Pages: [1] 2   Go Down

Author Topic: uSDR tests  (Read 1380 times)

HAMHOCK75

  • Member
  • Posts: 1297
uSDR tests
« on: October 24, 2022, 03:56:12 PM »



I recently bought a used uSDR which included a microphone and charger. Two features about this rig had my interest. First, it generates SSB signals but does not use a linear amplifier. Second, it’s amplifier is class E. Some reports indicate the amplifier around 80-85% efficient.

My first tests were of output power. The advertising says it has eight bands but it has power in 10 bands. It will allow tuning to 6M but 6M has no measurable output power. Power is controlled by menu “8.2 PA Bias max” which takes numbers from 1 to 255 with 255 the highest power. The curves of power output as a function of “PA Bias max” setting was done on 20M.

All power measurements were made with a LP-100A power meter with curves indicating what the LP-100A measured vs. what the uSDR displayed.





Harmonic tests done at full output power with an external +13.8V power supply except where otherwise noted. These tests were done with an HP8593E spectrum analyzer with a precision JFW 3 GHz, 40 dB, 100 watt attenuator between the uSDR and the analyzer. Since 10 watts = +40 dBm, a 0 dBm signal at the analyzer input is equal to 10 watts.













This is what the CW waveform on 20M’s looks like at full output power with an external +13.8V power supply.



Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #1 on: October 24, 2022, 06:02:38 PM »

Microphone gain is under menu “3.3 TX Drive” with a number value from 0-8. The following image has TX Drive set to 2. As can be seen my voice peaks hit the same peak amplitude as the maximum CW amplitude under the same conditions with an external +13.8V power supply and PA Bias max = 255.



There is a pop when the microphone disconnects. This may not a fault of the rig however. Yaesu microphones I have examined have timed switching with a set of contacts that disconnect the microphone element prior to the PTT disconnecting the rig so the sound of the rig disconnect cannot be picked up which may not be happening in the inexpensive OEM uSDR microphone.
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #2 on: October 24, 2022, 06:52:45 PM »

This is how it SSB and other modulations can be done with a class E non-linear amplifier.

https://core.ac.uk/download/pdf/148657773.pdf
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #3 on: October 25, 2022, 01:14:11 AM »

The uSDR did not come with a manual. This link is to the closest manual to the product I could find. There are differences.

https://view.officeapps.live.com/op/view.aspx?src=https%3A%2F%2Fmyosuploads3.banggood.com%2Fproducts%2F20210902%2F20210902235502SKUI41486uSDRMANUAL.docx
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #4 on: October 25, 2022, 03:41:09 PM »

The OEM microphone was disassembled.



It is a speaker microphone which appears to have been designed to perform more sophisticated functions but has been assembled to perform less well. The photo below shows a trace from the microphone that goes over to the PTT switch. The switch on the board is an SPST switch which leaves the microphone element live when it switches but the board appears to have been designed with a DPST switch in mind to cut off the microphone with the proper timing. The two red wires are connected to the speaker but one wire is not connected to anything. The other wire goes to a pad labelled -SP but nothing goes to that pad.



This is how the wires are connected to the microphone plug.

Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #5 on: October 26, 2022, 03:43:16 PM »

I started listening to the uSDR’s SSB signal using an Icom 705. The SSB signal sounded muffled. The uSDR needs a bit of calibration for SSB to sound better. From this curve of output power vs PA Bias max it is important to use menu 8.1 PA Bias min to set PA Bias min near the knee in the curve and menu 8.2 PA Bias max to just before the curve flattens out. The SSB signal amplitude then appears to then occupy the powers between these set points.



It’s not very clear from the manual but I believe this what they mean in the description of menu 8.1 where they say PA Bias min “represents 0 % of RF on output”

At the risk of oversimplifying, having a low value of PA Bias min. allows the audio to clip like this,



« Last Edit: October 26, 2022, 03:49:32 PM by HAMHOCK75 »
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #6 on: October 27, 2022, 12:50:43 PM »

Sensitivity Tests

These were run with an HP8656B signal generator set to 0.5 uV output connected to the uSDR via a semi-rigid cable. The uSDR was set to a volume of +11, the bandwidth to 500 Hz. 0.5 uV could be easily heard on all band froms 80-10M. A bandwidth of 200 Hz allowed measurement down to about 0.3 uV.

