Topics

uBitx Modulation only lows no highs in frequency response

Björn Pasteuning
 

I have the uBitx for a couple of days now, and have several "issues" with it.

One of them is modulation frequency response.

Monitoring my TX audio sounds like talking through a cloth or something, no high tones are audible.

Is this normal behaviour of this kit?

I have done all the alignments, Masterclock/BFO

 


--
73' Björn de PD5DJ
www.pd5dj.nl

 

Björn, If you are using the electret element that came with the uBitx, I would try a different one. I haven't been able to use the mics that came with any of my Bitx kits because of bad audio quality. I bought a whole bag of them from a eBay vendor for just a few dollars that work perfect.

Joel
N6ALT

Björn Pasteuning
 

Hi Joel,

Thanks I will try that, I think I have a little bag of gathered electreds lying around somewhere.

Will keep you posted.
--
73' Björn de PD5DJ
www.pd5dj.nl

ajparent1/KB1GMX <kb1gmx@...>
 

You may have to move the BFO down more for better audio. 
Its the first thing everyone tries, 100 to 200 cycles may help greatly. 

Also more than a few have narrower than expected 12mhz filters.  
Mine swept on the analyzer at under 1700hz wide and reducing
the parallel caps from 100pf to 91pf got it out to over 2100hz.

Allison

Björn Pasteuning
 

Hi Allison,

Well I think that is the culprit.

I have tried about 7 electrets after my last posting, some sounded a minor bit better others worse.

No significant big improvement.
Also tried playing with the BFO, not much help there either.

Do you have any link regarding the parallel caps? or schematic how this could be done?

--
73' Björn de PD5DJ
www.pd5dj.nl

Jerry Gaffke
 

Allison is suggesting that the 100pf caps at C217, C218, C219, C220, C221, C222
be swapped out with something smaller.

The board is laid out for 1206 sized surface mount caps, though 0805 sized parts could be used.
These sizes are in imperial units, something near that size in metric should work fine.
    https://en.wikipedia.org/wiki/Surface-mount_technology#Rectangular_passive_components
Dielectric should be C0G or NP0.
Voltage rating doesn't matter much here, anything 25v or more should be fine.
Use caps of 5% tolerance (preferably better), or measure them with a capacitance meter.
    https://groups.io/g/BITX20/message/52471https://groups.io/g/BITX20/message/52471
Standard values that might be tried include the current 100pf, or Allison's 91pf, or  82pf.
they lower the value the wider the bandwidth.
Though changing those capacitance values may change the shape of the filter response
with frequency, ideally we would sweep the filter shape somehow to make measurements
before and after the mod.

As you know, crystal filters in an SSB transceiver like this are a trade-off.
A wider filter may sound better, but may not be suitable when there are many
stations operating on nearby frequencies.
 
Would be interesting to measure the capacitance of any 100pf caps removed,
with 10% or worse tolerance parts we would see significant variations n performance.

Jerry


On Thu, Jul 19, 2018 at 01:20 PM, Björn Pasteuning wrote:

Hi Allison,

Well I think that is the culprit.

I have tried about 7 electrets after my last posting, some sounded a minor bit better others worse.

No significant big improvement.
Also tried playing with the BFO, not much help there either.

Do you have any link regarding the parallel caps? or schematic how this could be done?

Peter Parker
 

The carrier oscillator is an incredibly important adjustment that greatly affects transmitted audio.  This could be useful. 

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

RCC WB5YYM
 

With my radio/engineering skills very much lacking, I probably don't even need to be trying to contribute to this conversation, but here is what I ran into on my build. My tx audio was very distorted when I first tested. I then checked the transmit using the tones sent from WSJTx. I found that no tones were passing that were around 500hz or lower, but higher tones around 1600hz caused more power to be output from the radio. I adjusted the  BFO until I had maximum power out when sending a 1khz tone. This gave me power out in a range from 200hz to 1600hz. This seemed to clean up the audio on SSB. Apparently I have a very narrow filter, but I am getting good audio reports on SSB. The down side is that with it being so narrow, when operating FT8, I need to shift the VFO frequency to see all the transmissions. If I remember right, to shift the audio frequency down, I had to adjust the BFO frequency up, but please don't hold me to this statement . I did adjust about 100hz at a time, and checked the power out using different frequencies generated by WSJTx. I don't exactly know what the above means, but it did help me to have a good working radio. Hope this helps someone. 

