Mismatched antenna on Class-D & Class-E amplifier


Wolfgang OE1MWW
 

Hi all,

For portable QRP operation I am planning to use a small semi-automatic antenna tuner
at the feedpoint* of my vertical. Manual tuning takes some time - set carrier, walk over to
the antenna and press a button ;-)

Wonder how Class-E in the QCX series and the Class-D in QDX behave in such
'mis-matched antenna' situation? Are there any limits ?

73's de OE1MWW
Wolfgang

*an antenna tuner belongs to the feedpoint of the antenna and not inside or right behind a rig. 
Otherwise the [cable+antenna] is 'tuned' (matched) = losses could be remarkable high!.


Dan AI6XG
 

http://www.arrl.org/transmission-lines   'Another Look At Reflections' is an interesting discussion of SWR and loss on feedlines.  It made me reconsider how I was approaching feedline loss and matching.


Wolfgang OE1MWW
 

Dan, once again, my question is:

How do Class-D (QDX) and Class-E amplifier (QCX series) react/behave on
mismatched antenna, right before tuning and setting a match on cable end ?

73's de OE1MWW
Wolfgang

P.S. your link can not be opened, because I am not a member of ARRL
But maybe this has more info for you https://ac6la.com/tldetails1.html


Mike Perry, WA4MP
 

Some of the more clever designs for QRP antenna tuners have a tune mode that protects your transmitter by directing most of the RF into a 50-ohm load. Only a small part of the RF is sent to the antenna, so those few out-of-tune seconds do no harm. That eliminates the need for SWR protective circuitry in the transmitter.

—Mike Perry, WA4MP

On Oct 29, 2021, at 8:59 am, Dan AI6XG <phase2682@gmail.com> wrote:

http://www.arrl.org/transmission-lines 'Another Look At Reflections' is an interesting discussion of SWR and loss on feedlines. It made me reconsider how I was approaching feedline loss and matching.


Wolfgang OE1MWW
 

Thanks Mike, but my question is:

How do Class-D (QDX) and Class-E amplifier (QCX series) react/behave on
mismatched antenna, right before tuning and setting a match on cable end ?

73's de OE1MWW
Wolfgang


RHETT ISLEY
 

Wolfgang,

Does your auto tuner have a resistive bridge so the radio never sees more than a 2:1 VSWar during the tuning process?

Many of us use a resonant antenna or a “tuner” with a resistive bridge (turn/operate switch) to protect the PA in the transceiver.

Make sense?

73,
Rhett KB4HG

On Fri, Oct 29, 2021 at 12:52 Wolfgang OE1MWW <oe1mww@...> wrote:
Thanks Mike, but my question is:

How do Class-D (QDX) and Class-E amplifier (QCX series) react/behave on
mismatched antenna, right before tuning and setting a match on cable end ?

73's de OE1MWW
Wolfgang


Robert M0NVQ
 

Hi Wolfgang,

Rhett is spot on.  I use the resistive bridge circuit by the G-QRP Club at this link Resistive SWR Bridge (gqrp.com)
I had most of the bits in my junk box.
Switch position 1 lets you calibrate the moving coil meter.
Switch position 2 lets you test your antenna, subjecting your transmitter to a worst-case SWR of 3:1.  Also, with your antenna in one leg of the bridge, the hf power is split 4 ways, which means that lets say with a transmitter output power of 4 watts, you are only polluting the frequency with 1 watt. It also means you could still call QRL 'Is this frequency in use?', with a reduced risk of damaging your transmitter.
Switch position 3  connects the transceiver directly to the antenna.

It may seem difficult to buy 50 ohm resistors, but two 2 watt 100 ohm resistors in parallel (carbon not wire wound) work fine.  If you also put two 2 watt 100 ohm resistors in parallel in the fourth leg of the bridge in place of your antenna, you will have a 16 watt dummy load.

The resistive bridge should be connected to the transceiver with the shortest coax possible and any long coax run should be from the bridge to the antenna.

I can supply photos of my quickly built resistive bridge if required. 

Best regards,
Robert M0NVQ
     


Ronald Taylor
 

Wolfgang:

"How do Class-D (QDX) and Class-E amplifier (QCX series) react/behave on 
mismatched antenna, right before tuning and setting a match on cable end ?"

