(tr)uSDX - Alternative Band Setups available

[dl6sez]

Hello Giuseppe,

you can not buy a 8-Band board for (tr)uSDX!
8-Band was and is available only for uSDX without the SWR and Power measurments on it.
Different Project with different Software. Was the preceder of (tr)uSDX.
I have both build and yes also uSDX works great

73 de Chris DL6SEZ

[ik6den]

OK Chris mny thaks for info , today is arrived the 5band transceiver kit, nex week assembling and testing my equipment
73' ik6den

[wa2t]

hello, where can I buy 8-Band RF Board?
tks ik6den

That would be a uSDX (octoband?) variant and it is not covered in this forum.

73,

Robert, WA2T

[DL2RMM]

After doing a lot of tweaking, here are my results for the highband option with FDT86256 as Q5:

Power(W) Eff
10m (28,500) 3,5 70%
12m (24,950) 4,8 72%
15m (21,250) 4,7 54%
17m (18,150) 4,0 71%
20m (14,200) 5,0 80%

- all C's = 0805 C0G 100V
- all L's on T37-2 Amidon cores
- supply voltage 12V
- T2 changed to 3T/21T
- with Li-Ion battery pack (4S2P = 14.8V) output on all 5 upper bands is between 5 and 6 watts

These are my C&L numbers:

10M Band
C10//C11 100//27
C12//C13 330//22
C14//C15 220//68
C18//C19 220//68
C16//C17 82//12
L11 5 turns
L12 4 turns

12M Band
C20//C21 120//22
C22//C23 220//180
C24//C25 220//100
C28//C29 220//100
C26//C27 82//27
L21 5 turns
L22 4 turns

15M Band
C30//C31 100//68
C32//C33 470
C34//C35 330//47
C38//C39 330//47
C36//C37 100//27
L31 6 turns
L32 4 turns

17M Band
C40//C41 100//100
C42//C43 330//220
C44//C45 220//220
C48//C49 220//220
C46//C47 100//47
L41 7 turns
L42 5 turns

20M Band
C50//C51 100//150
C52//C53 470//220//12
C54//C55 470//82//10
C58//C59 470//82//10
C56//C57 180//10
L51 7 turns
L52 6 turns

I also experimented with the L15 drain inductor (19T=>22T=>25T) but no significant changes to output power or efficiencies observed.
Two issues left:
a) 15m efficiency not good enough ….might need to experiment with different core type (T37-6 instead of T37-2)
b) hum during RX from OLED getting stronger on upper bands (12m and 10m worst) despite charge pump mod. The hum changes with the letter count on the OLED display. So, display full of letters = strong hum and display with less letters (i.e. menu 1.8 AGC) almost no hum. I added a laminated aluminum foil as a "screen" between main board and RF-board without any improvement. Does anybody have a suggestion/cure for the OLED hum on 10 and 12meters ?

Anyway, I checked into the ANZA DX net this morning and worked a few VK's with the (tr)uSDX ....this is a fun project .... kudos to Manuel and Guido.

73 de Jens, DL2RMM
https://www.qrz.com/db/DL2RMM

[JJ1ILX]

Thanks for the details. I am working on making an updated Schematic for the RF board for the Classic bands based on your calculations and will also make an updated BOM file. I am planning to order 5 from JLCPCB for testing. I will share those files once I get them updated over the next few days and maybe someone would volunteer to check my work for typos / wrong input.

-David (Sorry for the late intro just realized on my callsign is up there) / JJ1ILX

Hi (your name here!),

Happy to share!

Yes, same PA FET's

Yes, 40/20 I just moved caps and toroids to new positions for the Classic band plan, no values changes needed other than the toroid turns changes I made for better Pout, see my subsequent post for that information.

Yes, I have made some efficiency measurements. I have not made harmonic suppression measurements (but easily could), I have adjusted/verified the tuning of the LX2 notch for Fc * 2 on each band.

