(tr)uSDX power, efficiency and filter optimisation

[hb9hem]

Regarding *Rxxxx*: Since I suspected that they keep selling faulty kits I used somewhat harsher words in a recent e-mail to them. This is the response:

We have stopped producing & selling the (tr)uSDX products 6 weeks ago (March 11th) since we noticed the power stage efficiency issue we have in our kits/assembled units and since we noticed that we do not have a solution on this. Since 11th of March there was no new order since we were aware of that and discussed with Manuel that on our company private chat with them. What you read on forum (instead of asking us officially) that we are producing and selling kits is not true.

I also asked *Rxxxx* to communicate on the forum directly.

[PD0MVP]

If this is true then why are they just sending a problem kit to me on April 11th.
Since *Rxxxx* is a commercial party, I am curious how they will compensate for this.
If you as a seller know there is a problem and you just send the kit, and without any mention or choice from the buyer this is not correct.

[hb9hem]

I think we all agree there. I just asked Andrei of *Rxxxx* again to come forward in the forum.

[DL2MAN]

This:

We have stopped producing & selling the (tr)uSDX products 6 weeks ago (March 11th) (...)

and this:

If this is true then why are they just sending a problem kit to me on April 11th.

does not exclude each other.
It would be in deed helpful, if *Rxxxx* would post here in public, and clearify what´s going on.
From their Homepage, one could interpret, that they are in deed not taking new orders, but still delivering what has been ordered and paid.

RoWaves.JPG (45.43 KiB) Viewed 13853 times

So maybe -knowing there´s unsolved problems- it would be wise to inform the waiting customers, and give them a choice:
Still receive it or bail out....
EDIT: This is on their Homepage:

RoWaves2.JPG (66.18 KiB) Viewed 13846 times

So they are acting exactly as suggested, just not here in the forum, but on their Homepage
Source:
https://dl2man.de/where-to-buy-trusdx//product/trusdx-5-ba ... assembled/


73 Manuel; DL2MAN

[PD0MVP]

Hi Manuel,

Today I had direct contact with *Rxxxx* via email and Whatsapp. I can return the kit and all costs including the return will be refunded. In any case, this is a neat solution for me. For anyone with a kit that does not work properly, I recommend Email or send a Whatsapp and try to find a solution. I like a bit of hobying with the soldering iron, but replacing capacitors that are smaller than my soldering iron tip isn't going to be it. I may order another kit once the issues are resolved. Too bad it doesn't do what I expected it to do. It is a nice little tranceiver with which I have made a few nice connections despite the low power. Northern Netherlands with 200mW SSB with only a USB power supply connected. And shortened EFHW to only 5 meters high.

[Iw2fpe]

hi
It would be interesting to be able to buy only the RF plate efficient with the chip and correct capacitors soldered, it would certainly be less expensive than buy and replacing only the capacitors.

73
Iw2fpe

[g7jur]

Hi
I have measured the R Shunt track on the Aliexpress (Up Tech Store) version, it is 16.6 MiliOhms.
The *Rxxxx* version is 18.8 MiliOhms.
This was done by passing one amp though the track, and measuring the voltage drop on the track. 10 mV is 10 MiliOhms.
Philip Lock G7JUR.

[ea5anx]

hi
It would be interesting to be able to buy only the RF plate efficient with the chip and correct capacitors soldered, it would certainly be less expensive than buy and replacing only the capacitors.

73
Iw2fpe

I have ordered five RF boards from JLCPCB for this purpose. I will report results.

[Iw2fpe]

Hi

ea5anx, you have ordered PCB with pre soldered chip and capacitors? or only PCB?
just for know

[ea5anx]

Hi.
I ordered rf pcb with all smd parts soldered following Manuel's instructions. Only had to replace a set of capacitors from 0805 to 0603 package form, maintaining type and manufacturer.

[JJ1ILX]

Hi everyone,

I have one of the Uptech Store Assembled kits and have made QSO's on it but thought things were a bit weak on the 40m band. I did some testing with my home station's dipole antenna and the power out on the low end of the band was showing over 7W and 90%+ efficiency, but it drops significantly over 7.1Mhz. External RF meter verified that these are accurate readings from the rig.

