Date   
Re: 80 Meters Harmonics Fix Proposal

Howard Fidel
 

Gordon:
Thanks for your concurrence. To prevent others from being confused,
I am raising not lowering the capacitance, which I am sure is what you meant.
And yes, I am trying to come up with an easy fix that all can implement including HF signals.

Howard

On 9/3/2018 4:05 PM, Gordon Gibby wrote:
Howard, I get your point. And a six DB improvement makes the device a much better transceiver per the regulations

With simple unshielded wiring to three external relays, in my tests I found the fifth harmonic of 80 meter CW was basically obliterated, but the seventh only went down by about six dB.  (I don’t have an explanation for the disparity ) I haven’t tried any shielded wiring to see if that would make any better — but it certainly might. 


In my case It might be simpler to add in a permanent low pass filter above 15 MHZ in my case.     Your idea of lowering the capacitance is interesting.  And it’s pretty simple also!   For many people that might be the easiest solution to the harmonic problem. 




On Sep 3, 2018, at 15:47, Howard Fidel <sonic1@...> wrote:

Warren:
The fifth and the 7th are at -45dB now. You are missing my point. The design of the filter directly effects the level of the harmonics. Yes, the stray  capacitance creates a "sneak" path around the filter. However, the input and output capacitors of the filter act as a divider to that "sneak" signal. If you double the output capacitance, you 1/2 the sneak signals amplitude, reducing it 6 dB. If you create a model of each  of the filters in a simulator, and put 8 pF from input to output, you can clearly see the improvement in the attenuation of the filter for the higher output capacitance.
All the data I have seen shows that a 6 dB improvement would just make the uBitx legal.

Howard

On 9/3/2018 2:51 PM, Warren Allgyer wrote:
Howard

When I measured 80 meter harmonics the particularly troublesome ones were 5th at 18 MHz and 7th at 25 MHz. In both cases the attenuation of the stock uBitx at those frequencies was limited by the layout of the board and by the coupling of relays sharing both filter input and output in the same frame. In such cases the characteristics of the filters, both existing and any modifications are immaterial. The attenuation of the stock filter supplied is more than adequate. But the harmonics bypass the filter and go via I/O coupling straight to the output. Please measure the 5th and 7th harmonics on CW. I think you will find you have not done much to improve the stock situation.

WA8TOD



Re: K5BCQ uBITX Relay Switched LPF/BPF board

Kees T
 

The QRP Labs LPF filters are rated at 10W according to Hans Summers, the BPF filters he sells are not, probably because of the series current the capacitors have to handle or larger inductors required. No reason you can't make up BPFs which CAN handle 10W. The filters you see in the pictures are blank boards I made up and you add what components you need to add, and design your own. Could come out with a different BPF filter layout if needed. With those little 1.5" x 0.5" boards you can get 16 per 100mm x 100mm panel .....that's 160 boards per 10 prototype panel order for only $2 plus $5? for scoring plus shipping.

73 Kees K5BCQ  

Re: uBitx Unfiltered

ajparent1/KB1GMX
 

Start with 1V is adequate as the dive is Current not Voltage for transistors.

Altering the resistors make the previous stage a bit happier and lower the loss
to the resistors but the basic issue is crappy transistors, transformers, and layout.

For L8 and L9 the common practice is bifiler wind on a common core 
so the current balance out to zero flux in the core. I used 2t bufilar #26
on FB43-202 to get about the same result.   Then again anything
did help compared to existing.  The driver and predriver transformers are
not so good either.  While your at it the q90 transformer is not as good
as could be.  T11 rewind it as something reasonable for 10mhz and up.
try 2:3 or even 2:2.  For better IMD run about 150 to 200ma per transistor
on the finals.

Trying to get more than 1.8W on 10 with basically 10W at lower end is repeating
the work I did 2.5 months ago.  Ditch the 3904s put in 2222s tell me that is not
remarkable difference.  Or go straight to 3866s or 5109s.  FYI everyone found
the MPSH10s to be rather marginal improvement.

Then add RLC feedback to the driver and Predriver to keep the gain fade
with frequency a bit flatter.  With that you should be to 12W on 80 and 40
and about 5-6W on 10M.  The spur at 17mhz (10M) will now plague you.