These particular amplifier designs do not have the capability to reduce their output in the presence of High SWR. So, to answer your original question, the way they react is to get hot and/or fail rapidly due to high voltages present in that condition. 

73 ... Ron

On Fri, Oct 29, 2021 at 10:29 AM Robert M0NVQ via groups.io <m0nvq=mailbox.org@groups.io> wrote:
Hi Wolfgang,

Rhett is spot on.  I use the resistive bridge circuit by the G-QRP Club at this link Resistive SWR Bridge (gqrp.com)
I had most of the bits in my junk box.
Switch position 1 lets you calibrate the moving coil meter.
Switch position 2 lets you test your antenna, subjecting your transmitter to a worst-case SWR of 3:1.  Also, with your antenna in one leg of the bridge, the hf power is split 4 ways, which means that lets say with a transmitter output power of 4 watts, you are only polluting the frequency with 1 watt. It also means you could still call QRL 'Is this frequency in use?', with a reduced risk of damaging your transmitter.
Switch position 3  connects the transceiver directly to the antenna.

It may seem difficult to buy 50 ohm resistors, but two 2 watt 100 ohm resistors in parallel (carbon not wire wound) work fine.  If you also put two 2 watt 100 ohm resistors in parallel in the fourth leg of the bridge in place of your antenna, you will have a 16 watt dummy load.

The resistive bridge should be connected to the transceiver with the shortest coax possible and any long coax run should be from the bridge to the antenna.

I can supply photos of my quickly built resistive bridge if required. 

Best regards,
Robert M0NVQ
     


Wolfgang OE1MWW
 

Hi all.
interesting discussion - but I never ever asked about hints for a tuner - resistive or not ;-)
OK, FYI, I do own an ATU-100 (Chinese clone), a mAT-10 (similar to the Electraft T1 tuner)
and a small manual MFJ tuner (somewhere in the junk box).

As the title of my first post says: "Mismatched antenna on Class-D 
& Class-E amplifier"
and it said nothing about tuner...and so on...


I wanted to know, if a mismatched PA will blow up in smoke or what will happen?
Found this document: https://ris.utwente.nl/ws/portalfiles/portal/59478591/TMTT_AliGhahremani.pdf
"Load-Mismatch Sensitivity of Class-E Power Amplifiers" (very academic paper) but I am not sure
if this will apply to the RF amplifier of the QCX-series or or QDX

PLEASE - I would like to discuss about mismatched Class-D and Class-E RF amplifier and experiences
someone had with this issue - and no hints and advises about any kind of tuner :-)

73's de OE1MWW
Wolfgang


Steve in Okinawa
 

I also want to read that article by Maxwell but am not an ARRL member. I do use TLDetails, often enough so it's on my taskbar. The big debate over whether reflected power is "lost" in a transmission line seems to be answered by AC6LA himself, as "yes". At least the way TLDetails calculates it, feedline loss is the sum of conduction loss, dielectric loss, and reflected loss, and the latter is usually by far the highest when VSWR is high.


Roelof Bakker
 

Hello Wolfgang,

My QCX+ draws large current peaks when using an automatic antenna tuner running the full
gallon. It tripped the current limit setting of my power supply and I did not like it.

Tuning the antenna was proceeded with my K2....

Maybe running 1W will work, but this tuner needed a bit more juice to function properly.

Eventually I measured the impedance, plugged the figures in an on-line L-network calculator
and build a small L-tuner, using 1 coil and 1 capacitor. Tuner loss was less than with the
automatic tuner, which showed in an increased antenna current.

73,
Roelof, pa0rdt


Gary O'Neil
 

Hi Wolfgang;

     First and foremost... God Bless you for your persistence. :-)




--
72

Gary, N3GO


Gary O'Neil
 


Secondly... Try not to hit send before finishing your response... ugh!   Sorry about that!...  Another simple and innocent question... sidetracked on a tangent.   LOL!