I then did some incremental testing with a dummy load and came away with the following data.
Mhz / Watt / Eff
7.005 = 4.6W 85.2%
7.025 = 4.47W 84.5%
7.050= 4.30W 84.1%
7.075= 4.15W 83.2%
7.100= 4.01W 83.5%
7.125= 3.85W 82.3%
7.150= 3.75W 79.5%
7.175= 3.62W 82.1%
7.195= 3.53W 81.9%

It seems that the most efficient and highest output point is just at 6.95Mhz. I am guessing the best step would be to adjust the number of turns of L32. It is interesting and I guess logical that the power drop is actually very linear, 0.15W for every 25khz .

All of the other bands have 5W + and 85% efficiency or better across the entire band.

I am curious if others have a similar experience on the 40M band.

de/JJ1ILX / David

[DL3MHT]

@JJ1ILX: Not helpful, but just my two cents:

I would be more than happy if I would see efficiency in the 80 - 85% range.
My build is more in the 70 - 75% range after tweaking the coils.
That is measured with a current meter and a scope.
I can get the TX power up to 5W in band middle @12.8V (running on 4xLiFePo cells)
despite lower efficiency though.

This is not as bad as the *Rxxxx* kits, but also not great for Class-E from my limited understanding.

73 de Tobias DL3MHT

[John]

Tobias,
You might try playing around with the number and wire gauge of turns on L15. A few more turns on L15 on my RF bd improved my RF output efficiency over 5% on the 20 meter band when running on 12V or 13.8V. I haven't had time to play with the other bands yet.

[VU3DEY]

I made the purchase through PayPal in *Rxxxx*, though it was hard on them but the RF board is faulty and it's difficult for me to repair. I opened a case in PayPal and *Rxxxx* refunded 25€ for the faulty RF board. I think it's a fair deal as i get to keep the faulty one. Will buy a new RF board. 73

[hb9hem]

For the "Up Tech Store" batch 2 kit I can report 75-78% efficiency for 20-60m and 56% for 80m. No tweaking was performed, input power was read off the bench power supply (RX power subtracted), output power was calculated from Vpp (using scope) over 50R load.
For the out-of-the-box good bands output power is 3.7W (40m) to 4.8W (30m).
A software calibration of R_shunt=14 gives +/-3% accurate efficiency and +/-10% accurate power readings from the (tr)uSDX.
Two extra windings on L51 brings the 80m efficiency to 77%.
Details and waveforms are here.

[EB5HDT]

try putting a metal shield between the CPU board and the RF, and a metal shield on the CPU to prevent RF blocking from the RF board. idea of EB5HDT.
73,
EB5HDT

[EB5HDT]

You have to think that the effectiveness of class E is a bit poor, I mean 65 to 73% in class E, comparable with a class AB of a 100-watt icom or yaesu equipment, with between 2 to 4 watts it is already very good it's great for this qrp equipment you can do several KM of distance in qrp, the operation of the transistors in class E is very ineffective compared to a class a or class ab, maybe you have to change the lpf RF output filter circuit that be a little more effective for class E transistors, or the possibility of some modification to change the way transistors work from class E to class ab, risking of course the temperature of the transistors and the output power as an experiment. a personal opinion. idea of EB5HDT.

https://www.electronics-tutorials.ws/am ... asses.html

73,
EBH5HDT

[John]

A class E amplifier properly tuned using good components is inherently more efficient than a class AB amplifier. Obviously these rigs in kit form have not been optimally tuned, nor have the manufactured units, because optimal class E tuning on 5 bands takes time that can't possibly be put into a $100 radio by a manufacturer, and these class E rigs have been built with inexpensive, imperfect parts. They usually do pretty well right out of the box thanks to the good work of the designers. But tuning a class E amplifier for maximum efficiency is a bit like tuning a piano. People need to consider that aspect of this project as simply part of their construction. Frankly, I find that to be the most fun aspect of the construction of a (tr)uSDX radio. I worked pretty hard at tuning my 20 meter band inductors. In fact I even replaced them with some good T50-6 toroids, then spent an evening tuning them and squeezed out about 87% output efficiency, if I recall correctly. I did that much without resorting to the use of any of my high-end lab gear, or even my NanoVNA. I'll work like that on each band until I've squeezed out the highest efficiency I possibly can because minimal battery consumption for this portable and emergency rig is my goal. For others, 65 to 70 percent efficiency may be good enough. That's up to them. These are great little inexpensive experimental radios, and they work quite well. But they aren't thousand dollar commercial rigs, and they will only work as well as we as hams choose to make them work.

