Topics

Is push pull hf amp has higher spectral purity

Simon
 

Is push pull generally has higher spectral purity than single end transister ?   Ubitx is push pull, bitx40 is single end.

I read comments stating that bitx40 has better spectural than ubitx, because it is single band, simpler circuit. Is it true for present batch, revision?  I meant, the comments could refer to many revisions over time.  

How "bad" is the situation? It is not too far from fcc spec., right? There were quite a number of posting on spectural purity? Is the latest batch solved the issue now. 

73 Simon

iz oos
 

The push pull has the advantage to deeply attenuate the even harmonics, all versions of the ubitx were good in that, the issue was on odd harmonics, 3rd, 5th.... With push pull a less steeper low pass filter can be used. Especially the 'corner'  frequency can be chosen higher than it would be otherwise. A single low pass filter can also be used for a couple bands. So you save a lot in filtering and relays, it's more a matter of cost and convenience I believe. Monoband qrps are much easier, the ones I had assembled don't have a push pull final, just a normal amplifier and a low pass filter to remove any harmonics. Maybe even I could design a monoband transceiver...


Il 02/dic/2018 20:18, "Simon" <simonhk7@...> ha scritto:
Is push pull generally has higher spectral purity than single end transister ?   Ubitx is push pull, bitx40 is single end.

I read comments stating that bitx40 has better spectural than ubitx, because it is single band, simpler circuit. Is it true for present batch, revision?  I meant, the comments could refer to many revisions over time.  

How "bad" is the situation? It is not too far from fcc spec., right? There were quite a number of posting on spectural purity? Is the latest batch solved the issue now. 

73 Simon

ajparent1/kb1gmx
 

Simon, 

The structure of the amp is not the sole element of the question.  I have a single ended that beats the push-pull
of the ubitx by miles.  I also have a QSX amp that is push pull and is very clean.  In the end its not if A or B
but how well the design was executed.

Right now the V4 ubitx solves no issues over the V3 save for now it does not use the TDA2822 audio amplifier chip.

Allison

Simon
 

So, for current version that if i buy today, is it basically true that the overall story is 

1. btx40 is just meeting fcc spec, i believe is 50dBc for lower freq,

2. Ubtx is just a bit short, but not too far

73 Simon


On Monday, December 3, 2018, ajparent1/KB1GMX <kb1gmx@...> wrote:
Simon, 

The structure of the amp is not the sole element of the question.  I have a single ended that beats the push-pull
of the ubitx by miles.  I also have a QSX amp that is push pull and is very clean.  In the end its not if A or B
but how well the design was executed.

Right now the V4 ubitx solves no issues over the V3 save for now it does not use the TDA2822 audio amplifier chip.

Allison

MadRadioModder
 

Garbage in, garbage out. 


MRM

 


On Dec 2, 2018, at 3:40 PM, iz oos <and2oosiz2@...> wrote:

The push pull has the advantage to deeply attenuate the even harmonics, all versions of the ubitx were good in that, the issue was on odd harmonics, 3rd, 5th.... With push pull a less steeper low pass filter can be used. Especially the 'corner'  frequency can be chosen higher than it would be otherwise. A single low pass filter can also be used for a couple bands. So you save a lot in filtering and relays, it's more a matter of cost and convenience I believe. Monoband qrps are much easier, the ones I had assembled don't have a push pull final, just a normal amplifier and a low pass filter to remove any harmonics. Maybe even I could design a monoband transceiver...


Il 02/dic/2018 20:18, "Simon" <simonhk7@...> ha scritto:
Is push pull generally has higher spectral purity than single end transister ?   Ubitx is push pull, bitx40 is single end.

I read comments stating that bitx40 has better spectural than ubitx, because it is single band, simpler circuit. Is it true for present batch, revision?  I meant, the comments could refer to many revisions over time.  

How "bad" is the situation? It is not too far from fcc spec., right? There were quite a number of posting on spectural purity? Is the latest batch solved the issue now. 

73 Simon


--

…_. _._

Brian
 

The stock Bitx40 has a problem with the 7.2MHz birdie not only on receive but also it produces a spur on transmit which does not meet fcc spec.
You could use it on 7.1MHz and below or change the vfo to high side and move the carrier oscillator to the other side of the filter (Allard's software). Whether you can do that depends on the xtal filter in your Bitx40. You may get lots of carrier and poor sideband rejection.
The Bitx is a good entry into homebrew. You might get away with using it barefoot but take care if you want to push to higher power.
73 Brian.

Jerry Gaffke
 

Brian,

So when operating between 7.1 and 7.3 mhz on the Bitx40 we have a transmit spur at 7.2mhz?
That would be weird, especially if the spur didn't move as we tuned across that range.

The receive birdie at 7.2mhz is due to the  12mhz - 7.2mhz = 4.8mhz low side VFO
having a 5'th harmonic at 5*4.8=24mhz which beats with the second harmonic of the 12mhz BFO.