     I think the answer you are seeking is that Class D and E amplifiers are assumed matched by design. It's kind of a risky statement for PA's designed as high efficiency amplifiers (e.g. class D, E, F, ... etc.) to be considered more efficient than class C; especially when designed without categorically declaring that the the efficiency only holds when terminated into a unique and specific load... In our case, and with the QCX, we view this as 50 ohms real. But here's the rub... Analogous to tweaking BPF components for a precise frequency response, the output LPF inductors in the QCX are tweaked for maximum output power into a 50 ohm load. What appears obvious also holds up in practice... change the output load to something different, and power is no longer maximum... and so goes the PA efficiency. When you find a way to fix the load such that it remains constant for all time you can tune the PA for max power/efficiency and never revisit re-tweaking the LPF ever again. :-)

     Beyond that... The QCX-Mini "appears" to be a reasonably robust design... not so much for being efficient... for reasons stated above, but rather for its having a degree of margin sufficient to keep the smoke from escaping the final's. It is not good practice in general to operate into a high VSWR, but with high efficiency designs, Drain/Collector/Plate current is often allowed to exceed maximum dissipation limits, while restricting turn-on time in order to maximize power output, while using stored energy (in the tuned circuits) to source the energy to the antenna while the PA device(s) is turned off... the PA is cooling off rather than dissipating energy during it's turned off time.

      I have subjected my QCX-Mini to some pretty nasty mismatches... to the point of shutting down my current limited (half amp) power supply which crowbars and turns off the QCX. It serendipitously also serves as a test that my 5V regulator fix works as I anticipated. :-) I am also careful  not to exceed 13.8V DC input, and normally run it between 9 to 12.6 Volts... principally out of paranoia over the early release 5V regulator issue. :-)  Coming up on a year of in-service use without any mishaps, and because of the Class-E PA design; I pretty much ignore the output power variation of between 3 and 5+ Watts (less than half an S-unit on the other end) due mostly to supply voltage variation. Interestingly... I also use my QCX-30 on 40 meters as well... where it operates something closer to Class C... a pretty inefficient version of Class-E to be sure. I consistently get 4 to 5 W output on that band where we can likely agree it should appear as a pretty severe load mismatch for a 30 meter rig. :-)  So while I will stop short of offering any guarantees; my experience with the QCX... which I have scrutinized considerably, I am comfortable that Han's products are designed with some of the best price/performance/quality that I've experienced in the industry, and all seem to have a comfortable margin built into them... especially in the context that they are successfully built and deployed by fat fingered curmudgeons like myself.

The net is that a few seconds for an auto-ATU or the slower (manual) one that I use, that are needed to settle in on a sweet spot VSWR for operating, shouldn't cause you any stress or concern, much less any grief.

Hope you find this helpful...  More importantly; I hope it answers your question. :-)



--
72

Gary, N3GO


ajparent1/KB1GMX
 

Wolfgang,

Most switch mode amplifiers (Classes C,D,E) are generally not happy with
excessive SWR and the specific cases are low impedance (higher than spec
power out may result) and high impedance or no load as they the output
voltage can ring up to a high voltage at the Drains.  Class E is more
susceptible that class D.

The problem with autotuner is that during the seek for a solution they will
go though some very poor combinations with some of those basically open
circuit.

Generally you want an antenna that is tuned but only needs "touch up"
to be good at the specific site.  FYI: vertials if not resoanant for the
frequency in use suffer far lower efficiency that a tuner cannot correct
and only masks.

Also with QCX/QDX since the frequency range for expected usage is
very narrow there will be not QSY induced issues.

Allison
-------------------------------
Please reply on list so we can share.
No private email, it goes to a bit bucket due to address harvesting


Steven Dick
 

A Class E amplifier is less forgiving than a class D amplifier because a class E is resonant while a class D is not.  Class D is slightly less efficient than class E, 
However, a class E circuit has higher peak voltages and a shorter duty cycle compared to class D.  Under poor, non-ideal loads, peak voltage can rise outside its maximum rating more easily. Also the QDX has a push-pull output stage.  It has several advantages over single-ended including large cancellation of even harmonics, simplifying the LPF filters and importantly,  lower peak voltages for a given power output and power supply voltage rail.   In fact, for an ideal lossless transformer, a push-pull class D output can provide 4 times the power output for a given supply rail compared to single ended. There is no tuning adjustment as in class E and works over a much wider bandwidth compared to class E, allowing ease of multi-band designs.

-Steve K1RF

------ Original Message ------
From: "ajparent1/KB1GMX" <kb1gmx@...>
Sent: 10/29/2021 6:16:27 PM
Subject: Re: [QRPLabs] Mismatched antenna on Class-D & Class-E amplifier

Wolfgang,

Most switch mode amplifiers (Classes C,D,E) are generally not happy with
excessive SWR and the specific cases are low impedance (higher than spec
power out may result) and high impedance or no load as they the output
voltage can ring up to a high voltage at the Drains.  Class E is more
susceptible that class D.

