Si5351A Synth kit board Rev 5 now available, fits TCXO module


Hans Summers
 

Hi all

I'm pleased to announce a minor PCB revision to the Si5351A Synth kit http://qrp-labs.com/synth so that the 25MHz TCXO module fits precisely (as originally designed for the QCX+, and later the QCX-mini). The Rev 5 PCB has been re-arranged so that the 25MHz TCXO sits on the board in a similar way to how it fits in the QCX+ and QCX-mini kits (see photos attached).

25MHz TCXO: http://shop.qrp-labs.com/qcxptcxo costs $8.25

This 25MHz TCXO runs at 3.3V, consumes 1.25mA, has 1.1V peak-peak amplitude, and by measurement typically is within 0.25ppm of 25MHz and has a drift of within +/- 0.25ppm across the temperature range 0 to 65C (see attached). Very importantly for weak signal (e.g. WSPR) purposes, it is an analog TCXO making its corrections in a proportional analog manner, not digital "step" corrections which can corrupt weak signal modes. Over 3,000 of these TCXO modules have been shipped with QCX+ and QCX-mini CW transceivers, and they have shown excellent reliability and performance. 

The Si5351A Assembly manual has been updated to include instructions to install the 25MHz TCXO option. The new Rev 5 Synth can be ordered here http://shop.qrp-labs.com/synth

From now onwards, all kits that include the Si5351A Synth kit (e.g. U3S, ProgRock, VFO) will be shipped with the Rev 5 PCB that is capable of being fitted with the 25MHz TCXO option, if you choose to add it to your order. 

73 Hans G0UPL
http://qrp-labs.com



GIUSEPPE
 

Good morning Hans, great job, QRP LAB is always constantly updated and improved for all of us ham.  73 and have a nice day.  iu8eun Joseph


Il sab 4 set 2021, 07:37 Hans Summers <hans.summers@...> ha scritto:
Hi all

I'm pleased to announce a minor PCB revision to the Si5351A Synth kit http://qrp-labs.com/synth so that the 25MHz TCXO module fits precisely (as originally designed for the QCX+, and later the QCX-mini). The Rev 5 PCB has been re-arranged so that the 25MHz TCXO sits on the board in a similar way to how it fits in the QCX+ and QCX-mini kits (see photos attached).

25MHz TCXO: http://shop.qrp-labs.com/qcxptcxo costs $8.25

This 25MHz TCXO runs at 3.3V, consumes 1.25mA, has 1.1V peak-peak amplitude, and by measurement typically is within 0.25ppm of 25MHz and has a drift of within +/- 0.25ppm across the temperature range 0 to 65C (see attached). Very importantly for weak signal (e.g. WSPR) purposes, it is an analog TCXO making its corrections in a proportional analog manner, not digital "step" corrections which can corrupt weak signal modes. Over 3,000 of these TCXO modules have been shipped with QCX+ and QCX-mini CW transceivers, and they have shown excellent reliability and performance. 

The Si5351A Assembly manual has been updated to include instructions to install the 25MHz TCXO option. The new Rev 5 Synth can be ordered here http://shop.qrp-labs.com/synth

From now onwards, all kits that include the Si5351A Synth kit (e.g. U3S, ProgRock, VFO) will be shipped with the Rev 5 PCB that is capable of being fitted with the 25MHz TCXO option, if you choose to add it to your order. 

73 Hans G0UPL
http://qrp-labs.com



Larry Christensen
 

After receiving your email Friday answering my questions and informing my of the rev 5 / txco information I ordered the 6 band U3S with the txco.  

Now just wait for it to be shipped and delivered. My QCX+ was shipped from MO. Hopefully this will be also. 

Thank for the great support. 

