Topics

a simple spur fix

Jerry Gaffke
 

Farhan reports that his coil and cap have a null 10dB better than needed to meet regs, no need for higher Q.
If the 90mhz trap can be done with a surface mount coil and cap on the back of the board, that's a very easy fix.


On Sat, Nov 10, 2018 at 09:16 AM, iz oos wrote:

I have no doubt Farhan solution might work but I guess that the Q of the 90Mhz and the null could be higher using a quarter wave open coax stub cut for the 90Mhz (likely a cap between 20 and 30pf in parallel would be needed).

 

STEVE
 

I remember a relatively simple fix using a 45 MHz crystal across 2 points with a ground on the center wire. Has anyone done this?  What were the connection points?  Thanks. Steve. KE7GO 

Vince Vielhaber
 

Take an existing raduino and disable the si5351. Bring the i2c lines off to an external si5351 board, like the adafruit. Put the buffers on that board. Get the display and nano away from the si5351. Audio beat problem is gone. Design of a new raduino or redesign of the old is no longer necessary.

Now you can find out of the buffers will solve any of the problems. If it does, moving the si5351 to the main board (with buffers) will solve the need of another board. Add a header to the raduino board for i2c for now for an external board.

Vince.

On 11/10/2018 10:37 AM, Jerry Gaffke via Groups.Io wrote:
This might be an opportunity for somebody to create yet another new Raduino
Add buffer amps to all three clocks from the si5351 to help solve the
spur problem.
Choose a processor clock that won't beat with the BFO, creating audio
tones cases
where the Nano's 16mhz ceramic resonator happens to fall on an unlucky
frequency .
Communicate with a host processor via well filtered 3v UART lines
that go off to a well shielded USB-to-UART dongle.
Processor is a little beefier, something like an STM32F*, not pressed
for memory space,
more IO pins, and with the horsepower for a few simple digital modes
such as PSK31.

This could easily be used to upgrade old v3 and v4 uBitx's.

Jerry, KE7ER

On Sat, Nov 10, 2018 at 06:41 AM, Jerry Gaffke wrote:

Unless we add an amp to each of the three clocks between the si5351
and the mixers,
we are trading off si5351 crosstalk, local oscillator injection
level, and reflections due to
driving the mixer from a source impedance higher than 50 ohms.
All three of these factors will contribute to spurs.

Jerry Gaffke
 

A quick recap, as I recall the spur fix thing thing thus far:

Allison noted that the LO's had too low of an injection level, the mixers were overloaded with signal,
the 45mhz IF amp was not amping quite enough, there was coupling from the power amp into the IF's
due to poor board layout, and that most multiband HF rigs would have a filter for each band of use at L1,2,3,4.
All of this contributing to spurs, and suggesting a significant redesign was needed.
 
Raj then found that the spurs were reduced by inserting a 45mhz crystal filter in series with C22.
Alternate placements and some impedance matching schemes were proposed, in part to bring the transmit power
back up to what it was.  Resonant freq of any new 45mhz filter would have to match the old 45mhz filter at Y1.
This fix seemed to bring spurs down to where they about met regs in most cases, but still borderline.

Then Raj found he got better results by replacing L5 and L7 with a particular brand of 1210 sized shielded surface mount inductor.
Then found that the same treatment to L1,2,3,4 made it even better.
So forget the 45mhz crystal filter.

Now Farhan says forget all the above, he's doing better still with a couple series LC traps
at TP13 and C202 for 90mhz and 12mhz respectively.
And the discussion continues from there.

By all means, jump in and try any of the above if so inclined
and you have the instrumentation to measure spurs. 
Otherwise, sit on your hands a bit longer.
Allison is right, a complete redesign of about half the uBitx is not a bad idea.
But Farhan and Raj appear close to finding a spur fix for v3 and v4 uBitx boards that is very easy and cheap.


Spurs are at frequencies more or less unrelated to the operating frequency.
On a stock rig, in some cases they can be stronger than allowed by regulations, though at QRP levels
are reasonably low compared to the typical 100W hf transceiver.

Harmonics are a separate issue, solved by either adding external low pass filters
or hacking the relays and traces around the existing transmit LPF's.

