|
I'm interested in trying to push my QCX into being usable on both the 20m and 30m bands, with some obvious reductions in efficiency.? The parts that would affect the performance seem to be the LPF, the BPF, and the number of turns on the T1 transformer.? The LPF will just be what it is, though I might try to use a slightly different configuration to get a steeper attenuation curve above 14 MHz, the BPF I can adjust by switching in some parallel capacitance the way that Steve G4EDG has done (see builders' gallery), but the primary windings on T1 are something I can only speculate about.?
The manual calls for 3 turns each on the primary and secondary T1 windings for 20m, and 4 turns each for 30m.? Considering this is a transformer and that lower frequencies call for even more turns I'm guessing that the extra windings are a matter of efficiency.? If I were to use 4 windings on T1 (not counting the 30 BPF turns) instead of 3, would this adversely impact the performance on 20m?? Conversely, how much efficiency would I drop if I were to try and operate on 30m with 3 windings, as per the 20m configuration?? In either case my guess is not much, but if anyone has any insight or experience trying this then I'd be keen to hear it.?
My kit is still in the mail and I'll be building it almost exclusively for portable SOTA work so having an extra band would really help for those days when the atmospheric conditions are challenging.? Steve has reported good results from his 17m/20m build, but I suspect with propagation the way it is these days I'd get more use out of a 20m/30m split.?
Cheers,
John VA7JBE
|
I'm interested in trying to push my QCX into being usable on both the 20m and 30m bands, with some obvious reductions in efficiency.? The parts that would affect the performance seem to be the LPF, John, Hans makes his RX kits quite narrow compared with other SDRs that use similar circuitry. The Softrock RXTX for example had a version covering 40 - 30 - 20m with only the need to add a LPF for 20m. I do not think you would find a wider BPF in the QCX RX would make any great difference. Nor would the small difference in the windings on T1. The PA is a different thing. It is designed as a single band class E. I'd imagine that the 20m LPF, again tighter than Softrocks, would still be adequate at 30m although that would need testing depending how concerned you are. As for optimising L4 and C30 for two bands I guess reasonable results could be achieved. In view of a few reports of catastrophic failure due to PA breakdown I would suggest isolating IC3 pin 3 with a capacitor. Use a resistor to keep the gates low when not driven. 73 Alan G4ZFQ
|
Hi John, Alan ?
Hans makes his RX kits quite narrow compared with other SDRs that use
similar circuitry.
;-)? ?I think a narrow BPF improves the performance. Even with a strong mixer such as in the QCX (high dynamic range and IP3), it is good to keep strong out of band signals away from the mixer. Having a wide BPF which covers several bands makes the radio cheaper and less complex (fewer filters) but it is a compromise, in my opinion.? ?
As for optimising L4 and C30 for two bands I guess reasonable results
could be achieved.
The Class-E resonance is quite broad. There should not be any major problem, operating the Class E PA stage on two adjacent bands such as 20m and 30m.? ?
In view of a few reports of catastrophic failure due to PA breakdown I
would suggest isolating IC3 pin 3 with a capacitor. Use a resistor to
keep the gates low when not driven.
It won't work... if you capacitatively couple IC3 pin 3 to the PA gate, then the peak-peak squarewave will be 5V BUT, it will be centered on 0V. The resulting peaks at the BS170 gate will be only 2.5V above ground and will not drive the BS170 into full conduction. The power output would be greatly diminished. One of the key criteria for obtaining Class-E operation is to drive the PA device hard off/on as a switch.?
73 Hans G0UPL
|
;-)? ?I think a narrow BPF improves the performance. Hans, Yes, I do not dispute that. But may I suggest the difference would not usually be noticed without careful comparison. I would suggest isolating IC3 pin 3 with a capacitor.
It won't work... if you capacitatively couple IC3 pin 3 to the PA gate, then the peak-peak squarewave will be 5V BUT, it will be centered on 0V.