Larry NA0F


Rob Giuliano
 

I am currently running a U3S with the non TCXO crystal. 
I've been debating changing my U3S to use a TCXO. 
   I find my current U3S setup with GPS correction does keeps the frequency steady, but 'off'. 
   It also seems to vary after when not in constant use:  Start at 28,126,135 will TX at 28,126,080 and stay there.  Shut down and restart and it may TX at 28,126,117 and stay there.
   Setting the crystal frequency and not using the GPS seems to be as steady and stay at the same frequency between power cycles.

So if I want to switch to TCXO, it appears the better route would be to purchase the new Rev5 board and QRP-Labs TCXO module? 
I say (ask) that for 3 reasons:
1.  Rev5 board is laid out specifically for the QRP-Labs TCXO and makes a nice board to install in the U3S
2.  I would need to remove parts from the existing, functional (Synth) board (Rev 4) and the jumper the connections to it.
     These would not be as nice and clean as using the Rev5 board.

Puedo 3: Purchasing a full Rev5 Synth and 25MHz TCXO would free up the non TCXO Si5351 board for other projects.
               Since the boards are interchangeable, I could swap the TCXO and nonTCXO board as desired.

--
Rob KB8RCO


Dave VE3GSO
 

Seems like you answered your own question.

I have taken part in the bi-annual ARRL FMT and have been sub hertz in my submissions.  Of course I have taken great interest in the accuracy of the TCXO that Hans sells and have been surprised and pleased that so far the TCXOs that I have tested, four so far, have been right on 25 MHz plus or minus my measurement uncertainty of a few milliHertz.  Quite remarkable!

And even if they are found to be a few Hertz off, that can be accounted for in the calibration settings.

Dave

On Sep 9, 2021, at 10:05, Rob Giuliano via groups.io <kb8rco@...> wrote:

I am currently running a U3S with the non TCXO crystal. 
I've been debating changing my U3S to use a TCXO. 
   I find my current U3S setup with GPS correction does keeps the frequency steady, but 'off'. 
   It also seems to vary after when not in constant use:  Start at 28,126,135 will TX at 28,126,080 and stay there.  Shut down and restart and it may TX at 28,126,117 and stay there.
   Setting the crystal frequency and not using the GPS seems to be as steady and stay at the same frequency between power cycles.

So if I want to switch to TCXO, it appears the better route would be to purchase the new Rev5 board and QRP-Labs TCXO module? 
I say (ask) that for 3 reasons:
1.  Rev5 board is laid out specifically for the QRP-Labs TCXO and makes a nice board to install in the U3S
2.  I would need to remove parts from the existing, functional (Synth) board (Rev 4) and the jumper the connections to it.
     These would not be as nice and clean as using the Rev5 board.

Puedo 3: Purchasing a full Rev5 Synth and 25MHz TCXO would free up the non TCXO Si5351 board for other projects.
               Since the boards are interchangeable, I could swap the TCXO and nonTCXO board as desired.

--
Rob KB8RCO


M0PWX
 

 

 

Hi Rob

 

Below is my mod’ed Rev 4 board with TCXO

 

Remove C3 and XTAL

 

Solder new C3  (.1uf cap) to TCXO board

 

Put legs through the holes C3 came out of on the SiSSynth board (these provides  power tot the TCXO)

 

Then add a wire from the output on the TCXO to the XTAL hole nearest the end of the SISSynth board (see side photo below)

 

On the rev 5 board the SIS chip and C3 are swapped round on the board so the cut out sits neatly over the chip and all the holes line up

 

Unless you are getting other bits from Hans  on the same order, then I would suggest if you are paying more on P&P then the TCXO board you might as well put a SISSynth in the basket as well

 

I did look at getting a SMD TCXO to fit the board between the BS170 and regulator but they are more expensive and not as stable as Hans little tried and tested TCXO boards

 

73

Peter

M0PWX

 

 

 

From: Rob Giuliano via groups.io
Sent: 09 September 2021 14:05
To: QRPLabs@groups.io
Subject: Re: [QRPLabs] Si5351A Synth kit board Rev 5 now available, fits TCXO module

 

I am currently running a U3S with the non TCXO crystal. 
I've been debating changing my U3S to use a TCXO. 
   I find my current U3S setup with GPS correction does keeps the frequency steady, but 'off'. 
   It also seems to vary after when not in constant use:  Start at 28,126,135 will TX at 28,126,080 and stay there.  Shut down and restart and it may TX at 28,126,117 and stay there.
   Setting the crystal frequency and not using the GPS seems to be as steady and stay at the same frequency between power cycles.