There is some uncertainty about efficiency measurements, at least in my mind, as the changes I came up with to the components between the SWR bridge transformers and the ATMEGA (for the high power, 9.99 SWR readings issue) changes the (tr)uSDX measured efficiency by 10-11%, for the worse.

73,

Robert, WA2T

[DL2MAN]

After doing a lot of tweaking, here are my results for the highband option with FDT86256 as Q5:

(...)
73 de Jens, DL2RMM
https://www.qrz.com/db/DL2RMM

Hello Jens, would you allow me, putting your Values in a Table and publish them on my website ? (with your Call in it of course)
P.S. I realized drop of Efficiency above 20m as well in my "Classical Bands" Setup. So maybe it would be wiser, using yellow Toroids with modified Turns instead. I have yet to try this....

EDIT: I just realized, you were using significantly different amount of turns in comparison 20/15/10 of my classical bands. What Toroids were you using ? In case of 20m you even used 3 parallel caps.... Maybe that´s not good to publish yet and just leave it here in the forum for now.

73 Manuel; DL2MAN

[DH2LAB]

- all L's on T37-2 Amidon cores

This is possibly due to the very high loss resistance of the Amidon cores.
I roughly compared the cores with the NanoVNA and the cores that comes with the chin. kit, have a loss resistance that is 7-8 times smaller than that of the Amidon T37-2.

However, as I said, it is only an inaccurate measurement with the NanoVNA, but it may be an indication.

[DL2MAN]

I´ve updated my Classic Band Table. I did not change any Values (exept for Typo -initially suggesting to use wrong Toroid Type on 15m), but I changed the measurement results, as I could achieve 80% and more on all Bands by Setting it to PA Bias Max 160. The FDT86256 FET has a higher gate threshold and requires more Bias.
The tested Board also has the T2 Modification (3:21)

Back.JPG (106.73 KiB) Viewed 4490 times

Front.JPG (74.57 KiB) Viewed 4490 times

73 Manuel; DL2MAN

[AC9YY]

I´ve updated my Classic Band Table. I did not change any Values (exept for Typo -initially suggesting to use wrong Toroid Type on 15m), but I changed the measurement results, as I could achieve 80% and more on all Bands by Setting it to PA Bias Max 160. The FDT86256 FET has a higher gate threshold and requires more Bias.
The tested Board also has the T2 Modification (3:21)



Back.JPG

Front.JPG

73 Manuel; DL2MAN

If someone is going to produce a Classic Band or High Band board I would order 1 or 2 of them, especially with the FDT86256 FET.

[Murphyv310]

I´ve updated my Classic Band Table. I did not change any Values (exept for Typo -initially suggesting to use wrong Toroid Type on 15m), but I changed the measurement results, as I could achieve 80% and more on all Bands by Setting it to PA Bias Max 160. The FDT86256 FET has a higher gate threshold and requires more Bias.
The tested Board also has the T2 Modification (3:21)



Back.JPG

Front.JPG

73 Manuel; DL2MAN

If someone is going to produce a Classic Band or High Band board I would order 1 or 2 of them, especially with the FDT86256 FET.

Yes I'd definitely do the same.
I'm somewhat struggling with the cap changes due to eyesight issues, a partially made RF board for the Classic bands would be ideal.

I also would like to know the best supplier of these capacitors please.

[DL2RMM]

I´ve updated my Classic Band Table. I did not change any Values (exept for Typo -initially suggesting to use wrong Toroid Type on 15m), but I changed the measurement results, as I could achieve 80% and more on all Bands by Setting it to PA Bias Max 160. The FDT86256 FET has a higher gate threshold and requires more Bias.
The tested Board also has the T2 Modification (3:21)



73 Manuel; DL2MAN

Thanks for the new input Manuel.
After spending a few more hours tweaking and changing some of the toroids I'm now getting the following results at my highband version:

Power(W) Eff
10m 5,3 75%
12m 4,3 77%
15m 5,2 68%
17m 4,6 79%
20m 5,6 81%

With the following settings:
- Bias for FDT86256 set to 160
- T2 changed to 3:21
- all L's on T37-2 Amidon cores, except L11 (10m) and L31 (15m) they are now T37-6
- supply voltage 12V
- no changes to the capacitor values (see original post)

This is still not as good as your numbers but has to suffice for now

Is anybody else getting a stronger hum during RX on 12m and 10m from OLED display ?

