Pages: 1 [2]   Go Down

Author Topic: uSDR tests  (Read 1377 times)

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #15 on: November 12, 2022, 03:14:40 PM »

These are some SSB transmission tests. The test set up is either the uSDR or uSDX+ is connected to a Wavetek function generator as the audio source. The Wavetek is connected via a 10 ufd tantalum capacitor ( removes the +5 volt supplied by the uSDX+ to the electret microphone normally used ) to a 3.5 mm TRS connector attached to the uSDX+. Transmit occurs when a jumper shorts the sleeve of the TRS connector to the ring. The Wavetek modulation frequency is 1 kHz.



This is an image of the output of the uSDR+ to a digital scope with its input terminated in 50 ohms via 40 dB attenuator. 196 mVp is about 3.84 watts at the input to the 40 dB attenuator in CW mode.



This is an image of the output of the uSDR+ in LSB mode on 40M with the audio level adjusted for the same output power as in CW mode.



As the audio level is reduced, the RF output drops as expected but with some noise.



As the audio level is reduced more, the unit begins to transmit random full power pulses.



It gets worse as audio level is reduced again.



When there is no audio drive at all, the pulses do not occur. The above images were all taken with Noise Gate = 4. Below is the same audio level as before except the Noise Gate = 0. For the audio level shown, it was found that a Noise Gate level of <4 ( in other words Noise Gate = 0,1,2, or 3 ) eliminated the pulses. As the audio level was reduced further, the Noise Gate value also had to be reduced to prevent the pulses. Since the human voice has a large dynamic range, it is not clear why the Noise Gate would ever be set to other than 0.

This might not be true for digital modes which use audio tones of constant amplitude in which case a Noise Gate value higher than 0 would still work.



The above addresses SSB issues in the time domain. Now let’s have a look in the frequency domain. The test set up is the same as above but an HP8593E is used to view the output of the uSDX+ through a 40 dB attenuator.

The signal in the center of the screen is the uSDX+ in CW mode. The signal to the left of it is the 1 kHz LSB signal. The important difference is the the LSB signal has a considerable noise floor which is many kHz wide.



For comparison, this is the Icom 705 under the similar conditions.

Logged

KB1GMX

  • Member
  • Posts: 2252
Re: uSDR tests
« Reply #16 on: November 13, 2022, 05:41:09 AM »

Look at what it does to SSB two tone....
Then vary the levels both, and each.

For data like FT8 it works as they are single tones and saturated.

Allison
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #17 on: November 14, 2022, 10:27:12 PM »

Quote from: KB1GMX
Look at what it does to SSB two tone....

It's on the list but a few other measurements first.

The transmit bandwidth of the uSDX+ and uSDR were both measured. The HP8593E is used to capture three traces. At the center is the CW signal, 1 kHz below is the LSB, then the Wavetek function generator is moved to determine the 3 dB bandwidth relative to the amplitude of the 1 kHz LSB. Below, the upper 3 dB bandwidth appears to be about 2.2 kHz from the CW carrier.



The lower 3 dB bandwidth is about 350 Hz from the CW carrier.



The transmit bandwidth is about 2,200 – 350 Hz = 1850 Hz. This is a bit on the narrow side compared to most SSB rigs with 2.1-2.4 kHz being fairly typical.

These are curves of output power for the uSDX+ with an external +13.8 volt power supply.



These are efficiency curves calculated from the current drawn during the above tests. The efficiency is calculated from the RF output measured with the current drawn from the +13.8 volt power supply with a correction. For PA Bias max = 0, the RF output is always zero measured on the LP-100A power meter but for 15M where it measured .01 watt. The current for each band at PA Bias max = 0 is subtracted from the total current to arrive at the current for the PA amplifier.



The efficiency is not a constant, varying considerably from band to band and from the PA Bias max used.

It may be a good reminder at this point the hear how the originators of the uSDR, uSDX see these rigs. In this video, one of the originators at about 49:30 minutes into the video says “you’re not buying a completed commercial rig…”

https://www.youtube.com/watch?v=nBPIdE0tRo0

Logged

KB1GMX

  • Member
  • Posts: 2252
Re: uSDR tests
« Reply #18 on: November 15, 2022, 08:35:10 AM »

I got involved with the development in the QCX as starting point days.
After much fussing on a monoband unit I I thought initially it has
promise.

After a few months of trying I got tired of "is that you, your audio is terrible.)
where "terrible" was anything from unreadable to sounds like RF on audio
or blown final. 

As to efficiency as a system for he power out it was a push. Made it up with
class E fed to offset the MPU and analog drive.