So if I want to switch to TCXO, it appears the better route would be to purchase the new Rev5 board and QRP-Labs TCXO module? 
I say (ask) that for 3 reasons:
1.  Rev5 board is laid out specifically for the QRP-Labs TCXO and makes a nice board to install in the U3S
2.  I would need to remove parts from the existing, functional (Synth) board (Rev 4) and the jumper the connections to it.
     These would not be as nice and clean as using the Rev5 board.

Puedo 3: Purchasing a full Rev5 Synth and 25MHz TCXO would free up the non TCXO Si5351 board for other projects.
               Since the boards are interchangeable, I could swap the TCXO and nonTCXO board as desired.

--
Rob KB8RCO

 


Mont Pierce KM6WT
 

On Thu, Sep 9, 2021 at 08:13 AM, M0PWX wrote:
I did look at getting a SMD TCXO to fit the board between the BS170 and regulator but they are more expensive and not as stable as Hans little tried and tested TCXO boards
Did you find any that calibrate freq in analog fashion like Hans does?

I bought a batch of 20 of the FOX924 from DigiKey before finding out these calibrate freq digitally...  which is Not good for WSPR.  They work great otherwise, just not appropriate for WSPR mode.


Most specs on all the TCXOs I've looked at, don't elaborate on "how" calibration is performed (i.e. digital vs analog adj).


tnx
km6wt


M0PWX
 

No I just looked at price, size and PPM for stability,

 

Didn’t even get as far as how they calibrated,

 

Cost and stability did it for me, cheaper even with P&P to get the TCXO board from Hans

 

73

 

Peter

M0PWX

 

 

 

On Thu, Sep 9, 2021 at 08:13 AM, M0PWX wrote:

I did look at getting a SMD TCXO to fit the board between the BS170 and regulator but they are more expensive and not as stable as Hans little tried and tested TCXO boards

Did you find any that calibrate freq in analog fashion like Hans does?

I bought a batch of 20 of the FOX924 from DigiKey before finding out these calibrate freq digitally...  which is Not good for WSPR.  They work great otherwise, just not appropriate for WSPR mode.


Most specs on all the TCXOs I've looked at, don't elaborate on "how" calibration is performed (i.e. digital vs analog adj).


tnx
km6wt

_._,_._,_

 


Rob Giuliano
 

On Thu, Sep 9, 2021 at 10:52 AM, Dave VE3GSO wrote:
Seems like you answered your own question.
It was meant to solicit thoughts and opinion on the new board revision. 
And probably more so, on others using the nonTCXO and GPS for calibration.
I find it VERY strange behavior to have the frequency change between power cycles.
   I would have thought it would always adjust to the same frequency. 
   My observed overall spread is about as wide as the WSPR band on 15m and up.

I do have another project to use a Si5351 (doesn't need the TXCO) and was thinking of getting something of eBay or Amazon. 
Then I saw Hans' post and thought about changing route.  As someone stated, kind of thin king of other things to combine to a 'reasonable' order.
 
