|
Curious what others are seeing for their RF sweeps, I'm not seeing anything that resembles what's in the manual.??
Thanks,
Dennis
|
Dennis,
my RF sweeps look reasonable (screenshots below) but the peaks
are a bit low on all bands. The 20m sweep is different, it dives
steeply almost 100dB near 18MHz... I noticed that the vertical
scale differs a lot between the bands. Top-to-bottom difference is
6dB on 80m, 9dB on 40m, 15dB on 30m and 97dB (!) on 20 meters.
I don't know how much the lowpass filters contribute to the high
frequency roll-off but I'd guess that the peaks could be moved by
adjusting the resonant frequency of the RX bandpass filters. They
are series resonant circuits consisting of the tapped L12 and
C28..C31, switched by IC3. Experimenting with L12 to change the
number of turns could eventually result in a mess with destroyed
PCB pads (5 pads to de-solder/re-solder) so I guess it is better
to adjust the values of C28..C31 (one capacitor per band). If the
peak is too low in frequency, add a bigger capacitor in series. If
it's too high, add a smaller capacitor in parallel.
No QSO yet but my TX is working well, 4 to 5 watts into a dummy
load on all bands with slightly less than 9V supply.
73 de SM6LKM
Den 2021-11-05 kl. 22:31, skrev Dennis
Atwood:
toggle quoted message
Show quoted text
Curious what others are seeing for their RF sweeps, I'm not seeing
anything that resembles what's in the manual.??
Thanks,
Dennis
|
Johan, thanks for the screenshots.? I figured out my problem was the switch mode power supply I foolishly thought I could use.??
My graphs look like yours now, same thing with the 20m band.
Cheers,
Dennis
|
When I first tested my QDX it showed traces just like Dennis's
Jagged, no real sign of cancellation.
I removed the trifilar even though I was sure I'd done it correctly, checked the windings and refitted. Got good traces.
I checked again today, bad trace again like Dennis.
I had "Listen to this Device" enabled and instead of a very low smooth noise it sounded "choppy".
I disconnected the QDX and its power and reconnected. Checked the filters again and got good results, monitoring I heard just a smooth low noise and a "chuff chuff" as the test progressed.
The QDX obviously does not initiate correctly every time. If you get bad traces try disconnecting and doing the test again. The Windows "Listen to this Device" helps confirm correct operation.
73 Alan G4ZFQ
|
Hi Alan?
I wonder, is it just the 1.01c firmware that doesn't initiate correctly every time? Or all firmware versions??
73 Hans G0UPL http://qrp-labs.com
-------- Original message -------- From: Alan G4ZFQ <alan4alan@...> Date: Sun, Nov 14, 2021, 10:42 PM To: [email protected]Subject: Re: [QRPLabs] QDX - RF Filter Sweep graphs? When I first tested my QDX it showed traces just like Dennis's
Jagged, no real sign of cancellation.
I removed the trifilar even though I was sure I'd done it correctly, checked the windings and refitted. Got good traces.
I checked again today, bad trace again like Dennis.
I had "Listen to this Device" enabled and instead of a very low smooth noise it sounded "choppy".
I disconnected the QDX and its power and reconnected. Checked the filters again and got good results, monitoring I heard just a smooth low noise and a "chuff chuff" as the test progressed.
The QDX obviously does not initiate correctly every time. If you get bad traces try disconnecting and doing the test again. The Windows "Listen to this Device" helps confirm correct operation.
73 Alan G4ZFQ
|
On 14/11/2021 19:47, Hans Summers wrote: I wonder, is it just the 1.01c firmware that doesn't initiate correctly every time? Or all firmware versions? Hans, The first time, I'm not sure what the version was but was well before 1.01c. But then I can't be sure whether there really was a fault with the trifilar windings. Would I get a jagged trace or a smooth one with zero cancellation? I could go back tomorrow and try an earlier version. 73 Alan G4ZFQ
|
On Sun, Nov 14, 2021 at 01:47 PM, Hans Summers wrote:
I wonder, is it just the 1.01c firmware that doesn't initiate correctly every time? Or all firmware versions??
