GPS RX Type TU35-D140-041


Ed <ed.murphy31@...>
 

Hello All,

Has anyone got a spare TU35-D140-041 GPS Rx for sale ?

1PPS and 10Khz o/p.


Its going to a good cause, a 10 Ghz JT4 Beacon.


Cheers Ed GM3SBC


Andy G4JNT
 

Does that one give a reliable NMEA output ?
 
Coz I've been thinking about the JT4 (and JT65 and WSPR) generation.  There is no need to actually read the time - the PPS alone is probably good enough.   You just synchronise it to the right x2:00 point and it'll run from there...
 
A nice 1 PPS based GPSDO would be a good idea too - the ZAZ is already excellent design and I have two of them, but does rely on having a good OCXO to start with.   An intermediate spec one  based on a good TCXO and giving a similar short term stability to the low cost Jupiter 10kHz design seems to be needed.  They're made commercially so it certainly possible.
 
We're all placing too much reliance on that Jupiter concept - and its a single point failure if the source dries up.
 
Andy


 

On 25 November 2011 14:38, Ed <ed.murphy31@...> wrote:
 

Hello All,

Has anyone got a spare TU35-D140-041 GPS Rx for sale ?

1PPS and 10Khz o/p.

Its going to a good cause, a 10 Ghz JT4 Beacon.

Cheers Ed GM3SBC



Ed <ed.murphy31@...>
 

--- In ukmicrowaves@yahoogroups.com, Andy Talbot <andy.g4jnt@...> wrote:

Does that one give a reliable NMEA output ?
Trouble is Andy that I have a system that is already built. I need a part that is a direct replacement otherwise I shall have to make an entirely new system . new box, psu etc ,etc. My present GPS part has just decided to fail. There must be lots of these lying about just rusting away. I need just ONE.
Come on guys, look in your junk box.

Thanks Ed GM3SBC


Coz I've been thinking about the JT4 (and JT65 and WSPR) generation. There
is no need to actually read the time - the PPS alone is probably good
enough. You just synchronise it to the right x2:00 point and it'll run
from there...

A nice 1 PPS based GPSDO would be a good idea too - the ZAZ is already
excellent design and I have two of them, but does rely on having a good
OCXO to start with. An intermediate spec one based on a good TCXO and
giving a similar short term stability to the low cost Jupiter 10kHz design
seems to be needed. They're made commercially so it certainly possible.

We're all placing too much reliance on that Jupiter concept - and its a
single point failure if the source dries up.

Andy
www.g4jnt.com



On 25 November 2011 14:38, Ed <ed.murphy31@...> wrote:

**


Hello All,

Has anyone got a spare TU35-D140-041 GPS Rx for sale ?

1PPS and 10Khz o/p.

Its going to a good cause, a 10 Ghz JT4 Beacon.

Cheers Ed GM3SBC



Joe McElvenney <ximac@...>
 

Hi,

If you mean the common Jupiter-T (TU30-D140), I am owed one after just
completing a GPSDO for a local beacon and so you can have that.

Somewhere I have the correct 20-way field connector but the simplest
way is to solder a ribbon cable directly onto the pins and to use a
tie-wrap around the board to secure it. I actually broke apart a
turned-pin DIL socket and used its individual sockets pushed over the
ends instead.

BTW, if you need the software for a 12F629 PIC to do the
divide-by-thousand trick, instead of using three 74HCT390 chips, I can
help there or can even send you a programmed chip. Apart from the
simplicity it makes a synchronous counter - much better than a series
of ripple-throughs.


Cheers - Joe G3LLV


Andy G4JNT
 

On 25 November 2011 21:55, Joe McElvenney <ximac@...> wrote:
 

Hi,

If you mean the common Jupiter-T (TU30-D140), I am owed one after just
completing a GPSDO for a local beacon and so you can have that.

Somewhere I have the correct 20-way field connector but the simplest
way is to solder a ribbon cable directly onto the pins and to use a
tie-wrap around the board to secure it. I actually broke apart a
turned-pin DIL socket and used its individual sockets pushed over the
ends instead.

