Topics

CW transmit bandwidth #ubitx #ubitxcw

Gary Schwartz
 

Just got a v3 board running today.  I was monitoring CW on my IC7300 and noticed that when transmitting, i.e., a string of dots or dashes, the bandwidth would widen to 1Khz then return to very narrow.   This occurred at a periodic rate if I held the dot or dash lever.  Internal keyer was used.  When sending text the same issue occurred but it was more difficult to spot.  Initial test was 40 meters, 12 watts out, but same results on 20, 8 watts out all into a dummy. 12.2VDC power supply. 

Any clues to why this happens and what can be done?

Gary K3OS

Jerry Gaffke
 

So it's narrow when the key is down continuously, but a string of dots is 1khz wide?
That would be envelope shaping.  Has to do with how quickly the transmitter turns on and off when you key it.
Here's somebody who has looked at this:
    https://groups.io/g/BITX20/message/51423


On Thu, Jun 7, 2018 at 12:37 pm, K3OS wrote:
Just got a v3 board running today.  I was monitoring CW on my IC7300 and noticed that when transmitting, i.e., a string of dots or dashes, the bandwidth would widen to 1Khz then return to very narrow.   This occurred at a periodic rate if I held the dot or dash lever.  Internal keyer was used.  When sending text the same issue occurred but it was more difficult to spot.  Initial test was 40 meters, 12 watts out, but same results on 20, 8 watts out all into a dummy. 12.2VDC power supply. 

Any clues to why this happens and what can be done?

Gary K3OS

 

I agree with Jerry that it sounds like maybe the CW keying is too sharp.  That would explain a wider than expected bandwidth. 

I haven't really looked at the uBITX schematic in any detail so I can't comment on what is being done with regards to CW Waveform shaping.
I just unpacked my uBITX V3 board and I am starting to look at initial wire up. Once it is working I can do a similar investigation to what I did for the Bitx40.

Cheers

Michael VE3WMB 

Gary Schwartz
 

The issue is not with the keyed waveform but an undulating bandwidth change.  Using the internal keyer hold the dot paddle (much easier to see than with dashes).  The bandwidth starts out properly narrow, but after a second or so it quickly spreads to 1KHz+, then back to normal.  If you continue to hold the paddle this pattern constantly repeats.  When you send words the same thing happens, but it is more difficult to discern.

Gary


Virus-free. www.avast.com

Tim Gorman
 

How did you measure the bandwidth?

Remember that a square wave is a infinite series of odd harmonics. Your
keying is basically a square wave unless modified to have a rounded
up-slope and down-slope. Even then you will still have some odd
harmonics that cause the bandwidth of a repeating on-off dot or dash to
have a bandwidth increase over just the carrier itself.

tim ab0wr


On Thu, 07 Jun 2018 12:37:33 -0700
"K3OS" <garysch69@...> wrote:

Just got a v3 board running today.  I was monitoring CW on my IC7300
and noticed that when transmitting, i.e., a string of dots or dashes,
the bandwidth would widen to 1Khz then return to very narrow.   This
occurred at a periodic rate if I held the dot or dash lever.
Internal keyer was used.  When sending text the same issue occurred
but it was more difficult to spot.  Initial test was 40 meters, 12
watts out, but same results on 20, 8 watts out all into a dummy.
12.2VDC power supply. 

Any clues to why this happens and what can be done?

Gary K3OS

Allan Mason
 

This email is for those interested in using the transceiver for CW. 

Being mainly a CW operator, I was concerned about the V3 board uBitx transmit CW wave shape being very hard with an almost square waveshape and as expected, key clicks could be heard on either side of the signal. A look at the circuit diagram and a little tinkering today vastly improved the situation.

All that was required was to increase the value of C1 from 0.1uf to 1uf. Attached are some CRO photos of the end result showing the leading edge and trailing edge of the transmitter output with a 1uf connected across C1. The other photo shows where a short wire was soldered to the hot side of C1 for the tests.  In reality the wave shape is now a little soft on the trailing edge, so 0.047 or 0.68uf may be sufficient for some people.   More work could possibly be done with the CW keying RC network to the 1st balanced mixer, however this one component simple fix will suit my needs.

The tests were contucted using the internal keyer set to 25wpm. The CRO connection point was at the antenna BNC connector with a 50ohm dummy load.

Regards,
Allan Mason, JP
Mobile: +61 401 324 058


On Fri, Jun 8, 2018 at 10:58 PM, Tim Gorman <tgorman2@...> wrote:
How did you measure the bandwidth?

