Radioberry PreAmp running hot - How to set the bias?

[Samuel Lourenço]

Hi,

I've assembled my own Radioberry PreAmp board, but I had to do some replacements:
- AFT05MS004NT1 for the finals;

- Used B62152A4X30 core for the final transformer, but had to use 0.8 m enameled wire, first 4 turns for the secondary, and on top 1+1 turns for the primary.

When transmitting the AM carrier on 20 meters, the current starts around 800 mA, and goes up to 900 mA. The output voltage is 5.9 Vrms or 17 Vpp, which yields 0.7 W of output power. My question is, does the transformer need more turns? Because I see an increased output power on lower bands.

Also, I guess the bias is not set correctly. How to set it?

Kind regards, Samuel Lourenço

[Yado-san]

Hi Samuel,

I post the following for your reference.

bias setting:
I used PowerSDR to set the bias.
-> https://groups.google.com/g/radioberry/c/IlDGZeUKBLc/m/CgEpCpykAQAJ
   see p.4   bias setting.  (Fig.6,7)

The amplification rates are Fig.11,12 on p.5.
  The power ratio between 3.5 MHz and 28 MHz was about 3.6dB.
  -> radioberry2_TXout: 3.5MHz/28MHz=2dB
      radioberry-preamp_output: 3.5MHz/28MHz=1.6dB

T301 core:
 The following is a useful reference.
 https://github.com/softerhardware/Hermes-Lite2/wiki/Hermes-Lite-2.0-Getting-Started
 "TX Transmit Balun"
 and
 B62152A4X30 is a conductor.
 Using UEW wire may cause shorts with the core, leading to reduced output.

Yado-san, jg1twp

2024年6月16日日曜日 7:08:31 UTC+9 Samuel Lourenço:

[Samuel Lourenço]

Thanks Yado-san! I've managed to set the bias using SparkSDR, albeit I'll have to revisit that setting, now following the instructions for the HL2 (i.e. setting the bias separately for each MOSFET, to get 100 mA of quiescent current on each MOSFET).

Regarding the transformer, I took care not to scratch the enamel from the wire, and I used double-enameled wire. However, probably it was not enough and I may have to revisit that. The recommended Teflon-coated wire is thin and hard to strip, even with a proper tool. The process breaks quite a few strands. However, I do notice reduced output on the higher HF bands (say, 10 meters), and I don't know if that is due to a short, or due to the extra capacitance provided by the thinner isolation.

Kind regards, Samuel Lourenço

[Samuel Lourenço]

As an important update, I will have to redo the transformer. I've measured a relatively low resistance between each coil and the core, which means that the enamel is not doing it's job. Before the installation, I've tested the transformer and saw a very high resistance between both coils, but the enamel probably got scraped when bending the wires outwards during the installation.

To solve this, I got teflon coated solid core wire from TME, reference 2854/1-BR005. I also have bought an heatsink, reference SK646/50/SA, which is clearly insufficient for the job but may improve things a bit. I will have to open a threaded hole and secure it with a screw and thermally conductive glue to the underside of the PCB.

Overall, this amp board clearly needs some changes. The final transistors are discontinued, and the blue connector that attaches to the Radioberry needs to be replaced with a different, taller one.

Kind regards, Samuel Lourenço

[Samuel Lourenço]

Hi Yado-san,

I've redone the transformer using PTFE-coated, solid core 24 AWG wire. It was easier to work with, and strips quite easily with an X-acto knife, if one is careful enough. Despite having no current escaping through the core, the behavior is pretty much the same, albeit I'm burning a tad less power to get more power out. Still, I see some reduction in the higher HF bands. But that may be explained not by losses in the amplifier itself, but because the input signal is already reduced in amplitude.

Kind regards, Samuel Lourenço

[Yado-san]

Hi Samuel, Group

Thanks for your message.
The radioberry-2 output seems to drop by about -2.1dB(abt x0.62) in the higher frequencies.

-> a graph measured with pihpsdr bellow

Yado-san, jg1twp

ps.

TX gain does not lineally change with the current GateWear. (under drive 34, with pihpsdr)
Just to be sure, inproved GWs with measured TX-output are attached.
see: https://groups.google.com/g/radioberry/c/LoCRvkD_wYk/m/O_tKMsJzAgAJ

How to swap the GW:
$ sudo systemctl stop radioberry
$ sudo mv /lib/firmware/radioberry.rbf  /lib/firmware/radioberry.rbf_org
$ sudo cp radioberry.rbf_gw73.3_cl025_txgain_ok  /lib/firmware/radioberry.rbf
$ sudo modprobe -r radioberry
$ sudo modprobe radioberry
$ sudo radioberry
//

2024年6月23日日曜日 3:09:43 UTC+9 Samuel Lourenço:

[Samuel Lourenço]

Hi Yado-san,

I get just about 8 V RMS on 40 meters, when TXing an AM carrier at maximum drive. It seems to be under-performing, even after redoing the transformer and setting the bias correctly. Do you confirm this?

Kind regards, Samuel Lourenço

[Yado-san]

Hi,

My Radioberry2 output circuit was changed, so I can't be measured under the same conditions.
However, if the AM carrier and Tune carrier have the same power, 8Vrms@7MHz(abt. 1.3W) seems to be low.

Tnx, Yado-san, jg1twp

2024年6月24日月曜日 0:25:15 UTC+9 Samuel Lourenço: