2 (nice sine wave 42Vpp) and low at output of L1 (24Vpp) all with 50 ohm dummy load.

Any other ideas out there?  Thanks,

Alan de G1FXB
 

Hi Jim,

I'm not sure if Serial No# 2857 is a Rev3 PCB if not, then you should apply the mods yourself.

http://qrp-labs.com/qcx/qcxmods


Further reading the www.qrp-labs.com/qcx/qcxtrouble.html document seems to provide pointers on this topic?

A cut and paste of the relevant paragraphs results in:-

??My GENERAL RULE is that we intuitively always want to suspect defective components, rather than our own mistakes in assembly; but when we find out what is wrong, it usually turns out to be OUR MISTAKE, not a defective component! We want to make this work, not massage our ego... so always suspect defective components ONLY as a last resort when all other avenues of investigation are exhausted.

ONLY measure power across a 50-ohm dummy load such as the QRP Labs dummy load kit, and only measure the power output at the proper RF output connector of the transceiver, after it has been filtered into a nice sinewave by the Low Pass Filter.

Note also that to get 5W power output does require a 15 or 16V supply voltage. You can see this from the power vs supply voltage curves in the manual.

Finally if you really think your power output is a lot lower than you expected - then you can look at the windings on L1, L2 and L3, the Low Pass Filter toroids. There can be variation in component value due to manufacturing tolerances, and variation in inductance values due to winding style and other factors. First check that you wound the correct number of turns for your band. Remember that what counts as one "turn" is every time the wire passes through the toroid ring's hole. Count the turns again to make sure you got the correct number.

Simply removing a couple of turns from each of L1, L2 and L3 is usually sufficient to raise the filter cut-off, if it was too low - and you should notice a power output increase.


Alan



On 20/02/2018 23:25, Jim - AC0OW wrote:
Thanks for all the suggestions today. ??Hans, any ideas?

Right now I am thinking C28 must be bad. ??Not likely, I suppose, but don't know what else it could be. ??Power is okay at output of L2 (nice sine wave 42Vpp) and low at output of L1 (24Vpp) all with 50 ohm dummy load.

Any other ideas out there? ??Thanks,

Alan de G1FXB
 

Hi Richard,
In the interest of providing an alternative view,

?if you observe capacitor heating in the LPF ?

? https://groups.io/g/QRPLabs/message/21483


If the capacitors get hot, It maybe because the LPF cut-off is too low?

http://qrp-labs.com/qcx/qcxtrouble.html#lowpower


Alan


On 20/02/2018 20:18, Richard Harris via Groups.Io wrote:
Hi Jim,

A builder of a 20m QCX mentioned that C25 and C26 (390pF) in the output filter were getting hot when transmitting.? I had just completed a 20m QCX and so I checked these capacitors with an IR thermometer and they were also warming up on transmit.? I cut them out and replaced with 330pF + 56pF NPO 100v ceramic capacitors (soldered on the back of the PCB) and the output power rose significantly? It may be worth checking these capacitors, particularly if you have the 20m version.?

Richard G3OTK??


On Tuesday, 20 February 2018, 19:49, Jim - AC0OW <ars.ac0ow@...> wrote:


As-built had power out of 1.44 watts (24vpp) at L1 out (with 50 ohm dummy load). ?Checked resistance via L1,L2,L3 - All okay. ?Power at BS170 drains was ~ 5 watts. ?Following fault-finding guidance, I took 2 turns off each of L1,L2,L3 and again checked continuity. ?Power went up to 2.25 watts (30 Vpp). ?Not much improvement. ?Checked power just prior to L1 and got 5 watts. ?So, Where should I go from here. ?Looks like something after L1 must be loading down the Xmit chain. ?Any suggestions would be appreciated.

Otherwise rig is great. ?Worked PJ4 for first contact. even with QRPp Hi. ?



Richard Harris
 

Hi Alan,

Many thanks for the link - I think Hans posted this after I posted about the warming up of C25/C26. I'm going to disagree with Hans that if the LPF frequency is too low then it much result in increased dissipation in the filter components. This would only occur if the currents within the filter increase significantly with, say, greater than optimum inductor values (i.e. lower cut-off frequency) and the capacitors have a significant loss component.