Anybody see any way that this might go out as a transmit spur?
Has anybody else seen this spur?

Jerry, KE7ER


On Mon, Dec 3, 2018 at 01:23 PM, Brian wrote:
The stock Bitx40 has a problem with the 7.2MHz birdie not only on receive but also it produces a spur on transmit which does not meet fcc spec.
You could use it on 7.1MHz and below or change the vfo to high side and move the carrier oscillator to the other side of the filter (Allard's software). Whether you can do that depends on the xtal filter in your Bitx40. You may get lots of carrier and poor sideband rejection.
The Bitx is a good entry into homebrew. You might get away with using it barefoot but take care if you want to push to higher power.
73 Brian.

Rob Bleumer <Bleumer@...>
 

In Europe the birdey on 7.2Mhz is a nice end of band signal🤓
Rob

Jerry Gaffke
 

There is a way in which the Bitx40 can produce a spur, as Brian reports.
Look over the last few posts in this thread:
    https://groups.io/g/BITX20/message/21998

The thread has to do with images received when using a low side VFO on the Bitx40.

Example:  
When receiving 7.1mhz, the VFO is at 12 - 7.1 = 4.9mhz.  
The 4'th harmonic of the VFO is at 4 * 4.9 = 19.6mhz, effectively a high side VFO.
We will receive an image if there is a shortwave broadcaster at 19.6 - 12 = 7.6mhz

Not stated in that particular thread, but this can happen on transmit as well.
When transmitting at 7.1mhz, we will have a faint spur at 7.6mhz.
When transmitting at 7.21mhz, we will have a faint spur at     4*(12 - 7.21) - 12 = 7.16mhz
Note that this transmitted spur can easily pass through the 7mhz bandpass filter
and the transmit low pass filter..
The only factor that keeps the spur in check is that the 4'th harmonic of the VFO 
is much weaker than the fundamental of the VFO, and this may vary from board to board,
the VFO drive level programmed into the si5351, and how the VFO wires get routed.

The curious thing is that when transmitting at 7.2mhz, the spur is at 7.2 mhz.
The spur is a valid SSB signal, but on the opposite sideband.
So there's a vague connection with the 7.2mhz Bitx40 birdie,
but the math behind the two is quite different.

I'd say the best solution is to use Allard's v2 firmware and mods, 
stay with a high side VFO up around 19.2mhz instead of down at 4.8mhz.
  https://github.com/amunters/bitx40-raduino-v2
As Brian points out, this cures the spur, but when transmitting an LSB signal
we have the not-so-sharp skirt of the 12mhz crystal filter facing the carrier and opposite sideband.
So we will have additional carrier leakage, and the opposite sideband is not suppressed quite as well.
However, that's a perfectly legal signal anywhere AM phone is allowed.
Especially if only transmitting 5 or 10 watts.

I am of the opinion that if transmitting at 5 or 10 watts, the spur with a stock low side VFO
is likely down far enough to be more or less acceptable.  
But to fully meet regulations with a Bitx40, use Allard's v2 firmware and mods, and a high side VFO.

Jerry, KE7ER


On Tue, Dec 4, 2018 at 07:36 PM, Jerry Gaffke wrote:
Brian,

So when operating between 7.1 and 7.3 mhz on the Bitx40 we have a transmit spur at 7.2mhz?
That would be weird, especially if the spur didn't move as we tuned across that range.

The receive birdie at 7.2mhz is due to the  12mhz - 7.2mhz = 4.8mhz low side VFO
having a 5'th harmonic at 5*4.8=24mhz which beats with the second harmonic of the 12mhz BFO.

Anybody see any way that this might go out as a transmit spur?
Has anybody else seen this spur?

Jerry, KE7ER


toggle quoted message. . .

 

On Mon, Dec 3, 2018 at 01:23 PM, Brian wrote:
The stock Bitx40 has a problem with the 7.2MHz birdie not only on receive but also it produces a spur on transmit which does not meet fcc spec.
You could use it on 7.1MHz and below or change the vfo to high side and move the carrier oscillator to the other side of the filter (Allard's software). Whether you can do that depends on the xtal filter in your Bitx40. You may get lots of carrier and poor sideband rejection.
The Bitx is a good entry into homebrew. You might get away with using it barefoot but take care if you want to push to higher power.
73 Brian.

Jerry Gaffke
 

Regarding the Bitx40 spur, I previously wrote:

>  I'd say the best solution is to use Allard's v2 firmware and mods, 
>  stay with a high side VFO up around 19.2mhz instead of down at 4.8mhz.

It's possible that a low pass filter on the VFO is sufficient, allowing through 4.8mhz
but not the harmonic at 19.2 mhz.  However, the diode ring mixer is non-linear,
might somehow recreate that 4'th harmonic for us.

Jerry, KE7ER


On Wed, Dec 5, 2018 at 09:56 AM, Jerry Gaffke wrote:
There is a way in which the Bitx40 can produce a spur, as Brian reports.
Look over the last few posts in this thread:
    https://groups.io/g/BITX20/message/21998

The thread has to do with images received when using a low side VFO on the Bitx40.

