ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
esc to dismiss
U3S drift/chirp - again!
|
|
Ian Liston-Smith has shared a?OneDrive?file with you. To view it, click the link below.
|
|
|
|
|
|
|
|
Not sure if the attached image will appear on the list, but it's me on 10m QRSS 200mW (the stray lines are aircraft reflections - please ignore) as grabbed by Chris's station a few miles away. His frequency calibration
is pretty accurate, as can be confirmed by other 10m QRSS stations.
My U3S has the GPS QLG1 module which appears to be working - at least I'm getting the locator on WSPR transmissions and the clock is accurate.
Two problems: Look at the switch-on drift. It drops by about 10Hz in the fist couple of seconds! It's got the separate power regulator for FET PA. A heatsink and ?1 coin glued on to the crystal and still it 'chirps'.
The second problem is the frequency. The U3S readout shows 28.000.770 MHz, but I'm clearly transmitting a full 20Hz low. (Remember, Chris's grabber is not 20 Hz out.)
I've tried all the tricks regarding park frequency (150.000.000)? and the ref frequency has settled to 27.004.555, cal step time 20 240. The rig is in the U3S case on a shelf with a stable temperature and a decent 3A
power supply.
Is there ANYTHING left to try to get the frequency right and reduce the drift? On 10m WSPR the drift is 3 to 4Hz and possibly higher, limiting my readability.
Thanks,
Ian, G4JQT
Two problems: Look at the switch-on drift. It drops by about 10Hz in the fist couple of seconds! It's got the separate power regulator for FET PA. A heatsink and ?1 coin glued on to the crystal and still it 'chirps'.
Ian, Have you tried with NO PA voltage? That looks more like a voltage change than temperature related. It seems to stabilise too quickly. The second problem is the frequency. The U3S readout shows 28.000.770 MHz, but I'm clearly transmitting a full 20Hz low. (Remember, Chris's grabber is not 20 Hz out.)
Is that related? It looks as if it might start correctly but quickly shoots down. I've tried all the tricks regarding park frequency (150.000.000)? and And all possible park frequencies? Once the rapid drop is cured maybe you'll get the heat drift down. 73 Alan G4ZFQ the ref frequency has settled to 27.004.555, cal step time 20 240. The rig is in the U3S case on a shelf with a stable temperature and a decent 3A power supply.
|
Ian, Have you tried with NO PA voltage? Of course I am assuming you can hear the U3S on a local RX rather than a remote report?? 73 Alan G4ZFQ
|
I assume you are using a 'standard' QRP-Labs synth board and not the
OCXO version.
What Park Mode type are you using Ian??
Examples:? In mode 2 the park mode is inactive during that rather
long calibration time you have, there may not be enough time after
calibration ends and the QRSS starts for the park mode to 'warm up'
the synth again.? In mode 4 the park mode is active all the time
(during TX, between transmissions, and during calibration).? Keeping
the Si5351A 'busy' with continuous park mode activity might help
that 'initial hook' of drift.
Another possibility if the QRSS transmission is the first in the
sequence, is to add a 'dummy transmission' as another "mode" before
the QRSS "mode" (e.g. maybe a CW ident?) to take up the initial
drift hook.
No idea why the transmission might appear to be 20Hz off.? The 20
240 Cal Time setting should set the U3S to within a couple of Hz at
28 MHz ... but now the Ref Freq has settled you might try something
a little finer, like 01 240? (and as above, 240 is longer than is
probably necessary).
73, Bob? ZL1RS
|
Ian
Additionally to what others have said...?
On your image it seems that the gap between transmissions is well under one minute. Is this actually two consecutive transmissions or did you switch off and on again or something?
If that is really how short your gap is, it may explain it. It is certainly not enough for the 4 minutes of calibration that you configured. So the calibration will never complete, and therefore will have no effect. This could explain why the frequency is off. Park modes will not have any effect either because there is never any idle time between cycles when the unit is not trying to calibrate.
Excuse me if I misread something and I'm writing nonsense. It's been another long day!
73 Hans G0UPL? ?
On Nov 30, 2017 10:32 PM, "Bob Sutton via Groups.Io" <zl1rs= [email protected]> wrote:
I assume you are using a 'standard' QRP-Labs synth board and not the
OCXO version.