Reset the bias to Q90, its a low level stage and even 10ma is high.
some of them are running at over 25ma for 1 milliwatt or RF.  BEst way
1K is series with a 2K with a parallel .01uf chip. The idea is 3K DC R
and 1K AC R.  Use a real transistor there too MMBR2369 is better
(flat to about 22mhz)  BFR106 (flat pas 30mhz)  work very well.
A mmbt5179 should work if kept under 10ma as well.

You will have to reduce the extremely high standing current of the
predriver and driver.  The 22ohm emitter resistors tend to cause voltage
limiting on the pre-driver and driver as well as much to great emitter
degeneration.  Using 8.2 to 10 ohms improved that.  Bypassing them with
220/330/470 pf helps get some upper frequency peaking.

Flattening the power was more or less doable.but spurs and harmonics made
it worthless.  Also if the gain gets to about 60db at 20mhz and up the amp is then
subject to oscillation from ground loops extending back to the 33mhz low pass filter.
the fix is slice the board and separate. There are secondary loops with DC power
lines having RF (TX-DC+) due to parallel paths under the relays and near the
power amp.   Look at Kees picture of the board without relays. 

 Allison

Re: 80 Meters Harmonics Fix Proposal

Gordon Gibby
 

Yep, sorry, you were lowering the impedance, raising the capacitance.   Slip of the tongue. (I use Siri)


You might tell us how you changed each value in those filters.   





On Sep 3, 2018, at 16:20, Howard Fidel <sonic1@...> wrote:

Gordon:
Thanks for your concurrence. To prevent others from being confused,
I am raising not lowering the capacitance, which I am sure is what you meant.
And yes, I am trying to come up with an easy fix that all can implement including HF signals.

Howard
On 9/3/2018 4:05 PM, Gordon Gibby wrote:
Howard, I get your point. And a six DB improvement makes the device a much better transceiver per the regulations

With simple unshielded wiring to three external relays, in my tests I found the fifth harmonic of 80 meter CW was basically obliterated, but the seventh only went down by about six dB.  (I don’t have an explanation for the disparity ) I haven’t tried any shielded wiring to see if that would make any better — but it certainly might. 


In my case It might be simpler to add in a permanent low pass filter above 15 MHZ in my case.     Your idea of lowering the capacitance is interesting.  And it’s pretty simple also!   For many people that might be the easiest solution to the harmonic problem. 




On Sep 3, 2018, at 15:47, Howard Fidel <sonic1@...> wrote:

Warren:
The fifth and the 7th are at -45dB now. You are missing my point. The design of the filter directly effects the level of the harmonics. Yes, the stray  capacitance creates a "sneak" path around the filter. However, the input and output capacitors of the filter act as a divider to that "sneak" signal. If you double the output capacitance, you 1/2 the sneak signals amplitude, reducing it 6 dB. If you create a model of each  of the filters in a simulator, and put 8 pF from input to output, you can clearly see the improvement in the attenuation of the filter for the higher output capacitance.
All the data I have seen shows that a 6 dB improvement would just make the uBitx legal.

Howard

On 9/3/2018 2:51 PM, Warren Allgyer wrote:
Howard

When I measured 80 meter harmonics the particularly troublesome ones were 5th at 18 MHz and 7th at 25 MHz. In both cases the attenuation of the stock uBitx at those frequencies was limited by the layout of the board and by the coupling of relays sharing both filter input and output in the same frame. In such cases the characteristics of the filters, both existing and any modifications are immaterial. The attenuation of the stock filter supplied is more than adequate. But the harmonics bypass the filter and go via I/O coupling straight to the output. Please measure the 5th and 7th harmonics on CW. I think you will find you have not done much to improve the stock situation.

WA8TOD



Re: uBitx Unfiltered

Gordon Gibby
 

Thank you, Alyson, for even more education

I’m just about finished building a UHF repeater for emergency work, and I’ll be able to turn back to this rig at some point.  I don’t need it to work past 20 m, so little bitty improvements will suffice for my needs.  I don’t have quite the needs that others have.


Gordon



On Sep 3, 2018, at 16:28, ajparent1/KB1GMX <kb1gmx@...> wrote:

Start with 1V is adequate as the dive is Current not Voltage for transistors.

Altering the resistors make the previous stage a bit happier and lower the loss
to the resistors but the basic issue is crappy transistors, transformers, and layout.

