开云体育

After getting my licence I want to take my first steps in transmitting so I bought an Ensemble II transceiver kit for 40/30/20 meters
Building is a ongoing but I'm wandering it this kit can be used in a more generic way.
Its basically an I/Q modulator/demodulator with a simple power amplifier.
Would it be possible to eliminate the band specific parts (including the BS170 power amplifier) and use the kit as an all band FT-8 transceiver with appropriate external band filters?
I'm able to design and measure filters with Elsie and a home build VNA and I have a home build SA to evaluate spectral purity and inter modulation of the transmitter output.
Output level will be a couple of milliwatt but for first experiments that should be enough. Later I can add an external 10 Watt al band amp if needed
Would this work???

Would it be possible to eliminate the band specific parts (including the BS170 power amplifier) and use the kit as an all band FT-8 transceiver with appropriate external band filters?

Eric?

Some have done this, you may search this group's archives. I do not remember how detailed the descriptions go but anyone who understands the principles should have no problems.
I guess I've gone part of the way, a 3 band LF Ensemble on this page Note the link to Roger's description for more.

73 Alan G4ZFQ

Alan,

This, will dig into the archive.

--
Erik, PD0EK

Eric,

I have built an all-band RXTX mod.
See
Meter RXTX Modification.pdf <>
There are other related files. See the main page

and scroll down to "Articles”.
73,
Roger

Roger, Thanks!
Very helpfull

One more question. I notices the RX has a substantial I/Q phase misalignment (ast least 2 degrees) not constant over the whole audio bandwith so difficult to optimize
I read C70/C71 should be increased to 0.1uF to improve phase alignment
I changed these but did not see the improvement.
Is there anything I can do?
--
Erik, PD0EK

And here is the measurement of the phase and amplitude error as done by Rocky.

Anything that can be done to improve this?
In particular the amplitude error discontinuity around zero hertz makes correction difficult
--
Erik, PD0EK

Eric,
What is the frequency scale in your plot?
Roger

48kHz audio bandwith

--
Erik, PD0EK

What is the leftmost, y axis intercept, frequency? I think it can not be zero, or I do not understand the plot method.

Sorry, this is the phase amplitude mismatch versus the audio frequency. The total audio bandwidth is equal to the audio sample rate and the center is the tuned frequency and zero Hertz in audio.

--
Erik, PD0EK

Erik

Thank you for explaining, now I understand the plot. I have never used Rocky.

Re. the RX amplitude problem around zero Hz.
I have not looked into this as I use PowerSDR which avoids the zero freq issues by using a 10 kHz offset.
However, looking at the schematic I thought that increasing C74, C75 to at least 1 uF, and using matched values, could help given that the load on the op amp output is typically 10 kOhm or less.

It is possible that replacing T5 will help the RX balance issue as well. I can not remember now for sure.
See page 2 of my mod pdf, namely.

4) Replace transformers as follows.
Use either BN43-2402 or BN43-1502 binocular cores.
I used BN43-1502 cores for T2, T3 and T4 because the larger holes made it easier to wind them
using larger size wire (#26). The smaller BN43-2402 cores (with #30 wire) should work also, as far
as I know. Do not use type 61 cores.
For all the ferrite core transformers, to avoid unwanted coupling make sure that
no leads (other than the primary and secondary wires) carrying significant RF
currents touch or are closer than 1/8 inch from the core.
a) Replace
T5 4T:2T bifilar (This is the same as used in the Ensemble II and III units.)

Note that there is a similar issue on TX
see page 5 of my mod pdf, namely.

Additional consideration.
C11, C12, C13, C14 are only 10 uF in the stock RXTX. This is adequate for audio drive
frequencies greater than 3 kHz. However, if you intend to use lower frequencies, e.g. for
digital modes close to the LO, or if you use PowerSDR or GSDR where the LO is shifted
during transmit such that the audio modulating frequencies will be within the filter bandwidth
(e.g., 150 to 2400 Hz), then these capacitors should be increased. I used 10 V 220 uF
electrolytic caps of low electrical series resistance (low ESR) type. The Panasonic FR line is
good. I also added a 220 uF cap on the 5 V line near the TX op amps.

Roger

And here is the measurement of the phase and amplitude error as done by Rocky.
Anything that can be done to improve this?

Eric,

This is a difficult question, I have known soundcards that vary greatly over their bandwidth.

In particular the amplitude error discontinuity around zero hertz makes correction difficult

It is not recommended to use frequencies near the centre. The best way is to use a constant offset, maybe 10KHz. Provided the software is able to compensate this is the best answer.

73 Alan G4ZFQ

On 13/11/2019 20:16,

I have never used Rocky.

Roger,

You should, I think you will find this is typical.
I would say it looks normal.

73 Alan G4ZFQ

I would say it looks normal.

And note that Rocky does not even attempt to compensate near the centre.

73 Alan G4ZFQ

This from another SDR I own. Much easier for rocky to compensate

I'd like to understand if it is improve the variation in phase and amplitude error so hdsdr will better

--
Erik, PD0EK

This from another SDR I own. Much easier for rocky to compensate

Eric,

What SDR?
Using the same soundcard?
In any case I do not know what you are worrying about, both are compensated it will never be perfect.

73 Alan G4ZFQ

Erik,

I looked at the schematic for the QRP Labs receiver kit. The op amp outputs for that have 1 uF coupling caps instead of the 0.1uF in the SoftRock.
Increasing the SoftRock C74, C75 to 1 uF may improve the phase and amplitude low frequencies (close to the zero Hz point). If you try it, please let me know if you see any effect.

Roger

Roger,

Thanks for all the excellent info (this and previous post)
Will try and report back

--
Erik, PD0EK

I looked at the schematic for the QRP Labs receiver kit.

Roger, Eric,

I missed the clue in the filename. So it is a QRP Labs RX.
With the same soundcard?

The op amp outputs for that have 1 uF coupling caps instead of the 0.1uF in the SoftRock.
Increasing the SoftRock C74, C75 to 1 uF may improve the phase and amplitude low frequencies (close to the zero Hz point). If you try it, please let me know if you see any effect.

If this is the case then in all the years of the Softrock I do not think it has never been mentioned before.
The QRP Labs RX is primarily designed to run at zero offset (and reduced audio bandwidth?) as a stand-alone receiver, no PC.
The Softrock has always been an IQ RX. The advice has always been to keep away from the centre to avoid the "hole" and noise around it. With software like HDSDR, no automatic cancelling, a constant offset where IQ is balanced is preferable.
It will be interesting to see Eric's results.

73 Alan G4ZFQ

Alan,

Yes, same sound card.
I'm using the QRP Labs RX only as wide band I/Q receiver with HDSDR (coupled with 24 bit 192kHz external sound card and I've removed the output transformers and reduced the low pass filters in the audio path) and I would like to use the softrock in the same way when not doing FT8. The softrock is much better in avoiding ground loops (isolated USB interface) so much cleaner around zero Hertz
If you look at the softrock schematic the opamp output capacitors in the TX path are 10uF, and yes, there they drive 50 ohm but I guess the line impedance of the PC could also be rather low (200 ohm?) so I plan to increase C74 C75 to 10uF and match the integration caps after the switch.
--
Erik, PD0EK