--
Rob KB8RCO


M0PWX
 

From what I saw as I was playing with my U3s waiting for GPS module

 

There are a number of temp related factors affecting frequency stability

 

The non GPS o TXCO SIS Synth will always reset to the value store for the 27mhz ref freq

 

The GPS seems to update the stored value for the reference frequency, but there is the issue, because the NON TCXO XTAL is temp sensitive once it has run long enough to calibrate accurately it has moves the store ref freq value some way from its original “cold” value

If you then turn the unit off and let it cool down, when you turn  it back on the stored ref freq value will be off by a fair margin off error until the XTAL warms up again

 

The TCXO is stable to about 0.25 PPM and a similar yearly drift, the XTALs seem to be nearer 5-10PPM stability and very temp sensitive

 

For some applications like WSPR they do drift but don’t exceed the limits of the wspr protocols (8hz over 2 mnutes I believe)

 

But very tight visual things like QRSS it is exceedingly visible as you are working with signals only 5hz difference at 28mhz etc

 

So for QRSS a TCXO and GPS are almost a necessity for the required stability

 

As to your point about the differences between the rev4 and rev 5 boards

 

Is only a minor layout  change by the look of it so C3, the output on the TCXO module match up with the holes on the SIS Synth board and the SIS Synth chip matches up with the cutout

 

Other than that (unless Hans corrects me) they are the same beast and as you can see from my previous post the TCXO sits neatly on the rev4 board

 

73

 

Peter

M0PWX

Sent from Mail for Windows

 

On Thu, Sep 9, 2021 at 10:52 AM, Dave VE3GSO wrote:

Seems like you answered your own question.

It was meant to solicit thoughts and opinion on the new board revision. 
And probably more so, on others using the nonTCXO and GPS for calibration.
I find it VERY strange behavior to have the frequency change between power cycles.
   I would have thought it would always adjust to the same frequency. 
   My observed overall spread is about as wide as the WSPR band on 15m and up.

I do have another project to use a Si5351 (doesn't need the TXCO) and was thinking of getting something of eBay or Amazon. 
Then I saw Hans' post and thought about changing route.  As someone stated, kind of thin king of other things to combine to a 'reasonable' order.
 
--
Rob KB8RCO

 


Rob Giuliano
 

Except that ISN'T what happens. 
The 'long term settled' frequency changes, not the immediate, "Turn ON" frequency.
For instance:
 Run for 8 hours with GPS CALIBRATION:  say it settles on 28,126,080 and stay within a Hz or 2 of that.
 Turn off for several hours (overnight since I was using 10m). 
 After another 8 hours or so the next day with GPS CALIBRATION, it may settle on 28,126,117 and stay within a Hz or 2 of that.

I would EXPECT it to settle to the same frequency over time. 
That is the point I am trying to see if other have experienced.

I am pretty much only using the U3S fro WSPR.
NOTE: during the WSPR TX, the worst drift reported at any time is 1 on nearby receiver.

--
Rob KB8RCO


M0PWX
 



Ok, that does look odd now we are getting into more detail

 

At the end of the calibration cycle what do you see ?

 

It will normally display the correction or an “=” if no adjustment required?




73


Peter

 

Sent from Mail for Windows

 

From: Rob Giuliano via groups.io

 Run for 8 hours with GPS CALIBRATION:  say it settles on 28,126,080 and stay within a Hz or 2 of that.
 Turn off for several hours (overnight since I was using 10m). 
 After another 8 hours or so the next day with GPS CALIBRATION, it may settle on 28,126,117 and stay within a Hz or 2 of that.

I would EXPECT it to settle to the same frequency over time. 
That is the point I am trying to see if other have experienced.



--
Rob KB8RCO

 


Rob Giuliano
 

Normally, I see an adjustment, but VERY small (only a Hz or 2).
I don't
Just started it again (from cold start).  Set TX to 28,126,135.  CAL 02 100
   27,004,148 RX: 28,126,160  Drift -1                  Starting frequency (after power up CAL adjustment)
   27,004,143 RX: 28,126,159  Drift 0                   10 minutes later (Start 02 Frame 10)
I'll let it run now for awhile.

--
Rob KB8RCO


Rob Giuliano
 

UPDATE:
So I am seeing =.
Started:   19:42 on Sept 9
1942    27,004,148  RX: 28,126,159 Drift 0
1952    27,004,143  RX: 28,126,160 Drift -1
2002    27,004,144  RX: 28,126,160 Drift -1
2012    =                  RX: 28,126,161 Drift 0
2022    =                  RX: 28,126,160 Drift 0
2032    =                  RX: 28,126,160 Drift 0
Clock Frequency is post adjustment from previous TX.

