Unstable output power

[S58J]

Hi

Today I have soldered toroids for 20m and 10m (I have Hi Band version RF PCB). Upon testing I noticed that the ouput power is unstable - with key down if is falling rapidly. On 20m it went from some 4W (@ 12V) to some 2.1W in about 10 seconds and then stopped falling. On 10m it went from some 2.7W to 1.2W and then .... puffff. My 86256 said goodbye - I assume. The SWR displayed for both bands was 1.05. I used the values as suggested by Manuel and PA Bias set to 160.

Before replacing 86256 (only 2 spares left) is there an idea where to search for a problem? Maybe arround T1 and T2?

I am looking forward to your thoughts. Kind regards.

Janez, S58J

[s57l]

Maybe for some reason you are far away from properly tuned filter. Did the output fet got warm during transmit? Are you using proper, np0 capacitors? Did you check the efficiency?
Maybe leave the bias lower for the tuning/tryout an than increase it?
73 s57l

[S58J]

Thanks Janko for your reply

The situation here:
- I am using only high quality NP0/C0G capacitors rated for at least 100V (this is my third uSDX).
- As stated I opted for the Hi Band option and currently only 20m (position K5) and 10m (position K1) bands have toroids (for testing reasons). Toroids are T37-6 (yellow) from Reichelt.
- I have to admit that I did not check the efficiency (later... when the FET already died... I found out the proper menu choice).
- I assume the FET got very hot (I could not reach it, but the other side of the PCB was very warm).

I checked all the connections, they are fine. I am missing something obvious...

Janez, S58J

[DL2MAN]

When Output Power changes over time / heat, it´s usually a sign of a bad component somewhere.
Or -as others pointed out- Filters not tuned to high efficiency.

I would suggest to do further tests (with new FET) with either USB Power Supply only or connect (tr)uSDX to 7V through main Power Supply, to reduce Output Power until you fixed it. This way it´s more unlikely, that your FETs burn.

Also check, if Power comes out of the SMA Connector (by external measurement). If something in that path has a bad connection, things go downhill pretty fast.

The fastest way to check what´s going on would be a thermal camera. The component, that heats up most is usually the bad guy

73 Manuel, DL2MAN

[S58J]

Thank you Janko and Manuel

I checked everything, here are the findings (I have lowered PA BIAS to 130):
- All capacitors have correct values and are at the correct positions (yes, some de-soldering was involved to confirm/measure that).
- All connections are ok.
- Super BAD efficiency (25% on 10m, 32% on 20m). Of course efficiency is falling even further with decreasing output power during key-down test.
- The FET (86256) is the only part that gets hot (REALLY HOT!).

Now about toroids: the measured values are way to high (100% to 200%). For example the L11 on 10m with 5 turns (0,4mm wire, 24 AWG) produces 0,27mH and the L12 with 4 turns produces 0,23mH. This means that by reducing one turn on each I'll be still far away from the desired inductance. This leads me to the suspicion about the quality of T37-6 provided by Reichelt.

This weekend I will do the test with T37-2, although it doesn't make sense, because in theory they produce even higher inductance with the same number of turns.

Could it be that bad toroid forces the FET to heat up? However, I used the toroids from the same supplier in my 8-band uSDX and they perform like a charm (together with IRLML2060). On my desk I have three different T37-6 batches from Reichelt. I might go the distance and compare them (by comparing inductances).

Of course there is also an option reverting my (tr)uSDX RF board from 86256 FET to BS170 (or IRLML2060) because they both work nicely and simply using the values (capacitors, toroids) as in the case of 8-band uSDX RF board. Unless there is another, well hidden problem within my current RF board...

Kind regards

Janez "JP", S58J

[DL2MAN]

(...)
Now about toroids: the measured values are way to high (100% to 200%). For example the L11 on 10m with 5 turns (0,4mm wire, 24 AWG) produces 0,27mH and the L12 with 4 turns produces 0,23mH. This means that by reducing one turn on each I'll be still far away from the desired inductance. (...)
Janez "JP", S58J

Hi Janez,

Bingo....
I can actually confirm that, as this also happened to me (while building the Classic Bands Board).
If you look at the theoretical Values:

L11: 0,17uH and L12: 0,12uH....

If I feed that into a Toroid calculators, it should be
between 6 and 7 Turns on T37-2 for L11 (depending on calculator)
between 5 and 6 Turns on T37-2 for L12 (depending on calculator)

In reality, I found those Values too high and way off in Terms of Inductance.
Now if you look at my Table of TESTED classical Bands Values:

I found
5 Turns for L11 and
4 Turns for L12
working for 10m with FDT86256

And that was with Sunnys Toroids.
Amidon / Micrometals can behave completely differently in either direction.
Let alone using different Core Material (like T37-6)....

So:
This is the reason, I give Amount of Turns and the Toroids I use, as measuring Inductance at Audio Frequency Levels, when we´re using that coil at 28MHz can add additional confusion.
But as starting point: Remove turns, until Inductance is close to theory and tune from there.
This will get you closer than any Toroid calculator.

73 Manuel; DL2MAN

Resources:
https://dl2man.de/2-trusdx-assembly/
https://miguelvaca.github.io/vk3cpu/toroid.html
http://toroids.info/T37-2.php

P.S. And to answer your question:
Yes, if the Filters are not tuned right (eG way too much Inductance) thean the FET gets really hot, as it´s a mismatch. Basically the same, as if you´d use an Antenna with really Bad SWR.

[S58J]

Hi Manuel

As usual THANK YOU for your suggestions and prompt replies.

Just one more question: the seller/source of Sunny toroids?

Kind regards

Janez "JP", S58J

[DL2MAN]

Just one more question: the seller/source of Sunny toroids?

Sunny.... They came with the Kits....
Please contact her, as I donßt know her source.

73 Manuel; Dl2MAN