For L8 and L9 the common practice is bifiler wind on a common core 
so the current balance out to zero flux in the core. I used 2t bufilar #26
on FB43-202 to get about the same result.   Then again anything
did help compared to existing.  The driver and predriver transformers are
not so good either.  While your at it the q90 transformer is not as good
as could be.  T11 rewind it as something reasonable for 10mhz and up.
try 2:3 or even 2:2.  For better IMD run about 150 to 200ma per transistor
on the finals.

Trying to get more than 1.8W on 10 with basically 10W at lower end is repeating
the work I did 2.5 months ago.  Ditch the 3904s put in 2222s tell me that is not
remarkable difference.  Or go straight to 3866s or 5109s.  FYI everyone found
the MPSH10s to be rather marginal improvement.

Then add RLC feedback to the driver and Predriver to keep the gain fade
with frequency a bit flatter.  With that you should be to 12W on 80 and 40
and about 5-6W on 10M.  The spur at 17mhz (10M) will now plague you.

Reset the bias to Q90, its a low level stage and even 10ma is high.
some of them are running at over 25ma for 1 milliwatt or RF.  BEst way
1K is series with a 2K with a parallel .01uf chip. The idea is 3K DC R
and 1K AC R.  Use a real transistor there too MMBR2369 is better
(flat to about 22mhz)  BFR106 (flat pas 30mhz)  work very well.
A mmbt5179 should work if kept under 10ma as well.

You will have to reduce the extremely high standing current of the
predriver and driver.  The 22ohm emitter resistors tend to cause voltage
limiting on the pre-driver and driver as well as much to great emitter
degeneration.  Using 8.2 to 10 ohms improved that.  Bypassing them with
220/330/470 pf helps get some upper frequency peaking.

Flattening the power was more or less doable.but spurs and harmonics made
it worthless.  Also if the gain gets to about 60db at 20mhz and up the amp is then
subject to oscillation from ground loops extending back to the 33mhz low pass filter.
the fix is slice the board and separate. There are secondary loops with DC power
lines having RF (TX-DC+) due to parallel paths under the relays and near the
power amp.   Look at Kees picture of the board without relays. 

 Allison

Re: K5BCQ uBITX Relay Switched LPF/BPF board

Kees T
 

David,

Manual switching is only one of the options. There is also a small board you can add which can mux the 3 bits from the Raduino to all 6 sets of relays......but it would require a small code change. I was suggesting to use the codes 111,110,100, and 000 which the existing Raduino code supplies via TxA, TxB, TxC to select the 4 LPFs as today and add the remaining 2 when you get around to it. There is also no magic to 6x filters....it could be 5x or 8x filters or 12x filters if you add another select bit. 6x happens to fit readily on a 100mm x100mm panel.

The 4x LPF board version decodes the existing TxA,TxB, and TxC drivers just like the uBITX does today (and I have to reinstall K3). Those boards should be here next week.

73 kees K5BCQ

Re: K5BCQ uBITX Relay Switched LPF/BPF board

Kees T
 

You can also see, in the picture with the uBITX board, that the 2N3904's have already been pulled, waiting for the 2N2222A's ..............and other incremental improvements.  

73 Kees K5BCQ

Re: K5BCQ uBITX Relay Switched LPF/BPF board

Kees T
 

You can also  modify the existing filter for a faster roll-off which MAY catch enough of the 15m spur .....for instance. The plot labeled "1" is the original LPF, the other plot is the modified LPF.

73 Kees K5BCQ

Re: 80 Meters Harmonics Fix Proposal

Warren Allgyer <allgyer@...>
 

Got it Howard. And I agree with both the strategy and the necessary amount of improvement. On mine, a six dB improvement would make all bands legal for CW harmonics. That would be a nice solution for a CW only radio. Of course you must not use it on SSB for 20 meters and above but perhaps that is not your objective. Good work!

 

WA8TOD

Re: K5BCQ uBITX Relay Switched LPF/BPF board

MadRadioModder
 

Let’s do it…

 

 

From: BITX20@groups.io [mailto:BITX20@groups.io] On Behalf Of Kees T
Sent: Monday, September 3, 2018 3:26 PM
To: BITX20@groups.io
Subject: Re: [BITX20] K5BCQ uBITX Relay Switched LPF/BPF board

 

The QRP Labs LPF filters are rated at 10W according to Hans Summers, the BPF filters he sells are not, probably because of the series current the capacitors have to handle or larger inductors required. No reason you can't make up BPFs which CAN handle 10W. The filters you see in the pictures are blank boards I made up and you add what components you need to add, and design your own. Could come out with a different BPF filter layout if needed. With those little 1.5" x 0.5" boards you can get 16 per 100mm x 100mm panel .....that's 160 boards per 10 prototype panel order for only $2 plus $5? for scoring plus shipping.