Since I saw 3 (actually 4) no adjustment periods, I will shut it down for awhile.
My expectation would be that when I restart again (from 27,004,144), it should adjust such that the output frequency would be 28,126,160 again.

I do not think the adjusted Ref. Freq. is saved after the CAL. 
Each time I have checked the Ref. Freq. shows 27,004,148.  I set it to 27,004,144, then shut it down.

--
Rob KB8RCO


Roger Hill
 

HI Rob:

I have not been following this thread in detail, but I am curious about your RX numbers. Where are you getting them from? Remember that any drift shown is the net effect of your transmission drift and the receiving station drift, and does not directly relate to the drift of the U3S.


73

Roger

8P6RX

---
***************************
Roger Hill
***************************


On 2021-09-09 16:38, Rob Giuliano via groups.io wrote:

UPDATE:
So I am seeing =.
Started:   19:42 on Sept 9
1942    27,004,148  RX: 28,126,159 Drift 0
1952    27,004,143  RX: 28,126,160 Drift -1
2002    27,004,144  RX: 28,126,160 Drift -1
2012    =                  RX: 28,126,161 Drift 0
2022    =                  RX: 28,126,160 Drift 0
2032    =                  RX: 28,126,160 Drift 0
Clock Frequency is post adjustment from previous TX.

Since I saw 3 (actually 4) no adjustment periods, I will shut it down for awhile.
My expectation would be that when I restart again (from 27,004,144), it should adjust such that the output frequency would be 28,126,160 again.

I do not think the adjusted Ref. Freq. is saved after the CAL. 
Each time I have checked the Ref. Freq. shows 27,004,148.  I set it to 27,004,144, then shut it down.

--
Rob KB8RCO


Alan G4ZFQ
 

I do not think the adjusted Ref. Freq. is saved after the CAL.
Rob'

No.
I think I'm correct, there is a minimum "save" value in the configuration menu.
It is best not set too low as prolonged use might exceed the number of permitted memory writes. Small increments are held in temporary memory.

73 Alan G4ZFQ


M0PWX
 

Perhaps Hans could clarify if the calibration process write to the ref freq

 

And if so when

 

73

 

Peter

M0PWX

 

> I do not think the adjusted Ref. Freq. is saved after the CAL.

Rob'

No.
I think I'm correct, there is a minimum "save" value in the
configuration menu.
It is best not set too low as prolonged use might exceed the number of
permitted memory writes. Small increments are held in temporary memory.

73 Alan G4ZFQ


 


Rob Giuliano
 

On Thu, Sep 9, 2021 at 04:46 PM, Roger Hill wrote:
Where are you getting them from? Remember that any drift shown is the net effect of your transmission drift and the receiving station drift, and does not directly relate to the drift of the U3S.
The point was not the actual values.  The point of the reported numbers (from a nearby receiver) was to show that the drift was MINIMAL.
AND ...  That the Decode Frequency (on the EXACT same receiver) shifts between power cycles - after the Ref. Freq. stablizes on a frequency.

So, the point there is that a drift of -1 (as you just stated) is very small - STABLE.  Especially when you take into account the components involved.
In other words - that is NOT an issue!
 
--
Rob KB8RCO


Rob Giuliano
 

I did leave out that I haven't restarted the U3S yet.  Tomorrow morning I'll start again.
--
Rob KB8RCO


W7QJQ
 

Hi Dave, perhaps I have mis-read what you wrote: 
   "... so far the TCXOs that I have tested, four so far, have been right on 25 MHz plus or minus my measurement uncertainty of a few milliHertz."  
To expect an inexpensive xtal clock to exhibit this level of  accuracy is simply totally unbelievable.
73,  Sid  W7QJQ
Radio Welling