开云体育

Almost identical results! very encouraging.

SMD Yellow shielded 331 x 2 + 51pf to ground



Toroids 330nf x 2 + 51 pf

--
Raj, vu2zap
Bengaluru, South India.

Sarma, that is the LPF to remove spurs between the IF, so it will

Raj, is - 0.65 the total loss on a 50 ohm load for this filter on SSB? What is the CW loss on various bands?

Sharmaji, this is on a test jig.

At 29-11-18, you wrote:

Guys,

This has nothing to do with the spurs issues. Just a test to see if the SMDs are as good as the
toroids. I am pleased that smd is a good alternative at signal stages.

For a cure all I stand by - change KT1/2/3 to Axicom relays, change L5/7 for good measure and
ad Farhans LPF filter for another good measure.

Raj

At 29-11-18, you wrote:

Raj and group,

Farhan was also suggesting a filter to remove the 12mhz spur in the transmitted output
that resulted from crosstalk from clk0 into clk2 at the si5351.
Should that be included in this recommended cure-all?
I still feel the best way to deal with the 12mhz spur might be a high pass filter on clk2
rather than a tuned (and finicky) notch filter.

Do the new inductors for the 45/90mhz LPF have to be a specific kind of shielded inductor?
Or can we get by with any old 1206 surface mount inductor?

How much better are the results when L5,L7 are replaced with shielded
surface mount inductors?? Are we legal without that change?

At one point, Raj saw better results with the torroids at L1,2,3,4 replaced with?
shielded surface mount inductors.? Is that still the case?

There was some discussion about different coils in the various Axicom relays,
has it been determined that the higher current coils are required
to reduce coupling sufficiently?

Ideally we would have web references to specific easily obtained
parts for the relays, caps, and inductors needed to clean up the uBitx.
Even better, perhaps somebody could provide a kit of appropriate parts.
But I think this should wait till a specific list of recommended mods is nailed down,
and we know that the results are repeatable.?

Would be very cool if there was a cheap way to see this trouble in a stock rig,
and to verify that it works better after the fixes have been applied.?
For example, use a second receiver (a second uBitx?) to check uBitx transmissions
for the 12mhz spur and the?45mhz-DialFreq spur?
I assume the transmitted signal would have to be attenuated by 40dB or so
to prevent damage to the receiver, attenuate by an additional 43dB when
listening for spurs.? And put rig under test plus dummy load in a Faraday cage.
Can the procedure include checks for harmonics, at least on the lower bands?

The previous method of adding a second 45mhz crystal filter is no longer recommended.?
This new 330uH-51pf-330uH low pass filter to knock out the 90mhz harmonic of the 45mhz IF
goes into the same position as that 45mhz crystal filter did, but works better.
It does not attenuate the desired signal so much as the 45mhz crystal filter did.


Sarma asked:
> will the smd inductors in post PA lpf mode, take that power?

Answer: NO

Raj said:
> This has nothing to do with the spurs issues.??
I believe that should read:
? ?This has nothing to do with the harmonic issues.

Leave the post PA LPF's with the stock torroids, the new smd inductors will not take the power.
The harmonics thru the post PA LPF's at L11 through L22 are fixed by just swapping out
the relays at KT1,2,3 with Axicom relays, the stock relays are not very suitable for RF.
There are other methods of fixing the post PA LPF's, these involve adding extra relays,
or?adding a band specific low pass filter external to the rig in line with the antenna coax.

We need to finish this up with a detailed and complete list of recommended mods
so folks can use their uBitx with confidence.? Very few have a spectrum analyzer.

Jerry, KE7ER

Jerry,

At 29/11/2018, you wrote:

Raj and group,

Do the new inductors for the 45/90mhz LPF have to be a specific kind of shielded inductor?
Or can we get by with any old 1206 surface mount inductor?

Any one wont do Jerry! has to be a shielded variety with about 2 ohms DC resistance.

How much better are the results when L5,L7 are replaced with shielded
surface mount inductors? Are we legal without that change?

Changing the relays KT1/2/3 to Axicon (thanks to Mike Doty) results in reduced harmonics
which reduces the signals fed back to the first BiDi and therefore I see less spurs.
You may be legal with only this but every board I have tested behaves differently.

Just changing L5/7 reduces the spurs on all higher than 20M bands.
I also found that Farhan filter also does the same thing. Now this filter is at the output
of the amp.. I'd rather clean the input with L5/7.
Using both mods will or should improve the spurs further.

At one point, Raj saw better results with the torroids at L1,2,3,4 replaced with
shielded surface mount inductors. Is that still the case?

The reason I changed them was that L1 toroid was causing instability if anything
like my finger was near it. With the SMDs - no issues.

This one board is clean and legal. Puts out 10W/40M with 15mV rms at the mic in!
I can push it to 14W. Another stubborn board needs 35mV for the same Po.

There was some discussion about different coils in the various Axicom relays,
has it been determined that the higher current coils are required
to reduce coupling sufficiently?

I think the current probably is not important. Axicom's under contact does
the trick.

Ideally we would have web references to specific easily obtained
parts for the relays, caps, and inductors needed to clean up the uBitx.
Even better, perhaps somebody could provide a kit of appropriate parts.
But I think this should wait till a specific list of recommended mods is nailed down,
and we know that the results are repeatable.

There is a seller for the SMD inductor parts, I requested him to sell a small
number of them uninterested in small sales. He asked me to buy the whole
lot! Like 3000 !!

The 45MHz filter is lossy but better than the LP filter IMHO.

Would be very cool if there was a cheap way to see this trouble in a stock rig,
and to verify that it works better after the fixes have been applied.
For example, use a second receiver (a second uBitx?) to check uBitx transmissions
for the 12mhz spur and the 45mhz-DialFreq spur?
I assume the transmitted signal would have to be attenuated by 40dB or so
to prevent damage to the receiver, attenuate by an additional 43dB when
listening for spurs. And put rig under test plus dummy load in a Faraday cage.
Can the procedure include checks for harmonics, at least on the lower bands?

The previous method of adding a second 45mhz crystal filter is no longer recommended.
This new 330uH-51pf-330uH low pass filter to knock out the 90mhz harmonic of the 45mhz IF
goes into the same position as that 45mhz crystal filter did, but works better.
It does not attenuate the desired signal so much as the 45mhz crystal filter did.


Sarma asked:

will the smd inductors in post PA lpf mode, take that power?

Answer: NO

Raj said:

This has nothing to do with the spurs issues.

I believe that should read:
This has nothing to do with the harmonic issues.

Leave the post PA LPF's with the stock torroids, the new smd inductors will not take the power.
The harmonics thru the post PA LPF's at L11 through L22 are fixed by just swapping out
the relays at KT1,2,3 with Axicom relays, the stock relays are not very suitable for RF.
There are other methods of fixing the post PA LPF's, these involve adding extra relays,
or adding a band specific low pass filter external to the rig in line with the antenna coax.

We need to finish this up with a detailed and complete list of recommended mods
so folks can use their uBitx with confidence. Very few have a spectrum analyzer.

Jerry, KE7ER

This test is NOT on a BITX board, the loss at 45MHz and 90MHz is a points for comparison
of both components. This was done on a test jig I made with 2 BNCs and hooked to a SA.