Oh yes, I see... But it seems there have been a couple of cases where PA device failure has resulted in 12V being sent back through IC3. One of those affected have reported failure of the Si5351 and another has said IC2 required replacing. If that really is the case then I'm wondering how to buffer IC3 to avoid this rare occurrence. 73 Alan G4ZFQ
|
Simple problem, Coupling C then resistor with parallel diode to ground. Fet gate goes to ground with no drive, drive loss is Vf of diode.? DC restoration circuit? used in? AC coupled video amplifiers from 70 years back . ? Alan G8LCO ? Sent from for Windows 10 ?
toggle quoted message
Show quoted text
From: Alan G4ZFQSent: 27 August 2018 08:38 To: [email protected]Subject: Re: [QRPLabs] Pushing the QCX (T1 question) #qcx ? > ;-)? ?I think a narrow BPF improves the performance. ? Hans, ? Yes, I do not dispute that. But may I suggest the difference would not usually be noticed without careful comparison. ? >>> I would suggest isolating IC3 pin 3 with a capacitor. > It won't work... if you capacitatively couple IC3 pin 3 to the PA gate, > then the peak-peak squarewave will be 5V BUT, it will be centered on 0V. ? Oh yes, I see... But it seems there have been a couple of cases where PA device failure has resulted in 12V? being sent back through IC3. One of those affected have reported failure of the Si5351 and another has said IC2 required replacing. If that really is the case then I'm wondering how to buffer IC3 to avoid this rare occurrence. ? 73 Alan G4ZFQ ? ? ?
|
Hi Alan ?
> ;-)? ?I think a narrow BPF improves the performance.
Hans,
Yes, I do not dispute that. But may I suggest the difference would not
usually be noticed without careful comparison.
Yes, I agree. But many people have a noisy receiver and don't ever know it... the mixture of lots of unwanted intermodulation products raising the noise floor sounds the same as any other noise. Unless there are particular strong spurii caused by certain stations' mixing products.?
Remember the QCX constructor who complained his QCX has very low audio, low volume... but, when he tuned to a station it did sound very loud? He was used to a more noisy receiver ;-)? ?Noise that wasn't even really there, just added by the radio itself.?? ? >>> I would suggest isolating IC3 pin 3 with a capacitor.
> It won't work... if you capacitatively couple IC3 pin 3 to the PA gate,
> then the peak-peak squarewave will be 5V BUT, it will be centered on 0V.
Oh yes, I see...
But it seems there have been a couple of cases where PA device failure
has resulted in 12V? being sent back through IC3. One of those affected
have reported failure of the Si5351 and another has said IC2 required
replacing.
If that really is the case then I'm wondering how to buffer IC3 to avoid
this rare occurrence.
Yes, an interesting question. Something between IC3 output and the BS170 gate, which does not alter the sharp edges of the waveform, nor does it reduce the 5V "on" voltage, since that would mean the BS170's not being driven ON so hard, which would reduce the efficiency and power output. It's a tall order...?
73 Hans G0UPL ?
|
RE: QCX PA Failures:-
there comes a time where you cannot protect for every
eventuality?
(If you not already seen it, have a look at the PA circuitry in the
NC2030 design,
I guess something similar was not included for reason of cost, both
component and also PCB? size increase?)
It's an example of when potentially it costs more for the protection
circuit than the likely failure scenario,
In the QCX It's Q1->3, Q6 and possibly IC3, they add up to
something sub $4 to replace??
(Q6 Should never fail spectacularly as often as it's being reported,
it's rated at 2A dissipation.
The QCX on TX consumes circa 500mA TOTAL. Why doesn't a fuse save
it?)
the failure of the si5351 is 1 in 5,500+ QCX's that have being built
occurrence?
The best protection is afforded by the operator him / her self,
perhaps think before pushing a QCX hard..
Remember there is no fancy SWR mismatch or over temp protection on
long duty cycle modes.
"Without any protection circuits the QCX? will try it's very best to
fulfil it's purpose,
right up to the point of failure."
Alan
On 27/08/2018 08:38, Alan G4ZFQ wrote:
toggle quoted message
Show quoted text
;-)? ?I think a narrow BPF improves the
performance.
Hans,
Yes, I do not dispute that. But may I suggest the difference would
not usually be noticed without careful comparison.
I would suggest isolating IC3 pin 3
with a capacitor.
It won't work... if you capacitatively couple IC3 pin 3 to the
PA gate, then the peak-peak squarewave will be 5V BUT, it will
be centered on 0V.
Oh yes, I see...
But it seems there have been a couple of cases where PA device
failure has resulted in 12V? being sent back through IC3. One of
those affected have reported failure of the Si5351 and another has
said IC2 required replacing.
If that really is the case then I'm wondering how to buffer IC3 to
avoid this rare occurrence.