Jerry, KE7ER
Jerry, KE7ER



On Sat, Nov 10, 2018 at 10:09 AM, <behlmer@...> wrote:

I remember a relatively simple fix using a 45 MHz crystal across 2 points with a ground on the center wire. Has anyone done this?  What were the connection points?  Thanks. Steve. KE7GO 

Glenn
 

Jerry, for what it's worth i did a small pcb for Si5351 buffering in an old post:- buffer board

It was designed to fit on the uBITX board adjacent to the "Raduino1" socket. Cut the 3 tracks from Si5351 and connect to the pads on the pcb. Then the outputs of the board are linked to adjacent points, now buffered.


Glenn

jim
 

A bit more elegant than my "interim" solution

Jim


On Saturday, November 10, 2018, 1:33:54 PM PST, Glenn <glennp@...> wrote:


Jerry, for what it's worth i did a small pcb for Si5351 buffering in an old post:- buffer board

It was designed to fit on the uBITX board adjacent to the "Raduino1" socket. Cut the 3 tracks from Si5351 and connect to the pads on the pcb. Then the outputs of the board are linked to adjacent points, now buffered.


Glenn

Glenn
 

Similar though Jim,. I used 74LVC1G126 devices. Plenty of level and I see you have put a pad in there. I am yet to actually fit my board as my uBITX is in bits right now..... pcb out of case etc.
glenn

Jerry Gaffke
 

Both the 74LVC1G126 and the 74LVC1G14 look good to me in this application.
Bypass each power pin well with a couple caps, and maybe even add an inductor at each,
you don't want crosstalk between those three chips.  Or those chips and the si5351.

Neither part expects a 50 ohm load.  PCB traces over a ground plane are usually well over 100 ohms,,
so a higher load impedance would actually cut down on the reflections.
In Jim's PI attenuator, I'd start with removing R101 entirely, set R102 to 150 ohms, and R103 to 75 ohms.
More or less as per this post:   https://groups.io/g/BITX20/message/61587
except that I assume there is not a 50 ohm internal source resistor in those 74LVC1G* parts.
(There will be some internal source resistance, I don't know what.  And the mixer ports are
not likely to be exactly 50 ohms either. )

Jim's post has four images, the last one shows something like a spectrum analyzer.
Can Jim tell us what we are looking at there?

Jerry


On Sat, Nov 10, 2018 at 04:48 PM, Glenn wrote:
Similar though Jim,. I used 74LVC1G126 devices. Plenty of level and I see you have put a pad in there. I am yet to actually fit my board as my uBITX is in bits right now..... pcb out of case etc.
glenn

Ashhar Farhan
 

using buffers seem to be an excessive solution. 
here is hownit goes :
there are three clocks from si5351 for the ubitx. of them, two are in the vhf region, only one is in the hf region, that is constantly at 12 mhz.
we have a spur that stays at 12 mhz all the time. when we tx below 12 mhz, it is swallowed by the LPFs.
at 14 and above, it shows up at around -40 dbc. we need to notch it down by a couple of dbs. 
the easiest way is to use a trap at 12 mhz. i guess even a 12 mhz crystal would work.
buffers will be need surely for any other project using the raduino. however, for ubitx, that single trap will kill the 12 mhz spur.

On Sun, 11 Nov 2018, 09:12 Jerry Gaffke via Groups.Io <jgaffke=yahoo.com@groups.io wrote:
Both the 74LVC1G126 and the 74LVC1G14 look good to me in this application.
Bypass each power pin well with a couple caps, and maybe even add an inductor at each,
you don't want crosstalk between those three chips.  Or those chips and the si5351.

Neither part expects a 50 ohm load.  PCB traces over a ground plane are usually well over 100 ohms,,
so a higher load impedance would actually cut down on the reflections.
In Jim's PI attenuator, I'd start with removing R101 entirely, set R102 to 150 ohms, and R103 to 75 ohms.
More or less as per this post:   https://groups.io/g/BITX20/message/61587
except that I assume there is not a 50 ohm internal source resistor in those 74LVC1G* parts.
(There will be some internal source resistance, I don't know what.  And the mixer ports are
not likely to be exactly 50 ohms either. )

Jim's post has four images, the last one shows something like a spectrum analyzer.
Can Jim tell us what we are looking at there?