Remember that a square wave is a infinite series of odd harmonics. Your
keying is basically a square wave unless modified to have a rounded
up-slope and down-slope. Even then you will still have some odd
harmonics that cause the bandwidth of a repeating on-off dot or dash to
have a bandwidth increase over just the carrier itself.

tim ab0wr


On Thu, 07 Jun 2018 12:37:33 -0700
"K3OS" <garysch69@...> wrote:

> Just got a v3 board running today.  I was monitoring CW on my IC7300
> and noticed that when transmitting, i.e., a string of dots or dashes,
> the bandwidth would widen to 1Khz then return to very narrow.   This
> occurred at a periodic rate if I held the dot or dash lever.
> Internal keyer was used.  When sending text the same issue occurred
> but it was more difficult to spot.  Initial test was 40 meters, 12
> watts out, but same results on 20, 8 watts out all into a dummy.
> 12.2VDC power supply. 
>
> Any clues to why this happens and what can be done?
>
> Gary K3OS






Virus-free. www.avg.com

Jerry Gaffke
 

Cool that all it took was a larger C1, that's a good result.

From the scope traces, the rising edge is much faster than the falling edge.
This is because the CW-KEY line from the nano is rising exponentially from 0v to a 5v asymptote
it is travelling very quickly through that first fraction of a volt where the mixer becomes unbalanced.
On the trailing edge, CW-KEY is starts travelling quickly as it makes its way exponentially down from 5v,
slowing as it asymptotically approaches 0v.

If you want the rising and falling edges to be symetrical,
try adding a 1k pot across C1, this forms a voltage divider with R104 so you can adjust
the maximum CW-KEY voltage that gets presented to C1.
Adjust the pot for symmetry in your scope display, then measure the pot and replace it with a fixed resistor.
resistor across C1.   First suggested in post  https://groups.io/g/BITX20/message/50445
 


On Fri, Jun 8, 2018 at 08:48 pm, Allan Mason wrote:
All that was required was to increase the value of C1 from 0.1uf to 1uf.

Jerry Gaffke
 

These are inverse exponential responses, it won't be symmetrical.
But the pot should help move it toward vaguely similar rise and fall rates.

The integrator mentioned in post 50445 is capable of creating a very nice trapezoid.
We had a discussion on this going in the AT-Sprint group over a decade ago:
    https://groups.yahoo.com/neo/groups/AT_Sprint/conversations/messages/1086
where we found a way to make the rise and fall slopes of the trapezoid equal,
and also how to make the slopes constant when the power supply voltage varied.
Way overkill, most of us should just stick with the cap and resistor.

 

On Fri, Jun 8, 2018 at 09:17 pm, Jerry Gaffke wrote:
If you want the rising and falling edges to be symetrical,

Tim Gorman
 

Nice job!

Did you try to capture two consective dits? Just to make sure the
trailing edge was being completed well before the next beginning edge?
If it is then just adding this capacitor would work for most purposes.

Thanks,

tim ab0wr

On Sat, 9 Jun 2018 13:48:04 +1000
"Allan Mason" <allan.vk2gr@...> wrote:

This email is for those interested in using the transceiver for CW.

Being mainly a CW operator, I was concerned about the V3 board uBitx
transmit CW wave shape being very hard with an almost square
waveshape and as expected, key clicks could be heard on either side
of the signal. A look at the circuit diagram and a little tinkering
today vastly improved the situation.

All that was required was to increase the value of C1 from 0.1uf to
1uf. Attached are some CRO photos of the end result showing the
leading edge and trailing edge of the transmitter output with a 1uf
connected across C1. The other photo shows where a short wire was
soldered to the hot side of C1 for the tests. In reality the wave
shape is now a little soft on the trailing edge, so 0.047 or 0.68uf
may be sufficient for some people. More work could possibly be done
with the CW keying RC network to the 1st balanced mixer, however this
one component simple fix will suit my needs.

The tests were contucted using the internal keyer set to 25wpm. The
CRO connection point was at the antenna BNC connector with a 50ohm
dummy load.

Regards,
Allan Mason, JP
Mobile: +61 401 324 058


On Fri, Jun 8, 2018 at 10:58 PM, Tim Gorman <tgorman2@...> wrote:

How did you measure the bandwidth?

Remember that a square wave is a infinite series of odd harmonics.
Your keying is basically a square wave unless modified to have a
rounded up-slope and down-slope. Even then you will still have some
odd harmonics that cause the bandwidth of a repeating on-off dot or
dash to have a bandwidth increase over just the carrier itself.

tim ab0wr


On Thu, 07 Jun 2018 12:37:33 -0700
"K3OS" <garysch69@...> wrote:

Just got a v3 board running today. I was monitoring CW on my
IC7300 and noticed that when transmitting, i.e., a string of dots
or dashes, the bandwidth would widen to 1Khz then return to very
narrow. This occurred at a periodic rate if I held the dot or
dash lever. Internal keyer was used. When sending text the same
issue occurred but it was more difficult to spot. Initial test
was 40 meters, 12 watts out, but same results on 20, 8 watts out
all into a dummy. 12.2VDC power supply.

Any clues to why this happens and what can be done?

Gary K3OS



<http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
Virus-free.
www.avg.com
<http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>