BTW, if you need the software for a 12F629 PIC to do the
divide-by-thousand trick, instead of using three 74HCT390 chips, I can
help there or can even send you a programmed chip. Apart from the
simplicity it makes a synchronous counter - much better than a series
of ripple-throughs.

Cheers - Joe G3LLV



Joe McElvenney <ximac@...>
 

Hi,

The same principle applies but I had programmed my first divider PIC a
couple of years before. I had seen a listing some time previous to
that for a divide-by-ten which gave me the idea for a by-1,000. I even
brought it up in "Time-Nuts" but they were ahead of me. As for the
rest of the GPSDO, I used a type-II phase-detector with a current pump
output and a Motorola VC-TCXO.

I've also just re-programmed an RDDS (call sign, etc.) for a local
24GHz beacon which, of course, uses your software. BTW, is the new
clock frequency (11-and a bit megs) something special or just the
result of a junk-box search.


Cheers - Joe G3LLV


Andy G4JNT
 

11.2896MHz   NO it wasn't a junk box crystal ! ! ! ! ! ! ! ! ! ! !
That frequency was very carefully selected after much thought and deliberation and painstaking back of envelope calculations and much burning of midnight oil
 
(Well, to be fair I just glanced though the Farnell catalogue looking for a cheap crystal not too far from 10MHz that divided nidcely by 44100, but the principle holds )
 
The underlying sampling rate in WSJT is that of old soundcards and CDs, ie. 44100Hz.   
 256 * 44100 = 11.2896MHz
 
The latest generation of WSJT / WSPR generating software I've written for PICs  is more versatile. The latest versions can free run and can be used with a variety of crystal frequencies - so long as the freqs meet certain divider criteria.  Of course, free running operation is not suitable for beacon use, but for short term runs - several hours to a day or two - its OK.  There are also traditional GPS times ones too.
 
At the moment code exists only to drive the PIC based DDS for optical subcarrier and baseband IQ conversion, but if/when time and need arises, its only a case of adding the appropriate subroutines for driving DDS chips.   
 
Andy


 

On 25 November 2011 22:32, Joe McElvenney <ximac@...> wrote:
 

Hi,

The same principle applies but I had programmed my first divider PIC a
couple of years before. I had seen a listing some time previous to
that for a divide-by-ten which gave me the idea for a by-1,000. I even
brought it up in "Time-Nuts" but they were ahead of me. As for the
rest of the GPSDO, I used a type-II phase-detector with a current pump
output and a Motorola VC-TCXO.

I've also just re-programmed an RDDS (call sign, etc.) for a local
24GHz beacon which, of course, uses your software. BTW, is the new
clock frequency (11-and a bit megs) something special or just the
result of a junk-box search.

Cheers - Joe G3LLV



Richard <richard@...>
 

On Sat, 26 Nov 2011 10:11:46 +0000
Andy Talbot <andy.g4jnt@gmail.com> wrote:

11.2896MHz NO it wasn't a junk box crystal ! ! ! ! ! ! ! ! ! ! !
That frequency was very carefully selected after much thought and
deliberation and painstaking back of envelope calculations and much burning
of midnight oil

(Well, to be fair I just glanced though the Farnell catalogue looking for a
cheap crystal not too far from 10MHz that divided nidcely by 44100, but the
principle holds )

The underlying sampling rate in WSJT is that of old soundcards and CDs,
ie. 44100Hz.
256 * 44100 = 11.2896MHz

The latest generation of WSJT / WSPR generating software I've written for
PICs is more versatile. The latest versions can free run and can be used
with a variety of crystal frequencies - so long as the freqs meet certain
divider criteria. Of course, free running operation is not suitable for
beacon use, but for short term runs - several hours to a day or two - its
OK. There are also traditional GPS times ones too.

At the moment code exists only to drive the PIC based DDS for optical
subcarrier and baseband IQ conversion, but if/when time and need arises,
its only a case of adding the appropriate subroutines for driving DDS
chips.
See http://www.g4jnt.com/OpticalComms/LF_DDS_Beacon_Source.pdf and
http://www.g4jnt.com/OpticalComms/LF_DDS_Beacon.zip

Andy
www.g4jnt.com

I don't know why Andy you didn't use the 10 MHz reference signal to clock the PIC, and that way the
clock which you depend on for accuracy of tones is locked back to the 10 MHz reference.
Just needs a bit of buffering. Thats what I did to avoid having two 10 MHz sources and intermod products
between the two.