An attempt to use 100 Hz bandwidth was less successful. The character of the background noise changed with more popping sounds vs the smooth white noise at wider bandwidths which made it difficult to hear the signal generator. An attempt was also made to measure the sensitivity on the 160M band but this proved impossible due to a buzzing sound present at most volume levels. Even a 100 uV signal at 1840 kHz could not be heard over this noise.

Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #7 on: October 29, 2022, 04:45:02 AM »

Recently, I viewed this video. My firmware version is R1.02v.

https://www.youtube.com/watch?v=57C8UgI_U9A&t=89s

According to the video, version R1.02w7 has these improvements.

1. support for units with more bands
2. resolve a distortion problem in SSB TX ( new menu 8.7 quad on/off  for that )
3. more functions like CAT on/off
4. ability add your call sign to the display
5. the tuning knob can tune bands up and down
6. if the default frequency and mode are changed, the unit will now remember those changes if bands are changed. Not sure if this is true if the unit is turned off.

This is not mentioned in the video but I notice with his unit the tuning knob increases frequency or the number of the menu items while turned clockwise. I hope that was a firmware change because my unit has it reversed.

An attempt will be made to update the firmware prior to any more SSB TX quality tests. The current firmware version seems to be R1.03.
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #8 on: November 01, 2022, 04:50:26 AM »

The uSDR firmware has an unusual set up for a straight key which is not mentioned in the manual  Like my Icom rigs, the uSDR uses a 3.5 mm trs ( tip, ring, sleeve ) plug connector. The connection for the key for Icom is between the sleeve and tip, the uSDR has the connection between sleeve and ring.

An iambic key was tested but it did not work. Only dit’s from the right paddle worked but dah’s from the left paddle did not work. “Menu 2.7 Keyer Swap” did work so that paddle producing dit’s would send dahs’ if Keyer Swap is set “ON”. An ohmmeter test indicated that the keyer paddles were closing the correct contacts as expected. The keyer also functioned perfectly with an Icom 705.
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #9 on: November 02, 2022, 03:27:56 AM »

The uSDR had to be disassembled to access the programming header.



The back side has the labels of the pins



The battery was examined at the same time. Some of the text was worn away but 55.5 watt hours translates to about 5,000 mAh.



The output power transistor is a Vishay Siliconix IRF510. The IRF510 has a rated power dissipation of 43 watts.

https://www.vishay.com/docs/91015/irf510.pdf

A quick calculation assuming a worse case efficiency of 70% for the class E amplifier shows that at 10 watts output, the IRF510 will only be dissipating about 4.3 watts or 1/10 of it rated power dissipation. As shown below, it is bolted to the aluminum extrusion.



Logged

KB1GMX

  • Member
  • Posts: 2252
Re: uSDR tests
« Reply #10 on: November 02, 2022, 06:56:13 AM »

Good luck getting a IRF510 to run class e at more than maybe a few Mhz.
Class C is easier and efficiency at that will  be low end maybe 60%

THe IRF510 is 43W saturated.  Class C its more like 25W due to
heat transfer limits for the TO220 package.   At 12V even less.

Poor design it becomes a dandy power oscillator and dies.


Allison
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #11 on: November 03, 2022, 02:45:53 PM »

While the uSDR is apart for firmware update, I received a different version called the uSDX+ for evaluation.



Here are some notes about differences between the two.

The uSDX+ seems to have newer firmware than the uSDR 1.02v. For example, it has new functions for setting the sidetone frequency and level. The newer firmware has the tuning knob increase frequency while rotating clockwise. The uSDR 1.02v has it reversed.

The key jack is conventional with a 3.5 mm TRS connector with the key between sleeve and tip. The uSDR has the key between sleeve and ring.

The uSDX+ has a 3.5 mm key/mic jack on the front panel but on the rear panel it accepts a Baofeng style speaker/mic with a 2.5 mm and 3.5 mm plug molded into a one piece plug as shown below. The uSDX+ does not have any LED’s that light up in transmit mode. The letter “R” in the lower right side of the display does change to “T”. If using the speaker/mic, a red LED on the mic lights up.



