Since you can upload the software, the USB connection is working and I specifically used the same baud rate as the upload to avoid any speed issue on the serial link.?

I re-downloaded the diagnostic software from the files section to make sure, and it works on my unit.

So I can't see where the problem is coming from.

Here are two screenshots of what to expect.

Screenshot one shows the main menu displayed at serial monitor launch, and after sending a "2" for the "Input tests".


Screenshot two shows the display when turning the encoder up and down.

Please note that in the main menu only options 1 and 2 are currently working.

73, John (VK2ETA)

Skip, that may be true.

A little thinking out loud, so to speak.? Stuff running in my head on why and
how of more uniform power.? Or -10dbm(.1mW) in and 40DBM(10W) out or
about 50DB of over all gain?out of 4 stages (an average of about 12.5db)
where the lower level stages are about 16-17db and the finals are closer
to 13-15db (in the circuit). If we assume the last stage as 13db that means
the prior 3 must deliver 37DB and also .5W minimum.? Not a tall order
but to do that from 3 to 30mhz means every part of that must play well.

First item: Mosfets are fairly gain flat despite drive issues.? They do not exhibit FT
that bipolar devices do.? The upper limits are more lead and case parasitics than?
the silicon.? IRF510s put out serious power at 6M with gains in the 13-16db range.
So other than the impedance matching on the input and output we know they can.
Besides every device I looked at at best gave a better IMD for the same power or
cost a lot of money and are not in a package that is uBitx friendly.? So changes
here are not explored.? While RD parts offer a better IMD the cost (5x higher)
for legit parts is out of consideration save for those that have other performance
goals.

What I think is going on why some do better than others compared to the bogey unit:
Adding inductors to reduce the feedback is good but when the transistors are?
at best giving a gain of 10 maybe 11 (10 to 10.4DB) you are running against
possible gain vs desired gain?(desired is about 39 (17db)!).? It only takes a
little bit more from the drivers to make that but with high Ic/Ie and only
300mhz Ft we run to?the corner of diminishing gain to less than enough to
hit the feedback selected value. So that means a? edge group of 3904s are
running higher current which depresses the HFE and also lowering gain
due to FT issues and we end up with 12W at 80m and 1.4W at 10M.

My thinking on what to do and why it may be a solution:
There are two ways to do this. Using 3904s we drop each stage gain down?
10-11db? that flattens the gain from 1-30mhz but we need more stages
and for a new design its reasonable to look at for a existing design?
its unreasonable to add so its out of? contention.? ?The other ways is
use transistors with higher FT and do not suffer diminishing Hfe with
increasing current (or less so).? ?Now we can get he required gain
and feedback limits it as we have?enough (or nearly so) at 30mhz.
A 600mhz part give us a shot and?1ghz parts?can easily do it.?
Those would be MPSH10 or similar and 2N5109 or similar.? For
production this may be too costly but there are other parts that
may be a good choice on a cost vs availability scale.? Oddly I've
looked at a classic the 2n2222A (TO18 metal part) as its been
my go to for years. It has a better HFE curve at higher currents
so it does not suffer that depression.? Its also higher power
(its the same DIE as the 2n2219A)? its FT is still only 300mhz
but experience says they do better so we have a candidate to try
as I have a bin full.? In the mean time an parts order is in order...

Optimizing the output transformer is also a factor but space and?
ease of performing the mod are a consideration.

It may take a lot or very few mods to get there.? Less is better.
I may try several routes and see what implements better, cheaper,?
and using parts that are not unobtainium.

My goal is not so much more power 10 to 12 W is fine but to make 10M near
that as well. Or constantly within 3db or less.? Also an effort to lower the
standing power drain some to make it more battery friendly.

Allison

After I wrote the last email, I realized the cascode configuration is much better.
It needs 2 more FETs but you drive the gate of the first one, and it provides the drive to the second one and the second one's gate is AC ground. I can't do it now, but I will see if I can find time to design one next week.
If you wanted to do this with uBitx, remove the exiting output xformer, add 2? IRF530 and a higher voltage supply, with a new output transformer. Should get up to 50W easily without changing the driver.
Howard
Quoting "Rahul Srivastava via Groups.Io" <vu3wjm=[email protected]>:

Thanks..!!? Alison for the detailed analysis. Just to scratch the itch maybe I will hook up one in mid air with a 25:1 transformer at input and standard bifilar on drain.?
I once made 3x3 pushpull on IRF510 running at 30V for? 27.120Mhz ISM band freq got pretty decent 150W output. Input was on binocular core and output on stacked Mix 6 toroids.?
73
Rahul VU3WJM