73 Kees K5BCQ  


Virus-free. www.avg.com

--

…_. _._

Re: uBitx Unfiltered

ajparent1/KB1GMX
 

On Mon, Sep 3, 2018 at 07:25 AM, Michael Babineau wrote:
This got me thinking that for another $20 I could just order the QRP Labs 5W PA kit and build a separate 160m to 10m CW TX with raised cosine wave shaping.
I do mostly CW anyway so giving up SSB wouldn't be a big loss for me.
Why use linear amp with lower efficiency for CW?  Class C can be shaped for keying and offer much mroe power with far less consumed.  However binding the two ubitx to a QSX amp is not easy.

>>Has anyone else thought  of just designing  a separate  TX board  for use with the V3/V4 ubitx? I know that this sounds like giving up ... but in many ways this could be a simpler solution than attempting major surgery on the existing ubitx.

Basically to do that you need to build a replacement fo the rear half of the board.  The problem is fourfold.
No input filtering for the amplifier,
The amplifiers inability to product power at 10M
low pass filter layout (filters would work otherwise)
Board layout.

Allison
We want a magik pill to fix is all.

Re: K5BCQ uBITX Relay Switched LPF/BPF board

Guy WB7SZI
 

Forgive me if this is way over simplified. I’m not an RF designer, 
but couldn’t you manually just insert an appropriate LPF from your ant. connection to the antenna to reduce harmonics on that specific band? I agree that a six band LPF system would be great but I for one only use one or two bands most of the time. Is that doable? 
73,
Guy WB7SZI 

Re: K5BCQ uBITX Relay Switched LPF/BPF board

ajparent1/KB1GMX
 

Kees,

Before you remount K3 remove the audio switching path from it.  Many get 
feedback from that path.  Also it can be inserted form the bottom.

Also the big issue is all the spaghetti traces. The LPF are good as is, just
the switching is a fail and its all layout.  Too many intereleaved lines.  The real
trouble maker is that all RF in to the filter and out from the filter go by was
of KT1.

In the drawing 3.jpeg The red is new wires yellow are bad wires.
This was built with relays on bottom and works well for 80 through 17m.
KT3 switches antenna to filter outptu or through coax direct to KT1 pin12.
Coax is under board.

In drawing 1jpg  The enclosed in red are the troublesome areas.
Yellow bad, green is audio mod, black means cut!, blue is the spaghetti to be removed/altered

That parallel DC lead (tx relay power) is parallel to RF line from KT3... If it looks like a
directional coupler it works like one. Its why there is a ton of RF on the switched 
DC+ line during TX.  Very bad!

If you don't care for above 17M the fix is messy but easy once the
relays are out.  Use one relay per band from 80/40/20 and it works.
RF goes in via KT1 and out via KT3.  Also modify the relays power
path to avoid that parallel line.  

Allison

Re: uBitx Unfiltered

iz oos
 

Regarding ' Has anyone else thought of just designing a separate TX board for use with the V3/V4 ubitx?' I agree completely considering that many have a separate receiver already.


Il 03/set/2018 16:26, "Michael Babineau" <mbabineau.ve3wmb@...> ha scritto:
W.r.t the current discussions on ubitx transmitter woes ... the route forward that I chose was to order LPFs plus the relay board from QRP Labs. My thoughts were to use a 2nd Arduino Nano clone configured as an I2C slave to allow the Raduino to communicate with the LPF board over I2C. This would require some very minor code changes to send band change info to the new filter board.

This solution would give me clean CW TX from 160m through 10m but does nothing to address the spur problem on the higher bands when using SSB nor the more recent concerns about other IMD. 

This got me thinking that for another $20 I could just order the QRP Labs 5W PA kit and build a separate 160m to 10m CW TX with raised cosine wave shaping.
I do mostly CW anyway so giving up SSB wouldn't be a big loss for me.

So now I am considering the option of just using my ubitx as an HF receiver. The RX works quite well so if I just forget the TX, at a little over $100 for my v3 ubitx is was still a great deal IMHO. 