Ralph Mowery
 

I found the filter in my ubitx to be about the same narrow range.  As I work ssb and no cw I modified the filter circuit slightly.  I replaced the 5 100 pf capacitors with some 82 pf capacitors and that seemed to broaden the audio response to a more normal ssb filter band width.

de ku4pt


On Sun, Jul 22, 2018 at 9:19 AM, RCC WB5YYM <curtis03@...> wrote:
With my radio/engineering skills very much lacking, I probably don't even need to be trying to contribute to this conversation, but here is what I ran into on my build. My tx audio was very distorted when I first tested. I then checked the transmit using the tones sent from WSJTx. I found that no tones were passing that were around 500hz or lower, but higher tones around 1600hz caused more power to be output from the radio. I adjusted the  BFO until I had maximum power out when sending a 1khz tone. This gave me power out in a range from 200hz to 1600hz. This seemed to clean up the audio on SSB. Apparently I have a very narrow filter, but I am getting good audio reports on SSB. The down side is that with it being so narrow, when operating FT8, I need to shift the VFO frequency to see all the transmissions. If I remember right, to shift the audio frequency down, I had to adjust the BFO frequency up, but please don't hold me to this statement . I did adjust about 100hz at a time, and checked the power out using different frequencies generated by WSJTx. I don't exactly know what the above means, but it did help me to have a good working radio. Hope this helps someone. 


Jerry Gaffke
 

You got it right.
Raising the BFO frequency brings it closer to signals coming from the filter, which lowers the audio frequencies..
We have a BFO below the 12mhz crystal filter passband  which mixes (at D5,T7) with signals in that passband creating audio.
Lets give it some numbers, the numbers for your particular rig might be a bit different.
I'm assuming we are in receive mode, transmit is the same math but in the opposite direction.

The BFO frequency in the stock firmware is at 11996500 hz.
The 12mhz crystal filter has a roughly 2000 hz 3dB passband, something like 11997000 to 11999000.
A signal coming in through the crystal filter on the bottom edge of the passband creates an audio tone of 11997000-11996500 = 500 hz.
A signal coming in through the crystal filter on the top edge of the passband creates an audio tone of 11999000-11996500 = 2500hz.

If we now raise the BFO frequency from 11996500 to 11996700 hz, 
the audio coming through would fall between 11997000-1196700 = 300 hz and 1199900-1199700 = 2300 hz.

Jerry


On Sun, Jul 22, 2018 at 06:19 AM, RCC WB5YYM wrote:
With my radio/engineering skills very much lacking, I probably don't even need to be trying to contribute to this conversation, but here is what I ran into on my build. My tx audio was very distorted when I first tested. I then checked the transmit using the tones sent from WSJTx. I found that no tones were passing that were around 500hz or lower, but higher tones around 1600hz caused more power to be output from the radio. I adjusted the  BFO until I had maximum power out when sending a 1khz tone. This gave me power out in a range from 200hz to 1600hz. This seemed to clean up the audio on SSB. Apparently I have a very narrow filter, but I am getting good audio reports on SSB. The down side is that with it being so narrow, when operating FT8, I need to shift the VFO frequency to see all the transmissions. If I remember right, to shift the audio frequency down, I had to adjust the BFO frequency up, but please don't hold me to this statement . I did adjust about 100hz at a time, and checked the power out using different frequencies generated by WSJTx. I don't exactly know what the above means, but it did help me to have a good working radio. Hope this helps someone. 