The problem with autotuner is that during the seek for a solution they will
go though some very poor combinations with some of those basically open
circuit.

Generally you want an antenna that is tuned but only needs "touch up"
to be good at the specific site.  FYI: vertials if not resoanant for the
frequency in use suffer far lower efficiency that a tuner cannot correct
and only masks.

Also with QCX/QDX since the frequency range for expected usage is
very narrow there will be not QSY induced issues.

Allison
-------------------------------
Please reply on list so we can share.
No private email, it goes to a bit bucket due to address harvesting

Virus-free. www.avast.com


Gary O'Neil
 


Back to answering the question "Are there any limits", the answer is yes. Performance degrades beyond its design profile, and if operated outside recommended limits may reach a point of self destruction.

If maximum output power at all times is your operational criteria, you have no choice but to establish and continuously maintain the output at 50 ohms real. ATU's  (automatic or otherwise) do not factor into this scenario... They are off limits beyond their "set and forget" use when the system was first installed.

If longevity is your goal, stay within, and never exceed the specified design specifications for maximum Voltage, Current, VSWR, Duty Cycle, and Environmental conditions.

If a smoke free environment is paramount, stay below the longevity limits above with good margin consistent with your personal risk tolerance and/or apriori knowledge of the consequences you deem acceptable in your application.   

My experience with the QCX-Mini (original) reveals that the 5V regulator failures that were being observed (no longer a factor in the current product version) were induced when input power was removed. The voltage on the input side of the regulator falls off almost instantaneously when power is removed, and reaches zero volts before the output (5V) voltage begins to decay. This produces a reverse (Output to Input) DC bias across the regulator maintained by the storage charge on the 470 uF output filter capacitor. The LDO regulator used in the original release version was insufficiently protected from this operational scenario, and when the radio was operated at input supply voltages above 12V it became vulnerable to an internal reverse current breakdown condition. A fast high current Schottky diode shunting the regulator Output to Input (Anode to output side) has proven to my satisfaction to be sufficient to protect the regulator over it's full operating range.

As Roelof mentions in post 73461, The supply current can and does spike excessively high under poor mismatch conditions, that are generally unavoidable in a less than sophisticated auto-ATU system... of even a manual ATU in the hands of an incessantly careless and inattentive operator such as moi.  :-)   Being very much aware of this flaw in my personna, I include current limiting in my power supply. To the point of that very flaw, I've discovered an interesting self-protecting QCX feature. While using a current limited switching wall wart, I observed that the switcher output drops to zero and shuts down... This is a common failure behavior for low-end switching supplies. The switcher stops when the output is shorted, and no output is produced... thus protecting the wall wart from coughing up smoke. The cool thing about the QCX is when it exceeds the current limit, the input voltage crowbars to zero, the QCX over-current condition forces a power-off/on reset, creating an extended lag time to react and correct the problem before it attempts to transmit again. While this "feature" may sound dull and obvious, I found it a cool way to stress test my voltage regulator fix. Holding the key down with no antenna connected causes the high PA Drain current to spike, cycles the power supply, boots to the previous operating condition, discovers the antenna is disconnected, and the cycle repeats ad infinitum. My previous comment on acceptable risk tolerance and consequences applies here,         

I have tested my QCX-Mini to the full extent of its specifications, but only for the purpose of validating my regulator protection modification, and a qualitative assessment of the heat dissipated by the PA FETs.

I am unable to offer any personal experience or comments on the QDX, but from my experience through several technical exchanges with Hans, he seems to have an unwavering aversion to compromising the quality of his products.     

A bit off topic, but...

As an aside... I also have an original 80 meter version of the QCX,,, which to some may seem a bit more robust in its design. There was no overt attempt at heat sinking the PA FETs in that version, and it performed flawlessly up to and slightly beyond where I did some overly aggressive hacking. One can only perform major mods and hacks and restore a 4" x 4"  PCB a finite number of times before it doesn't quite behave the way it once used to behave. :-(  The cycle of events that led to it's current (and yet remarkably operational) condition began when I thoughtlessly cranked up the power supply to 20V. What was I thinking???   T'was a major overhaul to be sure, but good radio designs don't go down without a good fight. And it's such fun bringing broken toys back to life. :-)
   
--
72

Gary, N3GO


Fred Spinner
 

Great analysis of the AMS1117 issue.  Makes sense. I actually bought some to try to destructively test before completing the Rev 1 mini, but I had a series of disasters in my personal life so the almost finished mini is still sitting in a box at my workbench.  Yes like 11 months later.  Just getting back to being able to have fun of any kind now. 