What Park Mode type are you using Ian??
Examples:? In mode 2 the park mode is inactive during that rather
long calibration time you have, there may not be enough time after
calibration ends and the QRSS starts for the park mode to 'warm up'
the synth again.? In mode 4 the park mode is active all the time
(during TX, between transmissions, and during calibration).? Keeping
the Si5351A 'busy' with continuous park mode activity might help
that 'initial hook' of drift.
Another possibility if the QRSS transmission is the first in the
sequence, is to add a 'dummy transmission' as another "mode" before
the QRSS "mode" (e.g. maybe a CW ident?) to take up the initial
drift hook.
No idea why the transmission might appear to be 20Hz off.? The 20
240 Cal Time setting should set the U3S to within a couple of Hz at
28 MHz ... but now the Ref Freq has settled you might try something
a little finer, like 01 240? (and as above, 240 is longer than is
probably necessary).
73, Bob? ZL1RS
|
No idea why the transmission might appear to be 20Hz off. Bob, My interpretation is that at the very start of TX the frequency is not far out. The trace shows a rapid drop maybe 20Hz from the exact frequency then not much variation. That is why I suggested a local RX with no PA voltage to eliminate any possible 5351 voltage drop. Ian does not seem to have one? I've never had any such problems, do you consider the rapid drop to be normal? 73 Alan G4ZFQ The 20 240 Cal Time setting should set the U3S to within a couple of Hz at 28 MHz ... but now the Ref Freq has settled you might try something a little finer, like 01 240? (and as above, 240 is longer than is probably necessary).
|
Alan, I'm waiting to see if there is comment from Ian about Hans
observation regarding the grabber screen, that potentially shows
insufficient time between the QRSS transmissions for the calibration
to work and also means the park mode would essentially be off when
it is needed for a minute or two before a transmission starts.? That
would shed more light on both the frequency accuracy and the
functionality of the park mode.? However, it is quite possible that
Ian does have sag on the synth IC supply despite using "decent 3A
power supply" and a separate regulator on the FET PA.
I haven't seen the rapid drop at start, but I've never had a U3S on
28 MHz and (usually) have the settings for Frame Start, Cal Time,
and Park mode operating in harmony.? The only time I ventured above
14 MHz with a U3S was to try one as a replacement for my 6m Beacon,
but I used the OCXO kit for the synth on that.
73, Bob? ZL1RS
On 1/12/2017 8:17 PM, Alan G4ZFQ wrote:
toggle quoted message
Show quoted text
No idea why the transmission might appear
to be 20Hz off.?
Bob,
My interpretation is that at the very start of TX the frequency is
not far out.
The trace shows a rapid drop maybe 20Hz from the exact frequency
then not much variation.
That is why I suggested a local RX with no PA voltage to eliminate
any possible 5351 voltage drop.
Ian does not seem to have one?
I've never had any such problems, do you consider the rapid drop
to be normal?
73 Alan G4ZFQ
The 20 240
Cal Time setting should set the U3S to
within a couple of Hz at 28 MHz ... but now the Ref Freq has
settled you might try something a little finer, like 01 240?
(and as above, 240 is longer than is probably necessary).
|
Thanks all for your suggestions.
Disconnection the power from the PA as suggested by Alan made do difference to the 28 MHz switch-on drift. Though interestingly my signal was still clearly visible on Chris's grabber a few miles away.
I have now set QRSS FSKCW6 (running only my c/s) to repeat every 16 minutes over night on park mode 4 to see what happens (operating on what the U3S readout says is 10.140.000 MHz) for calibration by the various grabbers tuned to 30m. If all goes well I'll
be back on 28 MHz to see what the frequency and drift are like there via Chris's grabber again.
Cheers,
Ian
|
Right, had my U3S have 6 minutes calibration between transmissions over night. Still the frequency is 20Hz low when frequency display is reading 28.000.890 and still has the 20Hz power-on 'tail'. Chris's grabber in not 20 Hz out.
I'm running out of options here. Do I assume there is something really wrong and chuck in the towel for 10m QRSS? I don't think there's anything left to adjust.