Has anyone else thought  of just designing  a separate  TX board  for use with the V3/V4 ubitx? I know that this sounds like giving up ... but in many ways this could be a simpler solution than attempting major surgery on the existing ubitx.

I have yet to package my ubitx in a permanent enclosure so for now I am going to put it aside and wait a month or two to see if there is a better option.

Cheers 
Michael VE3WMB 




Re: 80 Meters Harmonics Fix Proposal

Howard Fidel
 

Warren:
I'm peeling the onion, slowly. I thought the problematic band was 15 Meters for SSB?

Gordon:
I don't understand what you are asking me. I provided the simulation file for the 80 meter filter. I haven't touched the other bands yet, I was hoping someone would also try the mod and confirm 80 meter CW is now OK.

Howard


On 9/3/2018 5:05 PM, Warren Allgyer wrote:

Got it Howard. And I agree with both the strategy and the necessary amount of improvement. On mine, a six dB improvement would make all bands legal for CW harmonics. That would be a nice solution for a CW only radio. Of course you must not use it on SSB for 20 meters and above but perhaps that is not your objective. Good work!

 

WA8TOD


Re: K5BCQ uBITX Relay Switched LPF/BPF board

ajparent1/KB1GMX
 

Now if one does all that and really wants 20 and up spurs and all....

Simple add an inline (between radio and antenna jack) 28mhz low pass filter.
and leave it there.  If it well made it will be invisible (low loss) and at worst
it catches harmonics like the old school TVI filter.  It does not fix the spur
Issue.

So the deal for the relays then is:
           
KT1 selects 20m filter via TXA
KT2 selects 40m filter via TXB
KT3 selects 80m filter Via TXC
none selected is straight through
and the external always inline 28mhz LPF is doing the work.

 **
 * Select the tx harmonic filters
*/  
void setTXFilters(unsigned long freq){
  
  if (freq >= 20000000L){ 
    digitalWrite(TX_LPF_A, 0);
    digitalWrite(TX_LPF_B, 0);  // 10m straight through
    digitalWrite(TX_LPF_C, 0);
  }
  else if (freq >= 14000000L){ 
    digitalWrite(TX_LPF_A, 1);
    digitalWrite(TX_LPF_B, 0);  // 20M
    digitalWrite(TX_LPF_C, 0);
  }
  else if (freq >= 7000000L){  
    digitalWrite(TX_LPF_A, 0);
    digitalWrite(TX_LPF_B, 1);  // 40m
    digitalWrite(TX_LPF_C, 0);    
  }
  else {
    digitalWrite(TX_LPF_A, 0);
    digitalWrite(TX_LPF_B, 0);  // 80m
    digitalWrite(TX_LPF_C, 1);    
  }
}

Re: 80 Meters Harmonics Fix Proposal

ajparent1/KB1GMX
 

 peeling the onion, slowly. I thought the problematic band was 15 Meters for SSB? 

No its basically 20mhz and up as in all 15/12/10 for SSB.

Allison

Re: UBITX TX level diagramme

Lawrence Galea
 

Thanks Allison
Regards

On Mon, Sep 3, 2018 at 7:41 PM ajparent1/KB1GMX <kb1gmx@...> wrote:
Lawrence G.,

I tried a few things along that path.  made matters worse and that was also tried with
external level 7 mixer and level 17 external mixers.   It was then a rock fell on me
and I realized the output of the 45mhz amp with the TX producing power was filthy.
Add a filter module it got better but by then I had so many outboard modules
its was almost a breadboard radio.  Everything is talking to everything by sneak
paths.  

Allison
We need only one magik part.

uBITX40 filtered

Lev
 

Dear all,


I'm thinking about the adding low pass filters to the output of all
oscillators. The goal is to filter the square wave of the clock generator. I came up with this mixer solution. Using a JFET mixer is just my other experiment.

Question is if this circuit would clean the output signal?

Also, I'm thinking about modifying the IF amplifiers to be tuned.

Any ideas?



73s de HA5OGL

--
Levente Kovacs
Senior Electronic Engineer

W: http://levente.logonex.eu

Re: UBITX TX level diagramme

Glenn
 

Allison, is this the drawing (pdf) referred to?   Attached .BMP copy could be amended with MS Paint.



glenn