This fix is applicable to the QDX as well.  The difference at 9v to the 3.3v regulator is still pretty substantial and even with the input surge cap that might be a good idea to use that shottkey.  It certainly can't hurt.  

Personally I would build the QDX for 9V and not 12v with those regulators even with the fixes.  But that's me. 

Fred W0FMS 

On Fri, Oct 29, 2021, 6:49 PM Gary O'Neil <n3go@...> wrote:

Back to answering the question "Are there any limits", the answer is yes. Performance degrades beyond its design profile, and if operated outside recommended limits may reach a point of self destruction.

If maximum output power at all times is your operational criteria, you have no choice but to establish and continuously maintain the output at 50 ohms real. ATU's  (automatic or otherwise) do not factor into this scenario... They are off limits beyond their "set and forget" use when the system was first installed.

If longevity is your goal, stay within, and never exceed the specified design specifications for maximum Voltage, Current, VSWR, Duty Cycle, and Environmental conditions.

If a smoke free environment is paramount, stay below the longevity limits above with good margin consistent with your personal risk tolerance and/or apriori knowledge of the consequences you deem acceptable in your application.   

My experience with the QCX-Mini (original) reveals that the 5V regulator failures that were being observed (no longer a factor in the current product version) were induced when input power was removed. The voltage on the input side of the regulator falls off almost instantaneously when power is removed, and reaches zero volts before the output (5V) voltage begins to decay. This produces a reverse (Output to Input) DC bias across the regulator maintained by the storage charge on the 470 uF output filter capacitor. The LDO regulator used in the original release version was insufficiently protected from this operational scenario, and when the radio was operated at input supply voltages above 12V it became vulnerable to an internal reverse current breakdown condition. A fast high current Schottky diode shunting the regulator Output to Input (Anode to output side) has proven to my satisfaction to be sufficient to protect the regulator over it's full operating range.

As Roelof mentions in post 73461, The supply current can and does spike excessively high under poor mismatch conditions, that are generally unavoidable in a less than sophisticated auto-ATU system... of even a manual ATU in the hands of an incessantly careless and inattentive operator such as moi.  :-)   Being very much aware of this flaw in my personna, I include current limiting in my power supply. To the point of that very flaw, I've discovered an interesting self-protecting QCX feature. While using a current limited switching wall wart, I observed that the switcher output drops to zero and shuts down... This is a common failure behavior for low-end switching supplies. The switcher stops when the output is shorted, and no output is produced... thus protecting the wall wart from coughing up smoke. The cool thing about the QCX is when it exceeds the current limit, the input voltage crowbars to zero, the QCX over-current condition forces a power-off/on reset, creating an extended lag time to react and correct the problem before it attempts to transmit again. While this "feature" may sound dull and obvious, I found it a cool way to stress test my voltage regulator fix. Holding the key down with no antenna connected causes the high PA Drain current to spike, cycles the power supply, boots to the previous operating condition, discovers the antenna is disconnected, and the cycle repeats ad infinitum. My previous comment on acceptable risk tolerance and consequences applies here,         

I have tested my QCX-Mini to the full extent of its specifications, but only for the purpose of validating my regulator protection modification, and a qualitative assessment of the heat dissipated by the PA FETs.

I am unable to offer any personal experience or comments on the QDX, but from my experience through several technical exchanges with Hans, he seems to have an unwavering aversion to compromising the quality of his products.     

A bit off topic, but...

