开云体育

Upon completion of my QMX (#331) build I found that the sensitivity on 20 meters to be quite poor, about -92 dBm (5.6 microvolts) for a Minimum Discernible Signal. A topic soon appeared on the forum regarding this as a design problem so I entertained myself otherwise and waited for someone to find a solution.

Since there has yet to be a solution proffered I decided to try the obvious solution of replacing C403, a 22 pF capacitor in the band pass filter, with a 16 pF value. That seemed to be what a ratiometric guess would suggest. Nope! The band pass response did indeed get closer but the 14 MHz frequencies were still highly attenuated. A 10 pF capacitor did little better but confirmed that the tapped inductor ( L401) and the associated parasitic capacitances formed an effective trap for the entire 20m band.

?

It occurred to me that the simplest solution would be to just bypass the 20m band pass filter entirely. I live in a rural area with no broadcast stations or other strong HF signals. If I needed a band pass filter then I could add one externally or try to wedge in one separate from L401.

?

I removed the tap at the 19^th turn of L401 from the solder connection, soldering the two conductors together so to effectively remove the tap from the coil. I then removed C403. I installed a 1 nF (.001 uF) 0805 sized capacitor from the input signal side of C402 (the end shared with C403, C401, and C404) to the previously vacated solder pad of the removed tap. I positioned it a bit diagonally, reflowed the solder at the C402 pad so that it fastened the new capacitor. The other end was then soldered to the pad where the L401 19^th turn tap used to reside.

?

Progressing to the Step 8 tests, I confirmed continuity from IC3 pin 9 to pins 3, 4, and 5 but not 6. Visual inspection confirmed that no potential short circuits had been created.

?

The results have been satifactory. MDS is now -121 dBm (0.2 uV) which is comparable with the other four bands. No new birdies or images. No overloads. Just a 29 dB improvement in receive level. Terrestrial, solar, and galactic noise are now all stronger than the weakest signal that I can detect. Operation on all other bands were unaffected.

?

If new builders want to duplicate this fix might avoid making the 19^th turn tap on L401 and not install C403. The value of the new capacitor is not critical, it just needs to block the DC bias but pass the 14 MHz receive signal when that band is selected. It could be a leaded part, placed from the hole at the input end of the C403 position to the hole where the 19^th tap was supposed to go.

72, Don

Hi Don
Sounds like a much simpler solution?
When you do an RF sweep of 20m what is the level of the sweep in the 14 mhz range... I think the best I got was -12 db with an external toroid and coil
Dave VE3KCL

'Morning Dave
It shows -7 however I am not sure how that is based. Here's the produced chart:

Below is the original on this unit:


I'm pretty sure that these displays are meant to be just relative measurements since the amplitude parameters seem to change with various settings. The proof was the bench results. Those I can control the conditions. On-the-air performance is now quite nice, I am hearing the weak ones and the normal background noise. The only artifacts present are the ones that I had before the modification but, then again, I am in a rural area where I can expect to be able to sneak by without a receive band-pass filter.
72, Don

Hi Don?
It does look 5db stronger at 20m without having a resonant mechanism in the way.( the additional separate toroid bodge.)
It might be interesting to adjust the sweep starting at 12mhz to stepping at 60000 or something slower to see a bit more detail.
There may be a higher number in there with a narrow range ( or maybe another clue) .
In your situation it should work fine but it may not work for everyone having the bpf wide open like that.
Dave VE3KCL
??

Glad you found a fix.? Unfortunately, for those of us near urban areas this isn't really an option.

I'm probably going to cancel my order until this is resolved.? I already have an MTR3b, X6100, and 2 minis.? This was just a "I want to build another radio" purchase, not a necessity.? :)

Just to be clear: This modification only affects just the one band (20 meters) and has no effect on the other four.
There is still the low pass filter in the receive circuit on 20m so there IS filtering.

With so little room available in the tiny rig it is difficult to squeeze in another band pass filter. I have some surface mount inductors so I teamed up a 1.5 uH 1210 size with an 82 pF capacitor in series for a 14 MHz filter test. As expected, the filter is so broad that it doesn't make a difference. Only 3 dB down at 10 MHz. I could wind some AWG 34 on a T37-6 for a higher Q inductor but where would I put it? Would it make enough difference? I doubt it.
If someone were to experience a need for a receive filter then a multiple section external band pass filter would certainly work nicely. In the meantime, I can listen to shortwave broadcasts and WWV by just tuning there now so I even have some advantages.

It was fun trying to find a workaround. This is what I signed up for. I learned a few things and that, in itself was worth the price of the ticket!
72, Don

Hi Don?
Surprisingly there is room for stacking a 37-2 above L401 and it does seem to work as a temporary measure till a different solution is found.
Dave VE3KCL

Dave,
Nice!! Excellent patch. And it didn't affect the other bands? I wouldn't take the graph levels as gospel. Use the full operational signals as the real test. I am sure that the real peak is not flat like that (it would take at least two sections) so the loss at the real peak may not be anything at all. Impressive!
No downsides, then.
=Don=

Hi Don?
I did have to adjust all the L401 band windings to get them centered some details in this area??/g/QRPLabs/message/103622
Dave VE3KCL

Don,
If it works well, then I think having a wide bandwidth can be a good thing, as you suggested, for listening to commercial broadcasts. Narrow filters can be implemented in software as well. There were some builders here who were removing the bandpass filters on QCXs completely, so I think your experimenting with the QMX is extremely helpful!
--
73, Dan - W2DLC

There is a lot more room than I expected in there. I broke the inductor up into 5 size 25 toroids arranged standing in a radial pattern. I kept the circuit it the same. It was intended to be ?a test to see if moving the fields away from each other helped. I don’t have results yet because I have to debug a different issue created accidentally during the mod and haven’t had time to go back yet.?
--
Colin - K6JTH?

That should read -2 toroids.
--
Colin - K6JTH?

For a QCX without bandpass filters see:

/g/QRPLabs/files/Multiband_QCX_v2_en.pdf
page 3 and figure 4

73 Wolfgang DK4RW

On Wed, Jun 21, 2023 at 10:14 AM, Don, ND6T wrote:

If new builders want to duplicate this fix might avoid making the 19^th turn tap on L401 and not install C403. The value of the new capacitor is not critical, it just needs to block the DC bias but pass the 14 MHz receive signal when that band is selected. It could be a leaded part, placed from the hole at the input end of the C403 position to the hole where the 19^th tap was supposed to go.

Don, why not put the capacitor in place of C403 and short the L401 tap separately? Is something avoided by not using the PCB trace from the capacitor to the inductor tap?

I ask because I'm trying to do something similar on a QDX, and shorted 2 inductor holes and used an 820pf capacitor in a BPF capacitor spot, but my rf sweeps still all peak 1-2 Mhz higher than the target frequency.??

Nate
N8BTR

Nevermind, I should have looked harder at Han's explanation of how the BPF works and is switched, not just the schematic drawing itself.?

-Nate