Bill Cromwell
 

Hi Jerry,

That is a pretty good description. I have my BFO set so that the incoming signals start rolling off at 300 Hz just as you have described. A few decades ago when we were looking at "compandered" SSB we were shown that the human voice has three main 'bands' of audio spectrum that are required for intelligibility. The lowest band is at 300 Hz. there is a mid band (I don't recall the frequency range) and high band around 1500 to 1800 Hz. Compandering used some Rube Goldberg approaches to squeeze those bands all together (with multiple filters and mixers) so as to reduce the RF bandwidth used to transmit it. There was also some kind of pilot tone to help sort it back out at the receiver. Did I mention Rube Goldberg? It kind of worked in the lab. But not in the wild.

So our SSB filter would like to bridge all of those bands without mashing or bending them. Even with the narrower passband of the stock filters in the uBitx we can get all of that fitted in if the BFO presents that lowest frequency band at about 300 Hz (as you have described). The high end will be close to 19 or 20 kHz. Mine is right around 1900 Hz. "Narrow SSB filters are designed and installed with a 1.8 kHz bandwidth. So the uBitx filters are at the narrow side of SSB. They may sound a little better for armchair chatting at 2400 to 3000 Hz. The bottom end still wants to be rolling off stuff starting around 300 Hz and the wider passband used to to increase the high end of the audio spectrum.

That will probably be very good for digital operation too. I am a CW guy and it does nothing to hurt CW either. I have several times outlined how to set the BFO this way by using the noise shadow displayed on all of the digi-mode screens and the audio frequencies shown on the scale below it. If yours is working well for you just go take a look - without changing the BFO. The noise starts to drop off about 300 Hz on one end (or wherever your is actually set) and the other end it starts dropping off well above 1800 Hz. With CW (or other narrow) filters that noise band is obviously much narrower and the width of that 'noise shadow' is much less - as expected.

73,

Bill KU8H

On 07/22/2018 10:29 AM, Jerry Gaffke via Groups.Io wrote:
You got it right.
Raising the BFO frequency brings it closer to signals coming from the
filter, which lowers the audio frequencies..
We have a BFO below the 12mhz crystal filter passband which mixes (at
D5,T7) with signals in that passband creating audio.
Lets give it some numbers, the numbers for your particular rig might be
a bit different.
I'm assuming we are in receive mode, transmit is the same math but in
the opposite direction.

The BFO frequency in the stock firmware is at 11996500 hz.
The 12mhz crystal filter has a roughly 2000 hz 3dB passband, something
like 11997000 to 11999000.
A signal coming in through the crystal filter on the bottom edge of the
passband creates an audio tone of 11997000-11996500 = 500 hz.
A signal coming in through the crystal filter on the top edge of the
passband creates an audio tone of 11999000-11996500 = 2500hz.

If we now raise the BFO frequency from 11996500 to 11996700 hz,
the audio coming through would fall between 11997000-1196700 = 300 hz
and 1199900-1199700 = 2300 hz.

Jerry


On Sun, Jul 22, 2018 at 06:19 AM, RCC WB5YYM wrote:

With my radio/engineering skills very much lacking, I probably don't
even need to be trying to contribute to this conversation, but here
is what I ran into on my build. My tx audio was very distorted when
I first tested. I then checked the transmit using the tones sent
from WSJTx. I found that no tones were passing that were around
500hz or lower, but higher tones around 1600hz caused more power to
be output from the radio. I adjusted the BFO until I had maximum
power out when sending a 1khz tone. This gave me power out in a
range from 200hz to 1600hz. This seemed to clean up the audio on
SSB. Apparently I have a very narrow filter, but I am getting good
audio reports on SSB. The down side is that with it being so narrow,
when operating FT8, I need to shift the VFO frequency to see all the
transmissions. If I remember right, to shift the audio frequency
down, I had to adjust the BFO frequency up, but please don't hold me
to this statement . I did adjust about 100hz at a time, and checked
the power out using different frequencies generated by WSJTx. I
don't exactly know what the above means, but it did help me to have
a good working radio. Hope this helps someone.

--
bark less - wag more

jim
 



On Sunday, July 22, 2018, 6:20:06 AM PDT, RCC WB5YYM <curtis03@...> wrote:


With my radio/engineering skills very much lacking, I probably don't even need to be trying to contribute to this conversation,


Just a note ...

This is what the recieve passband looked like on my ubitx as recieved from the 'factory'

Jim