This is an addendum to posts? ?
/g/BITX20/message/73902and? ?
/g/BITX20/message/73909I'm trying to gather all my notes on v3,v4 mods into one place.
Here's an old post from KU4QO that makes clear how important it is to keep mike gain down.
? ??
/g/BITX20/message/66744If you raise the mike gain, the 45mhz IF distorts and starts creating spurs.
Mike found that spurs started getting bad on 10m when mike gain was turned up
such that the output power was only 1/4 Watt.? Not clear where RV1 was set
during his tests.? RV1 might be a better option for adjusting output power.
As mentioned previously, replacing Q20,Q21,Q22 with BFR93W's can help?
reduce the spurs.? An alternate transister that might be easier to obtain is the BFR106.
As mentioned previously, adding an outboard low pass filter is a good solution
for knocking down the harmonics.? This is probably the easiest and most thorough approach.
Harmonics on v3 and v4 boards can be reduced by replacing relays KT1,KT2 and KT3
with Axicom?V23105-A5403-A201.? ? No need to replace the other two relays.
Some forum members were using Axicom?V23105-A5303-A201, which have a 200ma
coil current instead of 400ma.? The lower coil current means more coil windings,
there was some discussion that this might mean more harmonics will get through,
though this was never fully sorted out conclusively.
A third option is to rip up a bunch of traces on the board and use the old relays
as per this wiki entry:??
/g/BITX20/wiki/uBITX-spur-and-harmonicsI assume that's Allison's work, though am not sure.
Jerry
On Sun, Dec 22, 2019 at 09:49 AM, Jerry Gaffke wrote:
On v5 (and v6), the changes Farhan made to reduce the 45mhz-DialFreq spur are as follows:
? ?v5 has a new low pass filter at L31,L32,C205 to block the 90mhz harmonic of the 45mhz IF
? ?R26 is now 220 ohms, to increase current (and reduce distortion on large signals)
? ?Transistors Q20,21,22 are now BFR93W for lower distortion (less harmonics) at 45mhz
I have listed those in what I consider to be their order of importance.
Changing out the transistors is the most difficult, and the other two changes may well be
sufficient to reduce the spur enough to meet FCC regs.?
I suggest using a low temperature alloy such as ChipQuik if removing those transistors:
? ??
You first wick away most of the old solder using solder braid and flux, then add a bit
of the ChipQuik as if it were normal solder.? If that ChipQuik is more than luke warm
(via either a soldering iron or a $20 hot air embossing gun) it will be molten, making
it easy to pluck the transistor off with tweezers.
ChipQuik will also be useful when removing the old relays, if replacing them with Axicom relays.
You must either replace those relays or use an external filter appropriate for each band
to prevent harmonics from going out on the air with a v3 or v4 rig.
Here's the discussion in the forum where Farhan developed those changes:?
? ? ?/g/BITX20/message/61645
Note that there are 70 messages in that thread, and there are other such threads.
That particular message doesn't mention the transistors, but the 2n3904's simply don't have
sufficient Gain-Bandwidth-Product to work well at 45mhz, and so distortion occurs?
which produces that 90mhz second harmonic.? But if the filter of L31,L32,C205 is successful
in preventing the harmonic from reaching the first mixer, that's most of the fix.
For better performance during receive, you might also want to change out Q10,11,12?
to the BFR93W's.??
How swapping out L5,L7 for shielded parts would reduce this particular spur
is a bit of a mystery to me, but could well be the case.? Shielded coils
would prevent high magnetic fields in the nearby final amp from getting back into
the IF amp, so worth doing if you have them on hand.? There were other?
problematic spurs beyond the 45mhz-DialFreq spur that this could cure.
There were a bunch of other changes around the audio amp for v5.??
If you have an LM386 or a non-DX branded TDA2822 working in the audio amp, that should be fine.
The discrete transistors of v4 worked well enough for many, but if you have trouble with distortion
or insufficient power it should be easy enough to patch in an LM386 or TDA2822.
That new 2n7000 is there in front of the audio power amp on v5 to reduce annoying PTT pops
in your received audio when keying the mike, but is not essential.
Lots of changes to the final power amp for transmitting, these bring up the transmit power?
on the higher frequencies.? Without these mods you may only get a watt or two at 30mhz,
while getting perhaps 10 watts on 3.5mhz.? But these mods are hardly worth the trouble.
I would just use an external linear such as the 10W linear from QRP-Labs if you need more power,
though you will need an appropriate attenuator between the uBitx and that amp.
Be careful of increasing mike gain beyond what's available in the design as shipped from hfsignals.
Increasing mike gain to compensate for loss of gain in the transmit power amp at higher frequencies means
larger signals in the IF amps, and thus more distortion in those amps and in the mixers,
and thus far far worse trouble with spurs which you won't be aware of unless you look it over good
with a spectrum analyzer.? The rig can sound fine on the air, and still be putting out significant
out-of-band spurs if the mike gain is too high.
Jerry, KE7ER