Jerry


On Sat, Nov 10, 2018 at 04:48 PM, Glenn wrote:
Similar though Jim,. I used 74LVC1G126 devices. Plenty of level and I see you have put a pad in there. I am yet to actually fit my board as my uBITX is in bits right now..... pcb out of case etc.
glenn

Jerry Gaffke
 

On the uBitx, not clear it is only crosstalk within the si5351 that is allowing some 12mhz to sneak through.
Traps or filters of some sort may well be the best way forward.
Using a crystal as a trap would be cool if it worked, affects the receiver only for a few hundred hz
at 12.000 mhz, a region that's useless now anyway because all you can hear is the BFO.



On Sat, Nov 10, 2018 at 07:51 PM, Ashhar Farhan wrote:
using buffers seem to be an excessive solution. 
here is hownit goes :
there are three clocks from si5351 for the ubitx. of them, two are in the vhf region, only one is in the hf region, that is constantly at 12 mhz.
we have a spur that stays at 12 mhz all the time. when we tx below 12 mhz, it is swallowed by the LPFs.
at 14 and above, it shows up at around -40 dbc. we need to notch it down by a couple of dbs. 
the easiest way is to use a trap at 12 mhz. i guess even a 12 mhz crystal would work.
buffers will be need surely for any other project using the raduino. however, for ubitx, that single trap will kill the 12 mhz spur.

jim
 

Picture of SA connected to junction of R101/R102 ..Shows clk02 and harmonics (appron -5 dbm fundamental) ubitx tuned to 30 mhz
Jim


On Saturday, November 10, 2018, 7:42:04 PM PST, Jerry Gaffke via Groups.Io <jgaffke@...> wrote:


Both the 74LVC1G126 and the 74LVC1G14 look good to me in this application.
Bypass each power pin well with a couple caps, and maybe even add an inductor at each,
you don't want crosstalk between those three chips.  Or those chips and the si5351.

Neither part expects a 50 ohm load.  PCB traces over a ground plane are usually well over 100 ohms,,
so a higher load impedance would actually cut down on the reflections.
In Jim's PI attenuator, I'd start with removing R101 entirely, set R102 to 150 ohms, and R103 to 75 ohms.
More or less as per this post:   https://groups.io/g/BITX20/message/61587
except that I assume there is not a 50 ohm internal source resistor in those 74LVC1G* parts.
(There will be some internal source resistance, I don't know what.  And the mixer ports are
not likely to be exactly 50 ohms either. )

Jim's post has four images, the last one shows something like a spectrum analyzer.
Can Jim tell us what we are looking at there?

Jerry


On Sat, Nov 10, 2018 at 04:48 PM, Glenn wrote:
Similar though Jim,. I used 74LVC1G126 devices. Plenty of level and I see you have put a pad in there. I am yet to actually fit my board as my uBITX is in bits right now..... pcb out of case etc.
glenn

 

Jerry,

Just replacing the KT1/2/3 not only fixes the harmonics and fixes the spurs too..

Raj

At 11/11/2018, you wrote:

A quick recap, as I recall the spur fix thing thing thus far:

Allison noted that the LO's had too low of an injection level, the mixers were overloaded with signal,
the 45mhz IF amp was not amping quite enough, there was coupling from the power amp into the IF's
due to poor board layout, and that most multiband HF rigs would have a filter for each band of use at L1,2,3,4.
All of this contributing to spurs, and suggesting a significant redesign was needed.

Raj then found that the spurs were reduced by inserting a 45mhz crystal filter in series with C22.
Alternate placements and some impedance matching schemes were proposed, in part to bring the transmit power
back up to what it was. Resonant freq of any new 45mhz filter would have to match the old 45mhz filter at Y1.
This fix seemed to bring spurs down to where they about met regs in most cases, but still borderline.

Then Raj found he got better results by replacing L5 and L7 with a particular brand of 1210 sized shielded surface mount inductor.
Then found that the same treatment to L1,2,3,4 made it even better.
So forget the 45mhz crystal filter.

