Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
Clicky sidetone
|
OK, having checked and corrected all the PA coils I have good TX output power and good RX sensitivity. Onward and upward...
Has anyone got really really silent QSK? I have reduced C22 from 10u to 100n. Replaced R39 with 6.8v zener. Even added 1K2 in series with diode D5 to slow the R->T transition a little. Replaced wimpy wire from PSU to PCB screw terminals with decent wire. But still there is a click on make (I mean key down). If I stop the transmit RF output by going into a menu I can still key the sidetone and the click is still there. If I go to semi-QSK the click goes away. So the click is not caused by RF or full power DC load on the PSU. It is caused by R->T changeover. Any ideas please? Could it be noise on the +5V line, which I just noticed is used to bias IC3A in the audio filter? It's late now, so something to check tomorrow. 73, Alan G3XAQ |
|
...and even weirder, the click is on the leading edge of the second and subsequent R->T transitions in string. The first dot has no click.
toggle quoted message
Show quoted text
73, Alan G3XAQ So the click is not caused by RF or full power DC load on the PSU. It is |
|
Richard Harris
Alan, Therein may lie a clue.? I have looked at the waveform in the semi-breakin mode, my preferred mode. When receiving, the keying transistor TR6 is on but the gates of the MOSFET PAs are at 0V and so no current flows through the PAs.? When the transceiver goes to transmit for the first dot or dash, TR6 is already on and IC3 pin 5 goes high so enabling the PA drive and the first dot or dash.?? At the end of this first dot or dash TR6 turns off and the gate drive remains,? This is the reason why some users have reported hearing a "spacer", which is caused by RF being coupled out via the gate-drain capacitance.? So TR6 only goes from off to on on the second dot or dash and so it may be the off-to-on transition that is causing the problem.? The waveform below shows TR6 collector voltage and the RF output. Others have suggested that the rise and fall times are too fast (and not equal) and I would concur.. I plan to make a change much as was recently described on this forum.? Slowing down the transition may reduce the side tone transient. The rise time of the first dot or dash isn't controlled and so will generate a key-click, but I don't think that is a problem.? It could be cured by a change to the software which controls the KEYOUT line, assuming it doesn't have some other impact.?? Richard G3OTK
On Thursday, 22 March 2018, 08:27:30 GMT, g3xaq <alan@...> wrote:
...and even weirder, the click is on the leading edge of the second and subsequent R->T transitions in string. The first dot has no click. 73, Alan G3XAQ > So the click is not caused by RF or full power DC load on the PSU. It is > caused by R->T changeover. > |
|
Thanks very much for this Sherlock Holmes analysis. I don't get any clicks at all with semi-QSK, only on dot #2 onwards when in full-QSK. Can you tell us your exact symptoms in both semi- and full-QSK please?
toggle quoted message
Show quoted text
I have already "corrected" my envelope rise and fall times by setting R41 to 180 ohms and C31 to 2uF, so this may change what you and I see. I'll try to look at Q6 collector but I only have an analogue scope. Maybe it is time to empty my piggy bank into Rigol's coffers? 73, Alan G3XAQ On Thu, Mar 22, 2018 at 07:29 am, Richard Harris wrote:
|
|
Richard Harris
Alan, I have set full break in .?The sidetone looks reasonable up to IC9 pin 1 i.e. the top of the volume control.?? Q7 shorts out the audio signal, which includes the side tone, so this gets through because Q7 has several ohms resistance. ?With the volume control set to about 20%, the waveform at IC10 pin 1 for a series of dots is: The transient after the side tone burst is probably Q7 turning off.? The sidetone burst is not symmetrical about nominal 0V, which may be another cause of the thump.? Please note that I have changed C22 to 1uF - others have suggested 0.1uF helps, although I haven't tried that value. Richard G3OTK
On Thursday, 22 March 2018, 14:54:25 GMT, g3xaq <alan@...> wrote:
Thanks very much for this Sherlock Holmes analysis. I don't get any clicks at all with semi-QSK, only on dot #2 onwards when in full-QSK. Can you tell us your exact symptoms in both semi- and full-QSK please? I have already "corrected" my envelope rise and fall times by setting R41 to 180 ohms and C31 to 2uF, so this may change what you and I see. I'll try to look at Q6 collector but I only have an analogue scope. Maybe it is time to empty my piggy bank into Rigol's coffers? 73, Alan G3XAQ On Thu, Mar 22, 2018 at 07:29 am, Richard Harris wrote: > >? Alan, > Therein may lie a clue.? I have looked at the waveform in the semi-breakin > mode, my preferred mode. > When receiving, the keying transistor TR6 is on but the gates of the MOSFET > PAs are at 0V and so no current flows through the PAs.? When the transceiver > goes to transmit for the first dot or dash, TR6 is already on and IC3 pin 5 > goes high so enabling the PA drive and the first dot or dash.?? > At the end of this first dot or dash TR6 turns off and the gate drive > remains,? This is the reason why some users have reported hearing a "spacer", > which is caused by RF being coupled out via the gate-drain capacitance.? So > TR6 only goes from off to on on the second dot or dash and so it may be the > off-to-on transition that is causing the problem.? The waveform below shows > TR6 collector voltage and the RF output. > > > > Others have suggested that the rise and fall times are too fast (and not > equal) and I would concur.. I plan to make a change much as was recently > described on this forum.? Slowing down the transition may reduce the side > tone transient. > The rise time of the first dot or dash isn't controlled and so will generate a > key-click, but I don't think that is a problem.? It could be cured by a > change to the software which controls the KEYOUT line, assuming it doesn't > have some other impact.?? > Richard G3OTK > > > >? ? On Thursday, 22 March 2018, 08:27:30 GMT, g3xaq <alan@...> wrote: > >? ...and even weirder, the click is on the leading edge of the second and > subsequent R->T transitions in string. The first dot has no click. > > 73, Alan G3XAQ > > > So the click is not caused by RF or full power DC load on the PSU. It is > > caused by R->T changeover. > > > > > > |
|
Fascinating. Thanks Richard (I really must get a digital storage scope!). I have 0.1uF for C22 so I don't see/hear that thump after the sidetone as the QCX goes to receive; I just hear a soft "squish" noise, which is completely acceptable.
I think I can see a hint of a spike at the beginning of the tone burst IC10 pin1 in your second screenshot. It's definitely not there at IC9. I suspect we need to wait for you to get time to try reducing C22 to see how the signal looks. I've also replaced R39 with a 6V8 zener but I don't think that has made much difference. I had a play with LTspice and even if there is no leading edge spike from the receiver front end just the act of turning on Q7 with +5v via D5/R60/C52 generates a 12mV spike at the wiper of the volume control. It seems to be caused by the FET's internal capacitances charging up via D5 and the two halves of the volume control in parallel. That is why changing the volume setting affects the click level. I don't see that much can be done about it because any attempt to significantly slow the spike will leave the mute gate open too long and allow a cycle or two of sidetone through at massive volume. 73, Alan G3XAQ |
|
OK, I've now fixed the QSK clicks.
All it needs is *both* 10uF electrolytics in the audio amplifier, C21 and C22 changing to something like 470nF. The root of the problem is the way the sidetone volume is controlled. As it says on p120 of the manual the mark/space ratio of the tone is reduced via Menu 4.9 to control the volume. So for any volume setting below 99 the M/S is not 1.000. My setting is 85. This introduces a shift in the average DC level all the way through the AF filter to C21, which therefore charges up. After the tone goes away at the end of a dot C21 discharges through the volume control with a time constant around 50ms. When the next dot begins, C21 has not fully discharged back to its resting voltage so the FET mute switch helps it by shorting out part of the volume control. The result is a big click on the leading edge of the second (and subsequent) dots. The same argument applies with C22. When the mute FET Q7 is turned on with a fast 5V edge from the CPU the Cgd+Coss charges and this causes an additional very fast spike at the volume pot at several MHz. It is rounded by the bandwidth of the opamp AF amplifier but still makes the leading edge of the mute very hard. I fixed it with 10nF from the wiper of the volume pot to deck. There may be some RF pickup around there too, because I have my pot off the PCB. Either way, 10nF across the tracks on the PCB fixes it. The QSK is now totally silent. It is almost like listening to someone else's signal on the band. 73, Alan G3XAQ |
|
Alan,
toggle quoted message
Show quoted text
Is that the only fix needed? -- or, was it in conjunction with some other, previous changes? If so, can you list all the changes needed? Thanks for your work and analysis of the fix! Paul, W9AC -----Original Message-----