[DL2MAN]

You have to think that the effectiveness of class E is a bit poor, (...)
73,
EBH5HDT

I would ask you, to only write about things you know for sure, as people come by here and take statements for granted. I observed, you have replied to a lot of topics here with -what seems to me- just guesswork, like adding a shielding, like alternative OPAmps, Like proposing to do high voltage reprogramming on a rig with unknown status.
While I really appreciate your commitment and attempt to help, I would like to ask you to not just randomly throw in theoretical ideas, as it appears more confusing rather than helpful when it comes to practice.
Do you even have a (tr)uSDX or is it just theory for you at this point ?

73 Manuel; DL2MAN

[Luca66]

Hello everyone, I did a quick check with nanovna on the 20m band. This is the result without touching the windings. I ask you how I can believe this result and if I have to act. Thanks

[DL2MAN]

Hello everyone, I did a quick check with nanovna on the 20m band. (...)

Your Picture shows a Logmag curve of S11 (so reflection only).
What we need to check our Filters is a Logmag S21 Curve (Through).
I´m not even shure what we´re seeing in your picture. Is it Passband or Notch ?
Please clearify, what exactly you wanted to do.

73 Manuel; DL2MAN

[Luca66]

Hello everyone, I did a quick check with nanovna on the 20m band. (...)

Your Picture shows a Logmag curve of S11 (so reflection only).
What we need to check our Filters is a Logmag S21 Curve (Through).
I´m not even shure what we´re seeing in your picture. Is it Passband or Notch ?
Please clearify, what exactly you wanted to do.

73 Manuel; DL2MAN

It is very likely that I have not used the nanovna correctly. I wanted to check the second harmonic as per your instructions and I can't see it. I connected the S11 output, via the sma adapter to the RF and Gnd pins on the RF board. And then I connected the S21 port (ch1) to the SMA outgoing from the RF card. Then I set the stimulus start 10 mhz stop 30 mhz and I activated the logmag function. Where did I go wrong? Thanks for your help.

[DL2MAN]

Ok, it's measuring logmag on ch0, and you want it on ch1.
So you need to find where you assign channel to trace. In case of my nanovna it's easiest to do it by disabling trace 1 and enable trace 2 and set it to logmag. Whatever you do: check if it shows ch1 in the upper left corner. Next to logmag.

73 Manuel DL2MAN

[Luca66]

Ok, it's measuring logmag on ch0, and you want it on ch1.
So you need to find where you assign channel to trace. In case of my nanovna it's easiest to do it by disabling trace 1 and enable trace 2 and set it to logmag. Whatever you do: check if it shows ch1 in the upper left corner. Next to logmag.

73 Manuel DL2MAN

I followed your instructions. Now that's what I wanted. Thanks

[DL2MAN]

Great !
Just for explaination to all the others, that might struggle with the same issue: (NanoVNA Crash Course)
Nano VNA has a Transmitter Port (Port 0 - S11) and a Receiver Port (Port 1 - S21), however, through internal directional coupler, it is also able to Receive back on TX Port (Port 0 - S11) -> Reflection.
What we saw in his fist picture was a Graph of port0 (S11), so NanoVNA was sending on Port 0 and Receiving at the same Port 0, so we got a graph of the Refelctions on the Input of the Filter only.
This kind of (Reflection) measurement is great for testing Antennas.

What we want to see instead is what happens on the other side of the filter, when we inject RF on one side. (Through Measurement) How does our Filter behave ? When I feed it with constant RF level though set frequency range, how are those RF levels affected on the output of my filter.

For this, we connect TX Port (Port 0/S11) to the RF Pin of the Inter PCB Connector and the SMA Jack to RX Port of the Nano VNA (Port1/S21). This is, what Luca did, however, he then printed a curve of Reflections on TX Port (using TX Port as RX Port at the same Time) -> Recognizable by the CH0 Indicator in the upper left corner.

Instead, you need to print a graph of RX Port of Nano VNA, so you see what´s coming from the output side of the filter (Indicated by CH1 Indicator in the upper left corner of the graph).
Filter measurement: Inject on one side, look what makes it to the other Side of the FIlter (S21)
Antenna Measurement: Inject into a Antenna and see what reflects back. (S11)
Makes sense ?

73 Manuel; DL2MAN