I had issues with 1.01a as well.? 1.01c seems to be better, although I now know how to get around the issue.? I did not install 1.01b. 73 Evan AC9TU
|
I have seen WSJT-X report clipping input audio (and no sensible reception) when first loaded up (QDX solid led before launching WSJT-X) on all versions of firmware (at least b and c but pretty sure it happened before too). I have a feeling it could be a related issue.
Closing WSJT-X and repowering QDX has fixed the issue each time. My hypothesis is that it happens when I launch WSJT-X almost immediately after the led lights solidly but I haven't had time to confirm this or whether only reloading WSJT-X is sufficient.
Running 9V, haven't looked at the serial traces for optimisation yet. Had plenty of QSOs on all 4 bands within Europe, one transatlantic to USA so QDX seems to be working great (note my antenna is a horrible compromise for now).
Modifications from standard: blew the PA up twice, first time noticed immediately and replacing 4 BS170s resolved the issue. Second time had to replace the 74ACT08 too (used TI SN74ACT08) but my transatlantic contact was post these mishaps. Pretty sure both times was due to tuning errors, I am now using an LER SWR resistor bridge type circuit during tuning and no troubles since.
(MS Windows 10, linear PSU @8.something volts)
Just thought I'd give you my observations in case they help.
73
Jack
|
firmware: 1.01c
OS: Ubuntu 20.04.3
xfrmr: 3:3 9v
power: 2S Li-ion ~8v
deviations:
1) went overboard with the magnet wire allowance and came up short on T2, so the B wire is from the scrap bin. No reason to expect it matters, except I had to scrape it the insulation.
2) no case. Still in my pcb vice
Here are my RF sweeps. Not pretty & not sure what's going on. I'm thinking to remove the first turn from L12 to shift the peaks up, but I don't think that will help the numbers or 20m.
¨CBrent WT4U
|
On 15/11/2021 01:14, Brent WT4U wrote: Here are my RF sweeps. Not pretty & not sure what's going on. I'm thinking to remove the first turn from Brent Do nothing until Hans comments. The scale of each trace is different, look at the figures on the left. The 20m one in particular has a big range. It seems to me these charts do not always work properly. I have seen different traces without making any changes to my QDX. 73 Alan G4ZFQ
|
Hi Brent,
I have quite similar rf filter sweep graphs. I re did L12 and some other toroids risking to damage the pads but almost nothing changed. I was concerned especially about 20M as I have a similar weird graphic on that too. But last weekend I connected QDX to a tuned 20M antenna and did very successful qso's. So I think the rf filter sweep info is not that accurate. Do not do anything until this is cleared out as Alan suggests.
Take a look at QDX issues topic you can see my rf filter graphs and pskreporter screenshots that shows the QDX is working fine with even that wierd rf filter sweer graph.
/g/QRPLabs/topic/86953388#74545
Br.
|
All,
I have found that the RF sweeps need to be connected to a good 50ohm resistive dummy load (shielded is better) to be repeatable (within accuracy limits).? Connecting to an antenna introduces received signals from the antenna which could be different each time you run the sweep.? Leaving it unconnected would cause issues with the lowpass filter that is ahead of the bandpass filter you are sweeping.
In my case, I had RF sweeps like Brent and on inspection of my L12 winding, I found that I had an extra turn in the first 19 turns of L12 which explains the lower peaks on the bandpass filter on my sweep.? Removing the turn moved all up, though 30 meters was still not centered.? The other bands were centered within reason.? My 20meters sweep did not seem to change much and is similar to Brent's before and after the turn removal of L12.
73
Evan
AC9TU
|
My sweeps were done into a little 4-resistor dummy load. They seem consistent enough, but I could try it on a shielded load.
I don't imagine I'd know there was a problem if I didn't have the sweeps. Given how weird the results are and how few things there to go wrong in the simple filter circuit vs how normal the radio acts, I'm wondering if the problem is on the sensing side of things. That definitely seems like a Hans question.