Richard G3OTK

On Tue, Feb 20, 2018 at 05:26 pm, Alan de G1FXB wrote:


Hi Richard,
In the interest of providing an alternative view,

?if you observe capacitor heating in the LPF ?

https://groups.io/g/QRPLabs/message/21483


If the capacitors get hot, It maybe because the LPF cut-off is too low?

http://qrp-labs.com/qcx/qcxtrouble.html#lowpower


Alan


On 20/02/2018 20:18, Richard Harris via Groups.Io wrote:
Hi Jim,

A builder of a 20m QCX mentioned that C25 and C26 (390pF) in the
output filter were getting hot when transmitting.? I had just
completed a 20m QCX and so *I checked these capacitors with an IR
thermometer and they were also warming up on transmit.*? I cut them
out and replaced with 330pF + 56pF NPO 100v ceramic capacitors
(soldered on the back of the PCB) and the output power rose
significantly? It may be worth checking these capacitors, particularly
if you have the 20m version.

Richard G3OTK


On Tuesday, 20 February 2018, 19:49, Jim - AC0OW <ars.ac0ow@...>
wrote:


As-built had power out of 1.44 watts (24vpp) at L1 out (with 50 ohm
dummy load). ?Checked resistance via L1,L2,L3 - All okay. ?Power at
BS170 drains was ~ 5 watts. ?Following fault-finding guidance, I took
2 turns off each of L1,L2,L3 and again checked continuity. ?Power went
up to 2.25 watts (30 Vpp). ?Not much improvement. ?Checked power just
prior to L1 and got 5 watts. ?So, Where should I go from here. ?Looks
like something after L1 must be loading down the Xmit chain. ?Any
suggestions would be appreciated.

Otherwise rig is great. ?Worked PJ4 for first contact. even with QRPp Hi.


Alan G4ZFQ
 

I'm going to disagree with Hans
Richard

I'm just an ignorant observer.
But it seems strange that capacitor heating is being noticed in these LPFs, standard design and components, copied through to the QRP Labs kits.
I suppose it could be sub-standard capacitors, does it happen on all 20m versions?

73 Alan G4ZFQ

In the interest of providing an alternative view,

?if you observe capacitor heating in the LPF ?

https://groups.io/g/QRPLabs/message/21483


If the capacitors get hot, It maybe because the LPF cut-off is too low?

http://qrp-labs.com/qcx/qcxtrouble.html#lowpower

Alan G4ZFQ
 

>I suppose it could be sub-standard capacitors, does it happen on all 20m versions?

I suppose if anyone has ONLY replaced the capacitors with silver mica and got appreciably more power out then it would confirm.
I've never been certain, most others have seemed to adjust the inductors as well.

73 Alan G4ZFQ

Alan de G1FXB
 

Hi All,

Like I said, posted in the interest of a alternative view.
People who search the threads and may not go beyond this one to read the others.....
So to provide comment from the kit supplier, on this subject but on another thread
Not intended as a definitive statement / cure all, for everyone.

We debated the unknowns,
No# of wire turns and torrid material reproducibility, compared the colour paint, to death previously......

As ever components have a 'tolerance' value, both in electrical value and quality.
capacitors, torroid material variations, and the greatest variable is our winding / turn spacing / mutual coupling?

(It's the RX BPF but
Didn't one builder check the inductance of the three identical windings on T1 and two were close and the third was wildly out.
I'm not sure that was ever solved?)

It all adds & subtracts up, we hope it all works in our favour. Sometimes it doesn't.
When it doesn't then it requires a change to the properties of the circuit.

Alan

On 21/02/2018 10:16, Richard Harris via Groups.Io wrote:
Hi Alan,

Many thanks for the link - I think Hans posted this after I posted about the warming up of C25/C26.  I'm going to disagree with Hans that if the LPF frequency is too low then it much result in increased dissipation in the filter components.  This would only occur if the currents within the filter increase significantly with, say, greater than optimum inductor values (i.e. lower cut-off frequency) and the capacitors have a significant loss component.