As an aside... I also have an original 80 meter version of the QCX,,, which to some may seem a bit more robust in its design. There was no overt attempt at heat sinking the PA FETs in that version, and it performed flawlessly up to and slightly beyond where I did some overly aggressive hacking. One can only perform major mods and hacks and restore a 4" x 4"  PCB a finite number of times before it doesn't quite behave the way it once used to behave. :-(  The cycle of events that led to it's current (and yet remarkably operational) condition began when I thoughtlessly cranked up the power supply to 20V. What was I thinking???   T'was a major overhaul to be sure, but good radio designs don't go down without a good fight. And it's such fun bringing broken toys back to life. :-)
   
--
72

Gary, N3GO


Hans Summers
 

Hi Fred

Gary and I had a friendly debate about this old AMS1117 issue (again all, please note, this was only an issue on the first batch of QCX-mini, it does not affect batch #2 and #3 QCX-minis including the latest Rev 3 minis, all of which use the indestructible 78M05 voltage regulator). 

According to the AMS1117 datasheet the internal diode is capable of handling surge currents in the range 50-100A and it specifically states that an external diode is not needed except under extreme conditions, the example given is a large value output capacitance in the range 1,000 to 5,000 uF and an instantaneous short at the input (crowbar). Such extreme conditions are not seen in the QCX-mini. 

I found that the 10uF input capacitor protects the AMS1117 against any failure due to intermittent supply such as hot-plugging the barrel connector. When I had the 10uF capacitor in place, despite a lot of effort, I was unable to cause failure. I also tested the AMS1117 at supply voltages far in excess of its rated maximum and had to go almost to 30V to induce failure. 

On the other hand the AMS1117 datasheet claims "protected against short circuit and thermal overloads" but certainly it has been the experience of several people that a momentary short caused failure. Such a momentary short could be caused by hot-plugging the PTT connector for example (the QDX has no external connectors that provide access to the regulated 5V or 3.3V rails so this is not an issue here). 

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

On Sat, Oct 30, 2021 at 7:02 AM Fred Spinner <fred.spinner@...> wrote:
Great analysis of the AMS1117 issue.  Makes sense. I actually bought some to try to destructively test before completing the Rev 1 mini, but I had a series of disasters in my personal life so the almost finished mini is still sitting in a box at my workbench.  Yes like 11 months later.  Just getting back to being able to have fun of any kind now. 

This fix is applicable to the QDX as well.  The difference at 9v to the 3.3v regulator is still pretty substantial and even with the input surge cap that might be a good idea to use that shottkey.  It certainly can't hurt.  

Personally I would build the QDX for 9V and not 12v with those regulators even with the fixes.  But that's me. 

Fred W0FMS 

On Fri, Oct 29, 2021, 6:49 PM Gary O'Neil <n3go@...> wrote:

Back to answering the question "Are there any limits", the answer is yes. Performance degrades beyond its design profile, and if operated outside recommended limits may reach a point of self destruction.

If maximum output power at all times is your operational criteria, you have no choice but to establish and continuously maintain the output at 50 ohms real. ATU's  (automatic or otherwise) do not factor into this scenario... They are off limits beyond their "set and forget" use when the system was first installed.

If longevity is your goal, stay within, and never exceed the specified design specifications for maximum Voltage, Current, VSWR, Duty Cycle, and Environmental conditions.

If a smoke free environment is paramount, stay below the longevity limits above with good margin consistent with your personal risk tolerance and/or apriori knowledge of the consequences you deem acceptable in your application.   

My experience with the QCX-Mini (original) reveals that the 5V regulator failures that were being observed (no longer a factor in the current product version) were induced when input power was removed. The voltage on the input side of the regulator falls off almost instantaneously when power is removed, and reaches zero volts before the output (5V) voltage begins to decay. This produces a reverse (Output to Input) DC bias across the regulator maintained by the storage charge on the 470 uF output filter capacitor. The LDO regulator used in the original release version was insufficiently protected from this operational scenario, and when the radio was operated at input supply voltages above 12V it became vulnerable to an internal reverse current breakdown condition. A fast high current Schottky diode shunting the regulator Output to Input (Anode to output side) has proven to my satisfaction to be sufficient to protect the regulator over it's full operating range.