¨CBrent WT4U
|
Brent,
Based on what I see in your RF sweeps, I would check that you do not have an extra turn in the first 19 windings of L12.? Your sweep is consistent with my sweeps before I removed a winding from L12.? NOTE: Verify that you do have an extra turn before you remove one.? Then do not cut the extra wire off until you solder it temporarily to the pad with the excess just above the pad and test the sweeps again.? If that centers your sweeps, then clear out the hole, trim the wire, and solder it in place.? I would only do this if I counted more than 19 turns to the first tap.
73
Evan
AC9TU
|
Thanks Evan. I just counted again for the upteenth time and it's still only 19. Of course, if I'm low, I'm low, and removing one may still be the answer.
73,
¨CBrent WT4U
|
Hi,
regarding L12, the AL tolerance i.e. inductance tolerance (regardless of winding style) is said to be +/-5% on Micrometals website:
although I don't know if the cores supplied in the kits are authentic Micrometals/Amidon cores. Hans is sourcing good quality components so it will probably not be far from this figure.
Considering the large number of turns on L12, adding or removing a turn or two could be worth the effort to put the resonance at mid band.
73 SM6LKM Johan
Brent WT4U wrote:
toggle quoted message
Show quoted text
Thanks Evan. I just counted again for the upteenth time and it's still only 19. Of course, if I'm low, I'm low, and removing one may still be the answer.
73,
¨CBrent WT4U
|
Hi Johan?
Yes we are using authentic micrometals cores. They and the ferrites are both from http://kitsandparts.com a.k.a. The Toroid King.?
73 Hans G0UPL http://qrp-labs.com
-------- Original message -------- From: Johan Bodin <jbodin@...> Date: Thu, Nov 18, 2021, 12:13 AM To: [email protected]Subject: Re: [QRPLabs] QDX - RF Filter Sweep graphs? Hi,
regarding L12, the AL tolerance i.e. inductance tolerance (regardless of
winding style) is said to be +/-5% on Micrometals website:
https://datasheets.micrometals.com/T50-2-DataSheet.pdf
although I don't know if the cores supplied in the kits are authentic
Micrometals/Amidon cores. Hans is sourcing good quality components so it
will probably not be far from this figure.
Considering the large number of turns on L12, adding or removing a turn
or two could be worth the effort to put the resonance at mid band.
73 SM6LKM Johan
Brent WT4U wrote:
> Thanks Evan. I just counted again for the upteenth time and it's still
> only 19. Of course, if I'm low, I'm low, and removing one may still be
> the answer.
> 73,
> ¨CBrent WT4U
|
"Diz" or Dieter, is one of the most reliable OM's I know! I have bought some of the "One-Watter",? he invented and sell. (Now he has the "5-Watter" on sale!)
A very smooth and reliable kit! - With only 1,2W, from DL to KP4 in CW on 30m, in the middle of the night, with this very small super-het TRX! - "Buildt-In-Keyer" is a fortune! It makes living even easier! And "Full-BK"!?
Unbelievibel what that small and easy "kit" can do! Nearly as good, as Hans QCX!
(A few years, before Hans 'invented' the QCX! )
I bought also a WA2WBY-Torioid-Kit, for the best price I could get!?
Shipping to EU/DL is a little expensive, but this is not Diz fault. The toroids I get, w?re absolute perfect for the filter of the WA2WBY 40W- PA.?
If you have any problem, Diz will help you, as soon as possible. (as quick as lightning ! ! ! ) He is a very good guy!?
i never regret to buy my 'kitsandparts' by 'the toroid king' !
73 de Markus ; db9pz?
|
My sweep graphs look almost identical to the ones in the manual. My question;? Is the effort and risk worth it to move the center freq in this case or is the potential improvement small enough to not make the effort worth it?
|
On Thu, Nov 18, 2021, at 06:43 AM, N3MNT wrote:
My sweep graphs look almost identical to the ones in the manual. My question;? Is the effort and risk worth it to move the center freq in this case or is the potential improvement small enough to not make the effort worth it?
I think that depends on how far up frequency you want to operate.? If you are getting good results now then I would not risk a change. I would try one of the other modes like JS8 at a higher frequency in the band to see how that is responding.? Otherwise, if it ain't broke don't fix it. 73 Evan AC9TU
|
Jack G5JDA
Markus
N3MNT