Example:  
When receiving 7.1mhz, the VFO is at 12 - 7.1 = 4.9mhz.  
The 4'th harmonic of the VFO is at 4 * 4.9 = 19.6mhz, effectively a high side VFO.
We will receive an image if there is a shortwave broadcaster at 19.6 - 12 = 7.6mhz

Not stated in that particular thread, but this can happen on transmit as well.
When transmitting at 7.1mhz, we will have a faint spur at 7.6mhz.
When transmitting at 7.21mhz, we will have a faint spur at     4*(12 - 7.21) - 12 = 7.16mhz
Note that this transmitted spur can easily pass through the 7mhz bandpass filter
and the transmit low pass filter..
The only factor that keeps the spur in check is that the 4'th harmonic of the VFO 
is much weaker than the fundamental of the VFO, and this may vary from board to board,
the VFO drive level programmed into the si5351, and how the VFO wires get routed.

The curious thing is that when transmitting at 7.2mhz, the spur is at 7.2 mhz.
The spur is a valid SSB signal, but on the opposite sideband.
So there's a vague connection with the 7.2mhz Bitx40 birdie,
but the math behind the two is quite different.

I'd say the best solution is to use Allard's v2 firmware and mods, 
stay with a high side VFO up around 19.2mhz instead of down at 4.8mhz.
  https://github.com/amunters/bitx40-raduino-v2
As Brian points out, this cures the spur, but when transmitting an LSB signal
we have the not-so-sharp skirt of the 12mhz crystal filter facing the carrier and opposite sideband.
So we will have additional carrier leakage, and the opposite sideband is not suppressed quite as well.
However, that's a perfectly legal signal anywhere AM phone is allowed.
Especially if only transmitting 5 or 10 watts.

I am of the opinion that if transmitting at 5 or 10 watts, the spur with a stock low side VFO
is likely down far enough to be more or less acceptable.  
But to fully meet regulations with a Bitx40, use Allard's v2 firmware and mods, and a high side VFO.

Jerry, KE7ER

Brian
 

Jerry,
I agree with what you have written and detailed. The transmit spur varies with frequency. Away from 7.2MHz it is outside of the band. Feeding the Raduino into the disabled vfo is not optimal as regards signal purity.
The start of this thread was a suggestion that the Bitx40 might be easier to get fcc compliant. Being single band that might be true but it is not without its own problems. There has been little reported on the Bitx40 since the uBitx was released.
Brian VK4BAP

Guy WB7SZI
 

Have you seen QST’s review of the BitX40? Seemed to pass their analysis. 
I think that was with an unmodded radio. 

Guy WB7SZI 

Jerry Gaffke
 

I think there was brief mention by somebody in early 2017 about that spur getting transmitted
now that Brian has jogged my memory. 
But mentioned only once more or less in passing, and then it was dropped.  
I think.

But I don't think anybody ever got around to measuring the spur,
Maybe Brian knows more about this.

Possible that the spur strength varies from rig to rig. 
Whatever might effect the strength of the VFO's 4'th harmonic.

I personally don't think it's a significant issue.
And if it were, going to Allard's v2 firmware and mods is a very easy fix.

Jerry


On Wed, Dec 5, 2018 at 05:36 PM, Guy WB7SZI wrote:
Have you seen QST’s review of the BitX40? Seemed to pass their analysis. 
I think that was with an unmodded radio. 

Guy WB7SZI 

Brian
 

From memory, I measured the spur to be 35dB down on the 7.15MHz signal. Nobody else reported measuring it which surprised me since quite a bit was written about modding the LPF to meet fcc on harmonics.
The 12MHz xtal filter in my rig was very bad on the LF side, perhaps one xtal is slightly off frequency. I did consider building a new 12MHz filter but never got around to it. I have several home built filters at 8.86MHz but they are just a bit too big to fit in.
I decided to leave it as is and only use USB with digital.
Brian.

Jerry Gaffke
 

Brian,

They were all busy looking at the harmonics, didn't think to check for that close in spur.
Would be interesting to see how strong it is on other rigs.

I believe the strength of the spur comes down to how strong that 4'th harmonic of the VFO is.
During transmit, the VFO mixes in after the signal passes through the crystal filter,
so a lousy crystal filter is not contributing to the spur.

Lesson learned:  Use high side local oscillators wherever possible.

Jerry


On Thu, Dec 6, 2018 at 02:56 PM, Brian wrote:
From memory, I measured the spur to be 35dB down on the 7.15MHz signal. Nobody else reported measuring it which surprised me since quite a bit was written about modding the LPF to meet fcc on harmonics.
The 12MHz xtal filter in my rig was very bad on the LF side, perhaps one xtal is slightly off frequency. I did consider building a new 12MHz filter but never got around to it. I have several home built filters at 8.86MHz but they are just a bit too big to fit in.
I decided to leave it as is and only use USB with digital.