On Tue, Sep 13, 2022 at 10:45 PM, Paul Allison wrote:
I really appreciate the way Fahran designs,? using low cost material and same kind of semiconductors everywhere, saving costs.
Still to finish mine, but when I will seriously rebuilt it, I will change something:
There is a lot of old design copied.? THe whole TX chain is the uBITX V3/4/5/6 repeated.
I have a very long memory.
1) If I use the second part of NE5532 ( I use lower noise LM4862) as low pass filter, I'll save 28 components, of course? adding
the small audio power circuit I prefer. This will mean some? more mA, but not a problem for me: 30% smaller PCB.
The Rs and Cs are going to be there if opamps are used for filters. might make a small difference?
and most modern opamps are low power.
2) Sorry, I do not understand the way? Q11, Q12 and Q13 are used, ( yes, Q11 collector is connected to R45 & R48 and Q12 base too, and R41 to +TX).
Why 2 emitter followers when collector of Q11 is only 330 Ohm? Why R52 in series with low impedance input of Q14?
It seems to me to have little gain and loot to loose.
Q11/12//13 as drawn is wrong..? See ubitx for correct.? However at 7mhz using three transistors to
get 20db of gain is going the hard way.? One 2n3904 can easily do that.
I suppose a single MiniCircuit Mar8 ( or similar) can do everything with a single component? : am I wrong?
Yes, incorrect.? MAR8 requires 50 ohm match and is conditionally stable. Poor choice and high
power consumption.? There are better MMICs for that, but they are power hungry and being UHF
or higher they can oscillate and cause havoc.
Oddly a 2n3904 when used for less than 100mW power level is pretty good gain and cheap.
Also doubling up devices eats up parts and board space and need not be used that way.
3) BFO: usually BFO frequency is 1.5KHz? upper? (USB) or lower (LSB ) than filter central frequency, here 5MHz, but I do not
like setting it desoldering-resoldering capacitors C44 & C45 till I have the right frequency: a SMD capacitive trimmer is better for me.How have you done it?
Generally A small trimmer cap is used maybe a series inductor to hit the corect frequency.
Is 1.5Khz high or low, maybe.? Hand made filters like that tend to be lower in general?
as in 5mhz crystal may have a center frequency of 4.9991mhz like the filter I made
meaning the 1.5KHz high side is really 1.4991.? Since the filter was less than 3khz wide?
I have the gear to sweep mine and it was barely 2.5khz wide with the specified parts.
So in that case .13 to a.4 khz above and below might sound better.? FYI you cannot
sweep the filter with NanoVNA as its resolution bandwidth is too wide.? I use the Rigol
DSA815T, the T is the tracking generator and allows for very precise sweeps.
Reality is most either just do nothing and maybe it sounds ok or they try to get
it by ear.
4) I'm using SMD 2222, so I think they are not the best choice as IRF510 driver,? I will try a single SMD transistor , such as? BFG97, with some
mm of PCB as radiator.
Terrible choice for driver. There are several flavors of SMD 2222 and some are very weak.
The driver needs to put ot .5W considering efficiency tit will dissipate easily 2 to 3x that.
None of them are good for .5W dissipation.? I'd consider not going SMT there.
5) Relays are expensive, require power, and lot of space on PCB.
I'm not shure if the very small PTT on my ultra-professional :_))) Baofeng can last forever: but if I will use some transistor to commute VCC to RX or TX
I can delete RL3 and still have a free commutation on RL2.
Not a fan of relays. They need current on the contacts get intermittent (what is
called contact wetting) and even 1 MA can make a difference over time.
My choice at build time a was diode switching using 1n4148.
Antenna switching was a smaller relay for only that.? DC switching
was a few transistors.??
6) I'm using dual side PCB, with bottom side as GND? only, connected here and there to upper side GND, giving me the best possible GND.
I will build it again on two 50x80mmPCB, with the power RF section and 2 relay and output low pass on a single PCB, so I can experiment
and change it without redoing all the other parts. I will built the PTO on the remaining 20x100mm PCB, so everything will only need a standard
100x100mm PCB, low cost and good quality from China.
Keep in mind the relays have RF (IF switching and Antenna switching) on them so they need to be kept apart or shielded.
7) I really appreciate this way to build PTO:?
But it works the reverse way, clockwise frequency goes down, that is not istintive, and it requires a custom knob; as the usual hole in knobs is 1/4=6mm
diameter, and is very difficult to find bigger .
I consider that one of those machining issues.? How they did it and the net results are alterable
but I thought it very creative use of materials.? Even PTO can have looseness and that makes
for frequency shifts and microphonics.? While simple appearing the details are critical.
Look back at the uBitx and Bitx 20 for history and what works and what needed attention.
?
While daylight has many good ideas there are sections that need review.?
Allison
--------------