As Roelof mentions in post 73461, The supply current can and does spike excessively high under poor mismatch conditions, that are generally unavoidable in a less than sophisticated auto-ATU system... of even a manual ATU in the hands of an incessantly careless and inattentive operator such as moi.  :-)   Being very much aware of this flaw in my personna, I include current limiting in my power supply. To the point of that very flaw, I've discovered an interesting self-protecting QCX feature. While using a current limited switching wall wart, I observed that the switcher output drops to zero and shuts down... This is a common failure behavior for low-end switching supplies. The switcher stops when the output is shorted, and no output is produced... thus protecting the wall wart from coughing up smoke. The cool thing about the QCX is when it exceeds the current limit, the input voltage crowbars to zero, the QCX over-current condition forces a power-off/on reset, creating an extended lag time to react and correct the problem before it attempts to transmit again. While this "feature" may sound dull and obvious, I found it a cool way to stress test my voltage regulator fix. Holding the key down with no antenna connected causes the high PA Drain current to spike, cycles the power supply, boots to the previous operating condition, discovers the antenna is disconnected, and the cycle repeats ad infinitum. My previous comment on acceptable risk tolerance and consequences applies here,         

I have tested my QCX-Mini to the full extent of its specifications, but only for the purpose of validating my regulator protection modification, and a qualitative assessment of the heat dissipated by the PA FETs.

I am unable to offer any personal experience or comments on the QDX, but from my experience through several technical exchanges with Hans, he seems to have an unwavering aversion to compromising the quality of his products.     

A bit off topic, but...

As an aside... I also have an original 80 meter version of the QCX,,, which to some may seem a bit more robust in its design. There was no overt attempt at heat sinking the PA FETs in that version, and it performed flawlessly up to and slightly beyond where I did some overly aggressive hacking. One can only perform major mods and hacks and restore a 4" x 4"  PCB a finite number of times before it doesn't quite behave the way it once used to behave. :-(  The cycle of events that led to it's current (and yet remarkably operational) condition began when I thoughtlessly cranked up the power supply to 20V. What was I thinking???   T'was a major overhaul to be sure, but good radio designs don't go down without a good fight. And it's such fun bringing broken toys back to life. :-)
   
--
72

Gary, N3GO


Fred Spinner
 

I certainly don't think the situation would be as bad with the QDX, but I still want to idiot proof the Rev 1 mini I have.   The operator probably will be an idiot in the field. 

Honestly there is probably room in the QDX case to build up a separate regulator board if one wanted to, but at 9V and it being more isolated I think it would be fine as is.  I have a drawer full of Shottkeys so I might as well add them. Maybe the only thing I question there is why the 3.3v regulator isn't feed from the 5V one instead of the 9V input, but maybe you'll tell me that it's because of current draw which is valid. I do want to add reverse polarity protection to the QDX though. 

The Shottkey thing is probably like me adding heat sink goo to the BS170s.  Probably not needed but available and "why not?" 

For the Rev 1 mini I saw where someone put in a heat shrinked 7805 sandwiched in the case.  I probably will do that.  I don't think thermal would be an issue with a full sized 7805.  Then it's idiot proof. 

It got shelved in January at 99% done because of a long unfortunate series of life changing events, but I do want to finish it by the end of the year.  2021 more or less completely sucked for me. 

Fred W0FMS 


On Fri, Oct 29, 2021, 11:46 PM Hans Summers <hans.summers@...> wrote:
Hi Fred

Gary and I had a friendly debate about this old AMS1117 issue (again all, please note, this was only an issue on the first batch of QCX-mini, it does not affect batch #2 and #3 QCX-minis including the latest Rev 3 minis, all of which use the indestructible 78M05 voltage regulator). 

According to the AMS1117 datasheet the internal diode is capable of handling surge currents in the range 50-100A and it specifically states that an external diode is not needed except under extreme conditions, the example given is a large value output capacitance in the range 1,000 to 5,000 uF and an instantaneous short at the input (crowbar). Such extreme conditions are not seen in the QCX-mini. 

I found that the 10uF input capacitor protects the AMS1117 against any failure due to intermittent supply such as hot-plugging the barrel connector. When I had the 10uF capacitor in place, despite a lot of effort, I was unable to cause failure. I also tested the AMS1117 at supply voltages far in excess of its rated maximum and had to go almost to 30V to induce failure. 

On the other hand the AMS1117 datasheet claims "protected against short circuit and thermal overloads" but certainly it has been the experience of several people that a momentary short caused failure. Such a momentary short could be caused by hot-plugging the PTT connector for example (the QDX has no external connectors that provide access to the regulated 5V or 3.3V rails so this is not an issue here). 

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

On Sat, Oct 30, 2021 at 7:02 AM Fred Spinner <fred.spinner@...> wrote:
Great analysis of the AMS1117 issue.  Makes sense. I actually bought some to try to destructively test before completing the Rev 1 mini, but I had a series of disasters in my personal life so the almost finished mini is still sitting in a box at my workbench.  Yes like 11 months later.  Just getting back to being able to have fun of any kind now. 

This fix is applicable to the QDX as well.  The difference at 9v to the 3.3v regulator is still pretty substantial and even with the input surge cap that might be a good idea to use that shottkey.  It certainly can't hurt.  

Personally I would build the QDX for 9V and not 12v with those regulators even with the fixes.  But that's me. 

Fred W0FMS 

On Fri, Oct 29, 2021, 6:49 PM Gary O'Neil <n3go@...> wrote:

Back to answering the question "Are there any limits", the answer is yes. Performance degrades beyond its design profile, and if operated outside recommended limits may reach a point of self destruction.

If maximum output power at all times is your operational criteria, you have no choice but to establish and continuously maintain the output at 50 ohms real. ATU's  (automatic or otherwise) do not factor into this scenario... They are off limits beyond their "set and forget" use when the system was first installed.

If longevity is your goal, stay within, and never exceed the specified design specifications for maximum Voltage, Current, VSWR, Duty Cycle, and Environmental conditions.

If a smoke free environment is paramount, stay below the longevity limits above with good margin consistent with your personal risk tolerance and/or apriori knowledge of the consequences you deem acceptable in your application.   

My experience with the QCX-Mini (original) reveals that the 5V regulator failures that were being observed (no longer a factor in the current product version) were induced when input power was removed. The voltage on the input side of the regulator falls off almost instantaneously when power is removed, and reaches zero volts before the output (5V) voltage begins to decay. This produces a reverse (Output to Input) DC bias across the regulator maintained by the storage charge on the 470 uF output filter capacitor. The LDO regulator used in the original release version was insufficiently protected from this operational scenario, and when the radio was operated at input supply voltages above 12V it became vulnerable to an internal reverse current breakdown condition. A fast high current Schottky diode shunting the regulator Output to Input (Anode to output side) has proven to my satisfaction to be sufficient to protect the regulator over it's full operating range.

As Roelof mentions in post 73461, The supply current can and does spike excessively high under poor mismatch conditions, that are generally unavoidable in a less than sophisticated auto-ATU system... of even a manual ATU in the hands of an incessantly careless and inattentive operator such as moi.  :-)   Being very much aware of this flaw in my personna, I include current limiting in my power supply. To the point of that very flaw, I've discovered an interesting self-protecting QCX feature. While using a current limited switching wall wart, I observed that the switcher output drops to zero and shuts down... This is a common failure behavior for low-end switching supplies. The switcher stops when the output is shorted, and no output is produced... thus protecting the wall wart from coughing up smoke. The cool thing about the QCX is when it exceeds the current limit, the input voltage crowbars to zero, the QCX over-current condition forces a power-off/on reset, creating an extended lag time to react and correct the problem before it attempts to transmit again. While this "feature" may sound dull and obvious, I found it a cool way to stress test my voltage regulator fix. Holding the key down with no antenna connected causes the high PA Drain current to spike, cycles the power supply, boots to the previous operating condition, discovers the antenna is disconnected, and the cycle repeats ad infinitum. My previous comment on acceptable risk tolerance and consequences applies here,         

I have tested my QCX-Mini to the full extent of its specifications, but only for the purpose of validating my regulator protection modification, and a qualitative assessment of the heat dissipated by the PA FETs.

I am unable to offer any personal experience or comments on the QDX, but from my experience through several technical exchanges with Hans, he seems to have an unwavering aversion to compromising the quality of his products.     

A bit off topic, but...

As an aside... I also have an original 80 meter version of the QCX,,, which to some may seem a bit more robust in its design. There was no overt attempt at heat sinking the PA FETs in that version, and it performed flawlessly up to and slightly beyond where I did some overly aggressive hacking. One can only perform major mods and hacks and restore a 4" x 4"  PCB a finite number of times before it doesn't quite behave the way it once used to behave. :-(  The cycle of events that led to it's current (and yet remarkably operational) condition began when I thoughtlessly cranked up the power supply to 20V. What was I thinking???   T'was a major overhaul to be sure, but good radio designs don't go down without a good fight. And it's such fun bringing broken toys back to life. :-)
   
--
72

Gary, N3GO