--
Best wishes / 73
Richard Bown

e-mail: richard@g8jvm.com or richard.bown@blueyonder.co.uk

nil carborundum a illegitemis
##################################################################################
Ham Call G8JVM . OS Debian Wheezy/Sid amd64 on a Dual core AMD Athlon 5200, 4 GB RAM
Maidenhead QRA: IO82SP38, LAT. 52 39.720' N LONG. 2 28.171 W ( degs mins )
QRV HF + VHF Microwave 23 cms:140W,13 cms:100W,6 cms:10W & 3 cms:5W
##################################################################################


Andy G4JNT
 

Hindsight is wonderful, isn't it ;-)
yes, of course it would have been the best solution, but I didn't design the RDDS PCB  and it wasn't originally designed with WSJT generation in mind.
 
Within reason the PIC can be clocked at anything, and in fact for WSJT generation 10MHz is not even the best choice, but it was convenient at the time and does give an acceptably low error for ALL modes including WSPR.   there aren't too many frequencies that can do all the modes
 
These days, I tend to go for 19.6609MHz for clocking PICs generally since they all run at 20MHz now (and I've had one rattling away at 35MHz to try it out)  Or for lower current consumption, 3.2768MHz  having bought a batch of 100 of those crystals.
 
Andy

On 26 November 2011 11:08, Richard <richard@...> wrote:
 

On Sat, 26 Nov 2011 10:11:46 +0000
Andy Talbot <andy.g4jnt@...> wrote:

> 11.2896MHz NO it wasn't a junk box crystal ! ! ! ! ! ! ! ! ! ! !
> That frequency was very carefully selected after much thought and
> deliberation and painstaking back of envelope calculations and much burning
> of midnight oil
>
> (Well, to be fair I just glanced though the Farnell catalogue looking for a
> cheap crystal not too far from 10MHz that divided nidcely by 44100, but the
> principle holds )
>
> The underlying sampling rate in WSJT is that of old soundcards and CDs,
> ie. 44100Hz.
> 256 * 44100 = 11.2896MHz
>
> The latest generation of WSJT / WSPR generating software I've written for
> PICs is more versatile. The latest versions can free run and can be used
> with a variety of crystal frequencies - so long as the freqs meet certain
> divider criteria. Of course, free running operation is not suitable for
> beacon use, but for short term runs - several hours to a day or two - its
> OK. There are also traditional GPS times ones too.
>
> At the moment code exists only to drive the PIC based DDS for optical
> subcarrier and baseband IQ conversion, but if/when time and need arises,
> its only a case of adding the appropriate subroutines for driving DDS
> chips.
> See http://www.g4jnt.com/OpticalComms/LF_DDS_Beacon_Source.pdf and
> http://www.g4jnt.com/OpticalComms/LF_DDS_Beacon.zip
>
> Andy
> www.g4jnt.com
>
>
I don't know why Andy you didn't use the 10 MHz reference signal to clock the PIC, and that way the
clock which you depend on for accuracy of tones is locked back to the 10 MHz reference.
Just needs a bit of buffering. Thats what I did to avoid having two 10 MHz sources and intermod products
between the two.

--
Best wishes / 73
Richard Bown

e-mail: richard@... or richard.bown@...

nil carborundum a illegitemis
##################################################################################
Ham Call G8JVM . OS Debian Wheezy/Sid amd64 on a Dual core AMD Athlon 5200, 4 GB RAM
Maidenhead QRA: IO82SP38, LAT. 52 39.720' N LONG. 2 28.171 W ( degs mins )
QRV HF + VHF Microwave 23 cms:140W,13 cms:100W,6 cms:10W & 3 cms:5W
##################################################################################



Ed <ed.murphy31@...>
 

Hi Joe,
Thanks for the offer. I think that part will fit into the present box layout. I already have a 20 way connector and cable assembly. I wonder if the part you can get is the TU30-D140-211 fitted with the SMB rt angle connector. I could fit one if needed. re offer of Pic that would be an interesting way to go. Possibly best way forward is to email via QRZ.com if that is ok with you.