From: [email protected] <[email protected]> On Behalf Of g3xaq Sent: Saturday, March 24, 2018 10:51 AM To: [email protected] Subject: Re: [QRPLabs] Clicky sidetone OK, I've now fixed the QSK clicks. All it needs is *both* 10uF electrolytics in the audio amplifier, C21 and C22 changing to something like 470nF. The root of the problem is the way the sidetone volume is controlled. As it says on p120 of the manual the mark/space ratio of the tone is reduced via Menu 4.9 to control the volume. So for any volume setting below 99 the M/S is not 1.000. My setting is 85. This introduces a shift in the average DC level all the way through the AF filter to C21, which therefore charges up. After the tone goes away at the end of a dot C21 discharges through the volume control with a time constant around 50ms. When the next dot begins, C21 has not fully discharged back to its resting voltage so the FET mute switch helps it by shorting out part of the volume control. The result is a big click on the leading edge of the second (and subsequent) dots. The same argument applies with C22. When the mute FET Q7 is turned on with a fast 5V edge from the CPU the Cgd+Coss charges and this causes an additional very fast spike at the volume pot at several MHz. It is rounded by the bandwidth of the opamp AF amplifier but still makes the leading edge of the mute very hard. I fixed it with 10nF from the wiper of the volume pot to deck. There may be some RF pickup around there too, because I have my pot off the PCB. Either way, 10nF across the tracks on the PCB fixes it. The QSK is now totally silent. It is almost like listening to someone else's signal on the band. 73, Alan G3XAQ |
|
Paul,
toggle quoted message
Show quoted text
Just change C21/C22 and add the 10nF to ground at the volume control wiper. I have made other changes: replace R39 with 6.8v zener (had no effect), move AF filter from 700Hz down to 575Hz, correct keying envelope to 4ms rise/fall by changing R41/C31. I suspect the collector PNP keying transistor tends to float above 0v during receive periods, messing up the shaping of the rising edge of the first dot. A few K resistance from collector to ground should fix it but I have yet to try this. Instead I've been enjoying operating. I worked 3C0W on Annabon island yesterday through a big pileup, which put a smile on my face! 73, Alan G3XAQ On Sat, Mar 24, 2018 at 08:07 am, Paul Christensen wrote:
|
|
Jess Gypin
Change C21 and 22 to what value?
toggle quoted message
Show quoted text
Jess On Mar 24, 2018, at 9:21 AM, g3xaq <alan@...> wrote: |
|
Thanks Alan - I have 3 QCXs that I had already replaced C22 with 0.1uF caps, and that had fixed the QSK thumps and clicks on all.?
? But the QSK click returned on the extra 40m QCX that I was going to sell. Replacing C21 with another 0.1uF has now removed the click on that one. I didn¡¯t want to sell it because of the click - now I can. My other 2 still sound great with just the C22 replacement. Go figure. ? 73, Steve N4IAG |
|
Steve,
since you noticed the clicks only on one QCX out of 3 after replacing C22, in this case Alan's explanation might tell the full story . I would not be surprised if this particular 10uF (C21) developped a high leakage current, causing a DC offset at the wiper of R36. Anyway, replacing both, C21 and C22 with low leakage 0.1uF .. 1uF capacitors seems to cure all the click issues. 73, Wolfgang DK4RW |
|
Hi Wolfgang,
? Yes, this makes sense. You guys are a lot smarter than I, and I¡¯m happy that you?share your knowledge with the group! ? It would be nice if Hans addressed this issue with future kits. I doubt that we are the only ones bothered by the QSK clicks and thumps. But the fix is a simple one. ? 73, ? |
|
I have just completed my QCX 40, and have been able to make a few contacts with it.? I had to turn down the volume on the side tone, and when I did I encountered this loud clicking problem.? I too like to work QSK, so do not want to go semi if it can be avoided.? I came to this forum for some research, and, Lo and Behold! I found this very topic of discussion.
My question is: can I parallel another cap with C21 and C22 to get the needed value, or do I need to pull them both and replace them.? To me this is a messy chore as I am not as adroit as I need to be sometimes.? And the final value needs to be 0.1uf, correct? Thanks for help and for all the great info on this forum! Tony, N0BPA |
|
Thank you Robert. I just bought 100 ceramic 0.1 mfd caps Digikey so I¡¯d have extras, so that¡¯s perfect. I think I left my stock at home and I¡¯m on the road again, but I will replace them as soon as I get back.
toggle quoted message
Show quoted text
73, Charles - NK8O On Jun 25, 2018, at 11:38 AM, rentwist@... wrote: |
|
Tony, N0PBA, no, you need to lower the value of C21 and C22 so paralleling additional caps with the originals will not get you there.? I also suspect that DC leakage that are a characteristic of electrolytic capacitors may be playing a part here as well, so the original C21 and C22 capacitors need to go.
73, Robert, WA2T |
Paul Christensen
to navigate to use esc to dismiss