Richard G3OTK



On Tue, Feb 20, 2018 at 05:26 pm, Alan de G1FXB wrote:

Hi Richard,
In the interest of providing an alternative view,

 ?if you observe capacitor heating in the LPF ?

https://groups.io/g/QRPLabs/message/21483


If the capacitors get hot, It maybe because the LPF cut-off is too low?

http://qrp-labs.com/qcx/qcxtrouble.html#lowpower


Alan


On 20/02/2018 20:18, Richard Harris via Groups.Io wrote:
Hi Jim,

A builder of a 20m QCX mentioned that C25 and C26 (390pF) in the 
output filter were getting hot when transmitting.? I had just 
completed a 20m QCX and so *I checked these capacitors with an IR 
thermometer and they were also warming up on transmit.*? I cut them 
out and replaced with 330pF + 56pF NPO 100v ceramic capacitors 
(soldered on the back of the PCB) and the output power rose 
significantly? It may be worth checking these capacitors, particularly 
if you have the 20m version.

Richard G3OTK


On Tuesday, 20 February 2018, 19:49, Jim - AC0OW <ars.ac0ow@...> 
wrote:


As-built had power out of 1.44 watts (24vpp) at L1 out (with 50 ohm 
dummy load). ?Checked resistance via L1,L2,L3 - All okay. ?Power at 
BS170 drains was ~ 5 watts. ?Following fault-finding guidance, I took 
2 turns off each of L1,L2,L3 and again checked continuity. ?Power went 
up to 2.25 watts (30 Vpp). ?Not much improvement. ?Checked power just 
prior to L1 and got 5 watts. ?So, Where should I go from here. ?Looks 
like something after L1 must be loading down the Xmit chain. ?Any 
suggestions would be appreciated.

Otherwise rig is great. ?Worked PJ4 for first contact. even with QRPp Hi.







Jim - AC0OW
 

Thanks for the response on this.  QCX is fixed and power is all okay at 4 watts.  Problem was L1.  I had to take an additional 3 turns off (for a total of 5) and it now works.  Looks like the toroid for L1 was significantly different than L2 & L3.  

Hans, you may wish to include in your troubleshooting guide.  To check if power loss is uniform across L1,L2, L3.  

Many thanks for a great little rig.  Look forward to using portable and buying another band or two.

Alan G3XAQ
 

Has anyone measured the inductance of L1-L3 when they had low output? The target values are in the build manual. In my (limited) experience inductors wound on Micrometals toroids pretty much never come up on the manufacturer's target value unless you squeeze or spread the turns while watching on an inductance meter.

For example, Micrometals do say 16 turns are needed on a T37-6 core for 770nH as is specified for the 20m LPF. I just pulled some T37-6 cores out. They were supplied by G-QRP club. 16 turns evenly spread around the core gave 880nH. 15 turns gave 710nH. 14 turns over 2/3rds the core circumference gave 756nH and I could get anywhere from 578nH to 846nH depending on how much or little I spread the 14 turns (see the photos). And to make matters worse Micrometals themselves say the permeability of their cores can vary by 20% between batches.

The theoretical design of the 20m LPF seems to be a 0.02dB ripple Chebychev with a cutoff frequency of 15MHz. It offers 46dB attenuation of the second harmonic and, probably more importantly, it presents an accurate 50R to the PA drains when the output is terminated in 50R. Increasing the inductors by 15% (770nH becomes 880nH, 900nH becomes 1.03uH) brings 14MHz to the very edge of the LPF cutoff and degrades the match to a return loss of 15.7dB (43+14j). I know nothing about class E PAs so cannot comment whether or not this would be expected to affect the output power. At 3.25W output the first 390pF in the filter is designed to carry 590mA of RF. If the inductors are 15% high the capacitor current increases 30% to 770mA and any heating due to ESR will increase by 70%.

I'm a bit surprised that the output seems sensitive to slightly high inductances but its very (very!) easy to miscount the number of turns, or the core permeability might be a bit high, and the capacitors could I suppose be on the edge of their current handling ability.

73, Alan G3XAQ