73 de Jens, DL2RMM

[Ohwenzelph]

What's the OLED hum from and how does it get out of the OLED?

[DL2RMM]

What's the OLED hum from and how does it get out of the OLED?

.... not sure. I did the suggested OLED charge pump mod and it's working fine on the low band version . However, on the high band version the OLED noise is audible and most pronounced on 12m and 10m.
The hum changes with the letter count on the OLED display. So, display full of letters = strong hum and display with less letters (i.e. menu 1.8 AGC) almost no hum. I added a laminated aluminum foil as a "screen" between main board and RF-board without any improvement. Does anybody have a suggestion/cure ?

73 de Jens, DL2RMM

[DL2MAN]

@Jens DL2RMM:
I did not whitness the presence of more OLED Noise on the Higher Bands.
In fact I do not recognize any OLED Noise AT All since introduction of the OLED Mod.
Maybe the mod has gone wrong ?
What was the Basis for your Hi Band Rig ? *Rxxxx*/Sunny/Group Buy ?
What OpAmp is installed ?

Regarding your High Band Values: Is there a reason why you would
Mix yellow and red Toroids in one Band ?

73 Manuel; DL2MAN

[DL2RMM]

@Jens DL2RMM:
I did not whitness the presence of more OLED Noise on the Higher Bands.
In fact I do not recognize any OLED Noise AT All since introduction of the OLED Mod.
Maybe the mod has gone wrong ?
What was the Basis for your Hi Band Rig ? *Rxxxx*/Sunny/Group Buy ?
What OpAmp is installed ?

Regarding your High Band Values: Is there a reason why you would
Mix yellow and red Toroids in one Band ?

73 Manuel; DL2MAN

Thanks for the feedback Manuel.
My hardware is from the first german group buy (DC8LZ) and the OpAmp is LT6231.
I mainly worked on the LX1 cores to improve efficiency and ended up with the T37-6 on 10 and 15meters. All other T37-2 toroids were tweaked and got me in the 70-80% range. So, I did not bother with changing the LX2 to T37-6 as well.
I think the main variable is the core manufacturer, Amidon vs. Micrometals vs. Sunny's chineses cores. I do not have any of the (good) chinese cores, so could not test those.

Anyway, besides the hum on 12 and 10meters I'm a happy camper. The higher bands are getting better every day now with the rising SSN numbers.
"The (tr)uSDX works some true DX"

73 de Jens, DL2RMM

[S58J]

Hi

Regarding T2 modification: I cannot find topic related to it. So now instead of 7 plus 1 turn the T2 should be 21 plus 3 turns, right? The T1 remains unchanged (7 plus 1)?

Kind regards

Janez, S58J

P.S.
I am doing the "all T37-6 toroid" high band version. The T37-6 toroids are from Reichelt. I'll inform you about the results.

[DH2LAB]

Regarding T2 modification: I cannot find topic related to it. So now instead of 7 plus 1 turn the T2 should be 21 plus 3 turns, right? The T1 remains unchanged (7 plus 1)?

viewtopic.php?t=395

[S58J]

Thank you Matt!

Best regards

Janez, S58J

[aa777888]

If I build the classic version will the software allow me to select a frequency in the 17M band and use the 15M LPF?. Or does the software lock out all bands other than those listed for the classic version?

If the software does lock out other bands, why? The classic LPFs as configured should be adequate to control 2nd and higher harmonics from the next band down in frequency.

[AC9YY]

If I build the classic version will the software allow me to select a frequency in the 17M band and use the 15M LPF?. Or does the software lock out all bands other than those listed for the classic version?

If the software does lock out other bands, why? The classic LPFs as configured should be adequate to control 2nd and higher harmonics from the next band down in frequency.

That won't work as it is more than just an LPF for class E amplifiers. The efficiency will be poor.

[k5vox]

Manuel,
What formula did you use to calculate the values for the LPF ?
I am using the WA0ITP spreadsheet and I am getting different values.

As for the 'theoretical' values of the LPF for the Hi bands I can confirm that these are correct for 3xBS170S.
I used combinations of serial or parallel capacitors to get the values as close as possible.
I used the method of adjustment with NanoVna you recomended.
I used T37-6(yellow) and measured the inductance with inductance meter.
This took me close enough.
When I adjust LX1 for maximum power it is more than 5W so I lowered it with Menu 8.2 PA Bias max.
This setting now is 110.
The results are:
20m 5W@84% efficiency
17m 5W@86%
15m 5W@92%
12m 5W@83%
10m 5W@86%

[Ohwenzelph]

Could you please post as much detail as possible on exactly what you did and used? Clearly you have gotten much much greater success than I with the same approach, so as much specific detail as you can come up with about how you made this work would be very much appreciated. And congratulations! Well done!

[John]

AC9YY, to the best of my understanding, the bias level is not a power adjustment, and shouldn't be used as such. Yes, it does have an affect on the absolute power output, but that's a side-effect, and not typically a good thing since the bias voltage is an offset used to set the point where the FET's are being operated in their most linear range, meaning generating the lowest distortion during transmit. Once there, it should be left there. Power output should then be adjusted by changing the output FET drive level, the toroid inductance, or the drain voltage.

I'm planning on feeding L15 (that goes to the output FET's drains) with an adjustable linear voltage regulator (probably a quasi-LDO) in order to gain good control of RF output so I can dial in the minimum power necessary and improve the reliability of the output FET's.

A linear regulator is hardly the most efficient way to reduce voltage, but it's much quieter RF-wise than a switcher, so I figure I'll take the efficiency hit in order to gain convenient control of power output without generating RF hash. For me and my needs I think it's probably a couple of bucks well spent.

[k5vox]

Power output should then be adjusted by changing the output FET drive level

John, how can I do that ?

Ohwenzelph, I followed Manuel instructions here:viewtopic.php?p=1686#p1686.
I had good T37-6(yellow) toroid cores to use. Bought some NPO 100V capacitors from Digikey. I checked them with capacitor meter before soldering, just to be sure they are ok. After adjusting the notch filter to the 2nd harmonic with the NanoVna, I set the frequency 100KHz down and transmit. Noted the power. Set frequency up 100KHz and check power. If power is more on the lower frequency strech the coils and check agian. If streched to maximum remove 1 coil and check again until you have maximum power in the middle of the band. This was my case with all the bands. LX1 was always bigger inductance, so I had to remove 1 coil for each band.I used the onboard indication and measured the current consumption with a multimeter. I tried to measure the RF voltage with my oscilloscope but discovered bad probes. So verified the output power with my Diamond SX-200 SWR meter, which I know is working good. Like I said, if I use PA Bias max@128 I get power about 6 to 7W with the same efficiency. I have already made a lot of QSOs FT8 on 20m band. I have good antenna on this band only at the moment.

[Murphyv310]

Hi.
I've now been able to just about complete my classic band conversion to the Sunny kit I got a few weeks back. I did the 3/21 SWR transformer mod. I've problems with two bands. 80 meters only 10% efficiency and virtually no output (all tests so far only on USB 5V) No output on 15, other bands around 62% efficiency. I'll need to recheck the caps in 80 & 15 and for possible shorts.
Made a mistake with the header in the respect of using the female on the processor board ( convention says power supply on female and male on unpowered side. Will make harmonic testing slightly more difficult.

I have to say working with these 0805 caps is a nightmare.

Will see if I can get some answers tomorrow.