Now Farhan says forget all the above, he's doing better still with a couple series LC traps
at TP13 and C202 for 90mhz and 12mhz respectively.
And the discussion continues from there.

By all means, jump in and try any of the above if so inclined
and you have the instrumentation to measure spurs.
Otherwise, sit on your hands a bit longer.
Allison is right, a complete redesign of about half the uBitx is not a bad idea.
But Farhan and Raj appear close to finding a spur fix for v3 and v4 uBitx boards that is very easy and cheap.


Spurs are at frequencies more or less unrelated to the operating frequency.
On a stock rig, in some cases they can be stronger than allowed by regulations, though at QRP levels
are reasonably low compared to the typical 100W hf transceiver.

Harmonics are a separate issue, solved by either adding external low pass filters
or hacking the relays and traces around the existing transmit LPF's.

Jerry, KE7ER

Jerry Gaffke
 

Raj,

Replacing the relays may reduce some of the spurs that are higher than the operating frequency.

But the problematic spurs are lower in frequency than the operating frequency and easily pass through the LPF's.
So fixing how harmonics can leak past the transmit LPF's (by replacing those 3 relays)
seems unlikely to fix our trouble with low frequency spurs.



Jerry


On Sun, Nov 11, 2018 at 07:34 PM, Raj vu2zap wrote:
Just replacing the KT1/2/3 not only fixes the harmonics and fixes the spurs too..

Nick VK4PP
 

Is anyone working on fixing the BIDI amps? or just the easy band-aids?
I would love to see the amps fixed with better transistors and associated changes, rather that quick fixes,
Unfortunately I dont have the understanding to do it myself yet...

Thanks and 73.

Arv Evans
 

Nick

What is wrong with the bidirectional amplifiers?  They have been succesfully used in all BITX transceivers for over 10 years.

Arv. K7hkl
_-_


Sent from my Verizon, Samsung Galaxy smartphone

-------- Original message --------
From: Nick VK4PP <nickpullen@...>
Date: 11/11/18 11:13 PM (GMT-07:00)
To: BITX20@groups.io
Subject: Re: [BITX20] a simple spur fix

Is anyone working on fixing the BIDI amps? or just the easy band-aids?
I would love to see the amps fixed with better transistors and associated changes, rather that quick fixes,
Unfortunately I dont have the understanding to do it myself yet...

Thanks and 73.

Nick VK4PP
 

Just using a device that performs better at 45mhz, BFR106? and associated bias changes....

AndyH
 

On Mon, Nov 12, 2018 at 07:15 PM, Nick VK4PP wrote:
Just using a device that performs better at 45mhz, BFR106? and associated bias changes....

Nick VK4PP
 

Thanks Andy.
I have read that, but its all over the place a bit. Probably easy enough if you have a thorough working knowledge of electronics, But for a NOOB like me, I get some of what is said., but really a step by step
replace this with that. and maybe a commoners explanation of what is happening. Could be a good lesson on how these amps work.

My apologies if I seem to be going round in circles. but hey.

Dave Space
 

Hi
I just bought a ubitx board.  Hasn't arrived yet.  Is this mod still needed on the current boards shipping?

Is there a schematic or closer zoom in for where the parts go across for both fixes? 
Even zoomed in on the photo having a little trouble making out where the connections are.



AndyH
 

No worries Nick - nothing wrong with circles.  hi hi  Revising the thread I see that you were part of it - sorry for the redundant link.

Here are Allison's suggestions for the two RF amps:
45mhz IF replace (Q20, Q21, Q22) with BFR106.  Change R23 to 18ohms.  Change R27 to 39 ohms.
12mhz IF amp change R43 to 18 ohma and put a .1uf in series with 680 ohms across R41.
https://groups.io/g/BITX20/message/51448

Cheers,
Andy

On Mon, Nov 12, 2018 at 08:36 PM, Nick VK4PP wrote:
Thanks Andy.
I have read that, but its all over the place a bit. Probably easy enough if you have a thorough working knowledge of electronics, But for a NOOB like me, I get some of what is said., but really a step by step
replace this with that. and maybe a commoners explanation of what is happening. Could be a good lesson on how these amps work.

My apologies if I seem to be going round in circles. but hey.