Is it worth getting the?OCXO/Si5351A synthesizer module kit? Is that still frequency stabilised by the
QLG1 GPS receiver module kit or is that not necessary?
Was having a lot of fun with the U3S, but maybe mine is starting to fail...
73,
Ian, G4JQT
|
|
|
Thanks for the replies Bob and Daniel.
I've changed the step time as suggested by Bob and it has significantly reduced the 'tail' on 28MHz QRSS transmissions. You both gave other useful info for me to consider.
Thanks again,
73,
Ian, G4JQT
toggle quoted message
Show quoted text
Hi Ian,
I built a 10m LPF kit for my U3S yesterday and ran it up on 28.000.089 FSKCW QRSS.??? With long calibration parameters, a similar 'power-on tail' and almost the same -20 Hz frequency error occurred, along with slow drift during the whole transmission just as
the grabber screen shot shows on your U3S.??? It seems that a long calibration time (like 20 240 as in your first posting) causes the standard synth module to cool down too much during the calibration period resulting in the calibration inaccuracy and then
frequency drift when it starts up/warms up again.??? Far better results were obtained with short calibration parameters like "Step Time" = 20 020 (along with Park Mode parameters = 2 150.000.000 to keep the Si5351a 'busy and warm' until the start of the next
transmission) ... so, despite what the U3S operating instructions say regarding calibration time, at 28 MHz it seems that "less is more"!??? After a couple of transmissions with those settings my U3S had settled to almost drift free FSKCW at .089 (see below).?????????
Perhaps try a "Step Time" = 10 020 which should provide a calibration accuracy of 1 Hz with up to 20Hz of frequency correction at 28 MHz which should be enough to keep a U3S on the display frequency in a stable temperature environment.
FWIW, one of my standard Si5351a synth boards had a bad reference xtal which caused 'jittery/wobbly' FSK tones and a calibration error no matter what the cal settings.????????? I also have a QRP-Labs OCXO/Si5351a module which performed extremely well keeping
the U3S right on frequency with no drift regards of the calibration time length.?????? While it can be done manually, the QLG-1 GPS makes initial U3S calibration easier with the OCXO synth, but after the Ref Frq parameters in the U3S are set to match the particular
OCXO, it should stay on frequency.??? In the end, the GPS is useful to keep the U3S clock accurate, but perhaps that is not so important for QRSS?
73, Bob??? ZL1RS
On 5/12/2017 5:43 AM, ian liston-smith wrote:
Right, had my U3S have 6 minutes calibration between transmissions over night. Still the frequency is 20Hz low when frequency display is reading 28.000.890 and still has the 20Hz power-on 'tail'. Chris's grabber in not 20 Hz out.
I'm running out of options here. Do I assume there is something really wrong and chuck in the towel for 10m QRSS? I don't think there's anything left to adjust.
Is it worth getting the???OCXO/Si5351A synthesizer module kit? Is that still frequency stabilised by the
QLG1 GPS receiver module kit or is that not necessary?
Was having a lot of fun with the U3S, but maybe mine is starting to fail...
73,
Ian, G4JQT
|
Ian,
I did some measurements on my U3S with OCXO tonight. Calibrated a warm TS590S (TCXO) on air 5MHz and to my rubidium 10MHz to within 1Hz in WSJT-X. Condx bad for 30/20m beacons tonight.
Listened with the 590 (small antenna on the bench) on the 27MHz crystal and found it to be slightly off for some reason, the U3S repeatedly calibrated to 26 998 040 and the 590/wsjtx said 26 998 048 (0.3ppm). This is about the same 3Hz someone reported for the GPS ProgRock at 10MHz, IIRC it was 2Hz low ? Anyways, this might be within drift of all components here. Started some WSPR2 transmission on 6m without LPF and into a dummy load, 60sec calibration and 60s idle (park mode 3). The 27MHz frequency variation was less than 1Hz, a fast jump up some 0.5Hz when the WSPR started and peaking in the end of the transmission. Slowly going down in the calibration (almost flattened out) and then lowered some 0.1Hz more when idling and returning to 0.2Hz above the last idling frequency. Repeated cycles made the frequency return to almost the same as the previous.
|
Daniel Ekman SA2KNG