Instead of two graphs, one showing output power vs PA Bias max for one band, another showing maximum output power for each band for the uSDR, I combined them by plotting output power vs PA Bias max for every band with the uSDX+.

There is so much variation between bands that the appropriate PA Bias max can range from 70 to 170. Also, I was surprised it is able to deliver over 6 watts on 10M. These tests were all done with the internal battery.



Harmonics are about the same between the uSDR and uSDX+. The following are uSDX+ harmonics at full output power using the internal battery.



Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #12 on: November 03, 2022, 03:14:58 PM »

A quick look inside the uSDX+



Three BS170’s are used for the final. The 6 pin firmware update header is visible but it looks like the speaker would have to be removed to access it.

Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #13 on: November 06, 2022, 01:10:05 AM »

The uSDR was successfully updated to firmware version 1.02x. This was not a simple task. The programming six pin header is located under the LCD display. Access requires almost complete disassembly. I made a cable to connect the Arduino UNO acting as a programmer to the uSDR which fits under the LCD display. I went through the programming process three times. The first time, the 1.02x firmware was uploaded without any changes.



I was disappointed 1.02x leaves the tuning knob so that it turns clockwise to decrease frequency. The second and third uploads involve changes made to the firmware as shown below. To have the tuning knob increase frequency in the clockwise direction, the line beginning “#define SWAP ROTARY…” had two forward slash marks in front of it which converts it to a comment. I removed the // marks to reverse the rotation direction.

The third upload was due to the line starting “//#define F_XTAL 27005000…”. This line did not have the // in front if it which meant it was expecting the uSDR to use a 27005000 Hz TCXO. The uSDX+ uses a 27 MHz TXCO but the uSDR has a 25 MHz TCXO so the // had to be added to the line “#define F_XTAL 27005000…” and removed from the line “//#define F_XTAL 25000000…”.



If another upload is necessary, it would be a good time to add a header in a location that does not require so much disassembly to access it. After compiling the firmware the Arduino programming environment shows the resources left for the Atmel mega328. Very little in terms of memory is left.

This is a re-examination of the speaker/microphone for the uSDR. From Internet searches, this appears to be a version of the Oppxum, KMC-21 which is available with various connectors.

As described earlier, the KMC-21 connector is wired as shown below,



However, an examination of how the uSDR connects is shown below. The uSDR reverses the positions of the speaker and ground connections from the KMC-21.



The KMC-21 works without modification as shown below. The uSDR grounds Ring 2 not the sleeve with the result that the PTT switch and microphone return to ground through the speaker.



In the photo below, if the black and white wire positions are reversed results in the modified version where the PTT and microphone do not return to ground through the speaker. The KMC-21 OEM configuration is standard for most microphones with the sleeve as ground. It is pretty unusual to use Ring 2 for ground.



The front panel speaker did not turn off when the speaker/microphone is connected but it does turn off if a speaker is plugged into the the separate speaker jack.

Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #14 on: November 10, 2022, 01:01:22 PM »

Firmware notes.

The uSDR identifies its firmware version at turn on in the upper right side of the display momentarily. For the uSDX+, the “menu” button needs to held down while the unit is turned on. The unit will reset and show the firmware version, again, in the upper right side of the display.

Both the uSDR and uSDX+ use an Atmel mega328 micro-controller with 32k of program space. The program space appears to have been exhausted some time ago. The only way to add improved code or new functions is to eliminate older functionality. This was discovered after v1.02v was updated to v1.02w or v1.02x.

The uSDR with v1.02v had “fast AGC” and “output power/swr” functions. Both of those disappeared with v1.02w which is said to have enhanced SSB capability and v1.02x which is said to have improved CW. The iambic keyer now works as it should under 1.02w and 1.02x.

Shown below are the “advanced configuration switches” in the firmware. The “//” marks eliminate that code from being compiled. Removing the “//” will allow the function to compile adding that feature to the firmware.



When the “//” were removed from “//#define FAST AGC” and “//#define SWR_METER” to add these functions resulted in the following error message.



Adding the “//” marks to “#define CW_DECODER” removing that function, provided enough memory space for the firmware to compile.
Logged
Pages: [1] 2   Go Up