73 Alan G4ZFQ
|
Hi Alan?
Yes, protection circuits would have added cost and complexity. And it is hard to protect against every possible failure mode. In my opinion it wasn't necessary. During all my QCX development, I never managed to break any IC3, Q6 (the earlier MPS2907 0.6A version), or any of the BS170s. Not by mismatched load, nor open load, nor short-circuit, nor continuous duty-cycle. Not even by temporarily increasing the supply voltage to around 20V and getting 10W out. Not a failure on any of the six bands. Not with a detuned Class-E resonant circuit either. If I couldn't cause any failures while I was working on the development... I didn't see the need to unnecessarily increase the cost and complexity if it was so hard for me to break something.
73 Hans G0UPL? ?
On Mon, Aug 27, 2018, 17:27 Alan de G1FXB via Groups.Io <g1fxb= [email protected]> wrote:
RE: QCX PA Failures:-
there comes a time where you cannot protect for every
eventuality?
(If you not already seen it, have a look at the PA circuitry in the
NC2030 design,
I guess something similar was not included for reason of cost, both
component and also PCB? size increase?)
It's an example of when potentially it costs more for the protection
circuit than the likely failure scenario,
In the QCX It's Q1->3, Q6 and possibly IC3, they add up to
something sub $4 to replace??
(Q6 Should never fail spectacularly as often as it's being reported,
it's rated at 2A dissipation.
The QCX on TX consumes circa 500mA TOTAL. Why doesn't a fuse save
it?)
the failure of the si5351 is 1 in 5,500+ QCX's that have being built
occurrence?
The best protection is afforded by the operator him / her self,
perhaps think before pushing a QCX hard..
Remember there is no fancy SWR mismatch or over temp protection on
long duty cycle modes.
"Without any protection circuits the QCX? will try it's very best to
fulfil it's purpose,
right up to the point of failure."
Alan
On 27/08/2018 08:38, Alan G4ZFQ wrote:
;-)? ?I think a narrow BPF improves the
performance.
Hans,
Yes, I do not dispute that. But may I suggest the difference would
not usually be noticed without careful comparison.
I would suggest isolating IC3 pin 3
with a capacitor.
It won't work... if you capacitatively couple IC3 pin 3 to the
PA gate, then the peak-peak squarewave will be 5V BUT, it will
be centered on 0V.
Oh yes, I see...
But it seems there have been a couple of cases where PA device
failure has resulted in 12V? being sent back through IC3. One of
those affected have reported failure of the Si5351 and another has
said IC2 required replacing.
If that really is the case then I'm wondering how to buffer IC3 to
avoid this rare occurrence.
73 Alan G4ZFQ
|
Thanks Hans,
I don't know how some manage it...
10W @ 20V, lets do the maths.?? :-[
But it lived.......
Alan
On 27/08/2018 15:41, Hans Summers
wrote:
toggle quoted message
Show quoted text
Hi Alan?
Yes, protection circuits would have added cost
and complexity. And it is hard to protect against every
possible failure mode. In my opinion it wasn't necessary.
During all my QCX development, I never managed to break any
IC3, Q6 (the earlier MPS2907 0.6A version), or any of the
BS170s. Not by mismatched load, nor open load, nor
short-circuit, nor continuous duty-cycle. Not even by
temporarily increasing the supply voltage to around 20V and
getting 10W out. Not a failure on any of the six bands. Not
with a detuned Class-E resonant circuit either. If I couldn't
cause any failures while I was working on the development... I
didn't see the need to unnecessarily increase the cost and
complexity if it was so hard for me to break something.
73 Hans G0UPL?
?
On Mon, Aug 27, 2018, 17:27 Alan de G1FXB via
Groups.Io <g1fxb= [email protected]> wrote:
RE: QCX PA Failures:-
there comes a time where you cannot protect for every
eventuality?
(If you not already seen it, have a look at the PA circuitry
in the NC2030 design,
I guess something similar was not included for reason of
cost, both component and also PCB? size increase?)
It's an example of when potentially it costs more for the
protection circuit than the likely failure scenario,
In the QCX It's Q1->3, Q6 and possibly IC3, they add up
to something sub $4 to replace??
(Q6 Should never fail spectacularly as often as it's being
reported, it's rated at 2A dissipation.