After a lot of fussing it could so really good AM.  Problem then was the RX
is SSB.

Over all I gave it a 4.5 for interesting, 2 for useful as SSB radio.
I then reverted the QCX to CW where its still a 5, as CW radios go.

Over all, it proves why that scheme has great difficulty meeting
its claims. 

Allison
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #19 on: November 22, 2022, 03:15:07 PM »

Quote from: KB1GMX
Over all I gave it a 4.5 for interesting, 2 for useful as SSB radio.
I then reverted the QCX to CW where its still a 5, as CW radios go.

These rigs bring back fond memories I have of the TenTec Century 21 CW only transceiver, but with much better frequency accuracy and stability. Listening to SSB is also very pleasant with the 3 kHz or greater bandwidth.

I have not given up on SSB just yet. These are tests with the supplied speaker/mic in both PTT and VOX modes.

This is how traditional VOX works in most rigs. Traditional VOX has three controls.

a. Anti-VOX which feeds back audio from the speaker with the correct phase to cancel speaker sounds from activating the VOX

b. VOX sensitivity which is self explanatory. This seems to be the closest function to what is called the Noise Gate.

c. VOX delay which sets how long after a person stops talking the VOX disengages and returns the rig to receive

Neither the uSDR or uSDX+ have anti-VOX or VOX delay. With no anti-VOX, if the Vol is set high with Noise Gate = 0 and VOX engaged, the uSDX+ will immediately to into transmit mode and stay there. Turning the unit off will return VOX to its default which is off at which point the uSDX+ will stop transmitting. To avoid this, the Vol must be turned down. If the unit does not go into transmit mode, a capital “V” will appear on the lower right side of the display to indicate it is in VOX mode.

VOX related functions do not normally affect PTT operation. This is not the case with the uSDR and uSDX+ where the Noise Gate has a noticeable mostly negative affect upon PTT operation. The uSDX+  was tested with it’s speaker/mic in PTT mode instead of using the Wavetek to bypass the microphone. As might be expected from the Wavetek tests, Noise Gate = 0 is best, Noise Gate = 9 has lots of pulses.

Below, PTT mode, Vol = 6, Noise Gate = 0



Below, PTT mode, Vol =6, Noise Gate = 9



These are images for the uSDX+ with it’s speaker/mic,

Vol = 0, Noise Gate = 0, VOX = on



Vol = 0, Noise Gate = 4, VOX = on



Vol = 0, Noise Gate = 9, VOX = on. For some reason at Noise Gate = 9 the pulses diminished with VOX where with PTT is was much worse.



Vol = 0, Noise Gate =9, VOX = on, same as above with voice added



Vol = 6, Noise Gate = 9, VOX = on. Raising the Vol from 0 to 6 brought this but there was a vast improvement in the last image if the volume was backed down to 5.



It does not require a scope to see these full power pulses during receive. They can be heard in the speaker/mic sounding like ignition noise. The developers may be aware of this problem with pulses on transmit. The attached note is embedded in firmware version 1.02x. I suspect they cannot develop VOX further or solve the pulse problem because they have run out of program memory space.



It looks like for the cleanest SSB audio at this time, it is best to avoid VOX; use PTT with Noise Gate = 0.

Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #20 on: November 30, 2022, 05:23:08 PM »

Additional tests on the uSDR



Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #21 on: January 09, 2023, 11:09:35 PM »

A relative helped with making audio recordings of the uSDR, uSDX, and Icom 7100 to the Icom 705.

It looks like there is no convenient way to upload those "wave sound" files to eHam. They open readily with Windows media player. The files are less than 231k.
Logged

HAMHOCK75

  • Member
  • Posts: 1297
Re: uSDR tests
« Reply #22 on: January 27, 2023, 02:42:07 PM »

The uSDX+ has a very useful function. It is the only rig I have which has a CW decoder. It can not only decode off the air but when it is set up as a code practice oscillator for straight key, iambic A or B, it can decode the internal code practice oscillator. It has helped my sending. I did not realize that I was leaving excess space between characters so the decoder would show the word REPLACED as RE P L A CE D. It is showing correctly in the photo below.

The uSDX+ firmware allows setting the CW Tone ( menu 2.2 ), Keyer Speed ( menu 2.5 ), Keyer Mode ( menu 2.6 ), Keyer Swap ( menu 2.7 ) which reverses the paddle for dots and dashes, and Practice ( menu 2.8 ) The uSDR firmware does not have CW Tone.

Logged
Pages: 1 [2]   Go Up