The QCX on TX consumes circa 500mA TOTAL. Why doesn't a fuse
save it?)
the failure of the si5351 is 1 in 5,500+ QCX's that have
being built occurrence?
The best protection is afforded by the operator him / her
self, perhaps think before pushing a QCX hard..
Remember there is no fancy SWR mismatch or over temp
protection on long duty cycle modes.
"Without any protection circuits the QCX? will try it's very
best to fulfil it's purpose,
right up to the point of failure."
Alan
On
27/08/2018 08:38, Alan G4ZFQ wrote:
;-)? ?I think a narrow BPF
improves the performance.
Hans,
Yes, I do not dispute that. But may I suggest the
difference would not usually be noticed without careful
comparison.
I would suggest isolating IC3
pin 3 with a capacitor.
It won't work... if you capacitatively couple IC3 pin 3
to the PA gate, then the peak-peak squarewave will be 5V
BUT, it will be centered on 0V.
Oh yes, I see...
But it seems there have been a couple of cases where PA
device failure has resulted in 12V? being sent back
through IC3. One of those affected have reported failure
of the Si5351 and another has said IC2 required replacing.
If that really is the case then I'm wondering how to
buffer IC3 to avoid this rare occurrence.
73 Alan G4ZFQ
|
Hi Alan..
Yep... that's why the Power vs Supply Voltage curves in the manual go up to 20V supply which produced 9 or 10W on each of the bands, in my prototype units. NOT recommended. But, nothing died...
73 Hans G0UPL? ?
On Mon, Aug 27, 2018, 17:48 Alan de G1FXB via Groups.Io <g1fxb= [email protected]> wrote:
Thanks Hans,
I don't know how some manage it...
10W @ 20V, lets do the maths.?? :-[
But it lived.......
Alan
On 27/08/2018 15:41, Hans Summers
wrote:
Hi Alan?
Yes, protection circuits would have added cost
and complexity. And it is hard to protect against every
possible failure mode. In my opinion it wasn't necessary.
During all my QCX development, I never managed to break any
IC3, Q6 (the earlier MPS2907 0.6A version), or any of the
BS170s. Not by mismatched load, nor open load, nor
short-circuit, nor continuous duty-cycle. Not even by
temporarily increasing the supply voltage to around 20V and
getting 10W out. Not a failure on any of the six bands. Not
with a detuned Class-E resonant circuit either. If I couldn't
cause any failures while I was working on the development... I
didn't see the need to unnecessarily increase the cost and
complexity if it was so hard for me to break something.
73 Hans G0UPL?
?
On Mon, Aug 27, 2018, 17:27 Alan de G1FXB via
Groups.Io <g1fxb= [email protected]> wrote:
RE: QCX PA Failures:-
there comes a time where you cannot protect for every
eventuality?
(If you not already seen it, have a look at the PA circuitry
in the NC2030 design,
I guess something similar was not included for reason of
cost, both component and also PCB? size increase?)
It's an example of when potentially it costs more for the
protection circuit than the likely failure scenario,
In the QCX It's Q1->3, Q6 and possibly IC3, they add up
to something sub $4 to replace??
(Q6 Should never fail spectacularly as often as it's being
reported, it's rated at 2A dissipation.
The QCX on TX consumes circa 500mA TOTAL. Why doesn't a fuse
save it?)
the failure of the si5351 is 1 in 5,500+ QCX's that have
being built occurrence?
The best protection is afforded by the operator him / her
self, perhaps think before pushing a QCX hard..
Remember there is no fancy SWR mismatch or over temp
protection on long duty cycle modes.
"Without any protection circuits the QCX? will try it's very
best to fulfil it's purpose,
right up to the point of failure."
Alan
On
27/08/2018 08:38, Alan G4ZFQ wrote:
;-)? ?I think a narrow BPF
improves the performance.
Hans,
Yes, I do not dispute that. But may I suggest the
difference would not usually be noticed without careful
comparison.
I would suggest isolating IC3
pin 3 with a capacitor.
It won't work... if you capacitatively couple IC3 pin 3
to the PA gate, then the peak-peak squarewave will be 5V
BUT, it will be centered on 0V.
Oh yes, I see...
But it seems there have been a couple of cases where PA
device failure has resulted in 12V? being sent back
through IC3. One of those affected have reported failure
of the Si5351 and another has said IC2 required replacing.
If that really is the case then I'm wondering how to
buffer IC3 to avoid this rare occurrence.
73 Alan G4ZFQ
|
Hi Hans and Alan,
I manufactured a product for many years. Each time one was returned we tried to get down to the cause. Eventually our product was seen as the most reliable one in the market. Lots of work, but a pretty good payoff.
With a kit it is much harder!!
I think the real question is the one that you have asked. How did our customer get the QCX to fail that way? Is there something simple we can do or add to fix the 'problem'?
My QCXs are both still OK, but I wish that a CAT feature worked.
John
toggle quoted message
Show quoted text
On Mon, 27 Aug 2018, Hans Summers wrote: Hi Alan..
Yep... that's why the Power vs Supply Voltage curves in the manual go up to 20V supply which produced 9 or 10W on each of the bands, in my prototype units. NOT recommended.
But, nothing died...
73 Hans G0UPL?
On Mon, Aug 27, 2018, 17:48 Alan de G1FXB via Groups.Io <g1fxb@...> wrote:
Thanks Hans,
I don't know how some manage it...
10W @ 20V, lets do the maths.?? :-[
But it lived.......
Alan
On 27/08/2018 15:41, Hans Summers wrote:
Hi Alan?
Yes, protection circuits would have added cost and complexity. And it is hard to protect against every possible failure mode. In my opinion it wasn't necessary.
During all my QCX development, I never managed to break any IC3, Q6 (the earlier MPS2907 0.6A version), or any of the BS170s. Not by mismatched load, nor open
load, nor short-circuit, nor continuous duty-cycle. Not even by temporarily increasing the supply voltage to around 20V and getting 10W out. Not a failure on
any of the six bands. Not with a detuned Class-E resonant circuit either. If I couldn't cause any failures while I was working on the development... I didn't
see the need to unnecessarily increase the cost and complexity if it was so hard for me to break something.
73 Hans G0UPL?
On Mon, Aug 27, 2018, 17:27 Alan de G1FXB via Groups.Io <g1fxb@...> wrote:
RE: QCX PA Failures:-
there comes a time where you cannot protect for every eventuality?
(If you not already seen it, have a look at the PA circuitry in the NC2030 design,
I guess something similar was not included for reason of cost, both component and also PCB? size increase?)
It's an example of when potentially it costs more for the protection circuit than the likely failure scenario,
In the QCX It's Q1->3, Q6 and possibly IC3, they add up to something sub $4 to replace??
(Q6 Should never fail spectacularly as often as it's being reported, it's rated at 2A dissipation.
The QCX on TX consumes circa 500mA TOTAL. Why doesn't a fuse save it?)
the failure of the si5351 is 1 in 5,500+ QCX's that have being built occurrence?
The best protection is afforded by the operator him / her self, perhaps think before pushing a QCX hard..
Remember there is no fancy SWR mismatch or over temp protection on long duty cycle modes.
"Without any protection circuits the QCX? will try it's very best to fulfil it's purpose,
right up to the point of failure."
Alan
On 27/08/2018 08:38, Alan G4ZFQ wrote:
;-)? ?I think a narrow BPF improves the performance.
Hans,
Yes, I do not dispute that. But may I suggest the difference would not usually be noticed without careful comparison.
I would suggest isolating IC3 pin 3 with a capacitor.
It won't work... if you capacitatively couple IC3 pin 3 to the PA gate, then the peak-peak squarewave will be 5V BUT, it will be
centered on 0V.
Oh yes, I see...
But it seems there have been a couple of cases where PA device failure has resulted in 12V? being sent back through IC3. One of those affected
have reported failure of the Si5351 and another has said IC2 required replacing.
If that really is the case then I'm wondering how to buffer IC3 to avoid this rare occurrence.
73 Alan G4ZFQ
|
(If you not already seen it, have a look at the PA circuitry in the NC2030 design, Alan, No, I had not. Having looked I'm not sure I want to again:-) the failure of the si5351 is 1 in 5,500+ QCX's that have being built occurrence? And an IC2, although maybe a few more that were trashed without being reported here. But thanks all for making me realise once more that things are never as simple as I imagine. 73 alan G4ZFQ
|
Hi Folks,
Great discussion and some interesting revelations there, thanks for the feedback!? Any thoughts on the primary/secondary windings on T1?
Cheers,
John VA7JBE
|
Alan de G1FXB