Thanks. Ed GM3SBC IO85IX
( Seems I have started some discussion here ! )

--- In ukmicrowaves@yahoogroups.com, Joe McElvenney <ximac@...> wrote:

Hi,

If you mean the common Jupiter-T (TU30-D140), I am owed one after just
completing a GPSDO for a local beacon and so you can have that.

Somewhere I have the correct 20-way field connector but the simplest
way is to solder a ribbon cable directly onto the pins and to use a
tie-wrap around the board to secure it. I actually broke apart a
turned-pin DIL socket and used its individual sockets pushed over the
ends instead.

BTW, if you need the software for a 12F629 PIC to do the
divide-by-thousand trick, instead of using three 74HCT390 chips, I can
help there or can even send you a programmed chip. Apart from the
simplicity it makes a synchronous counter - much better than a series
of ripple-throughs.


Cheers - Joe G3LLV


Richard <richard@...>
 

On Sat, 26 Nov 2011 11:19:56 +0000
Andy Talbot <andy.g4jnt@gmail.com> wrote:

Hindsight is wonderful, isn't it ;-)
yes, of course it would have been the best solution, but I didn't design
the RDDS PCB and it wasn't originally designed with WSJT generation in
mind.

Within reason the PIC can be clocked at anything, and in fact for WSJT
generation 10MHz is not even the best choice, but it was convenient at the
time and does give an acceptably low error for ALL modes including WSPR.
there aren't too many frequencies that can do all the modes

These days, I tend to go for 19.6609MHz for clocking PICs generally since
they all run at 20MHz now (and I've had one rattling away at 35MHz to try
it out) Or for lower current consumption, 3.2768MHz having bought a batch
of 100 of those crystals.

Andy

On 26 November 2011 11:08, Richard <richard@g8jvm.com> wrote:

**


On Sat, 26 Nov 2011 10:11:46 +0000
Andy Talbot <andy.g4jnt@gmail.com> wrote:

11.2896MHz NO it wasn't a junk box crystal ! ! ! ! ! ! ! ! ! ! !
That frequency was very carefully selected after much thought and
deliberation and painstaking back of envelope calculations and much
burning
of midnight oil

(Well, to be fair I just glanced though the Farnell catalogue looking
for a
cheap crystal not too far from 10MHz that divided nidcely by 44100, but
the
principle holds )

The underlying sampling rate in WSJT is that of old soundcards and CDs,
ie. 44100Hz.
256 * 44100 = 11.2896MHz

The latest generation of WSJT / WSPR generating software I've written for
PICs is more versatile. The latest versions can free run and can be used
with a variety of crystal frequencies - so long as the freqs meet certain
divider criteria. Of course, free running operation is not suitable for
beacon use, but for short term runs - several hours to a day or two - its
OK. There are also traditional GPS times ones too.

At the moment code exists only to drive the PIC based DDS for optical
subcarrier and baseband IQ conversion, but if/when time and need arises,
its only a case of adding the appropriate subroutines for driving DDS
chips.
See http://www.g4jnt.com/OpticalComms/LF_DDS_Beacon_Source.pdf and
http://www.g4jnt.com/OpticalComms/LF_DDS_Beacon.zip

Andy
www.g4jnt.com

I don't know why Andy you didn't use the 10 MHz reference signal to clock
the PIC, and that way the
clock which you depend on for accuracy of tones is locked back to the 10
MHz reference.
Just needs a bit of buffering. Thats what I did to avoid having two 10 MHz
sources and intermod products
between the two.

--
Best wishes / 73
Richard Bown

e-mail: richard@g8jvm.com or richard.bown@blueyonder.co.uk

nil carborundum a illegitemis

##################################################################################
Ham Call G8JVM . OS Debian Wheezy/Sid amd64 on a Dual core AMD Athlon
5200, 4 GB RAM
Maidenhead QRA: IO82SP38, LAT. 52 39.720' N LONG. 2 28.171 W ( degs mins )
QRV HF + VHF Microwave 23 cms:140W,13 cms:100W,6 cms:10W & 3 cms:5W
##################################################################################



Just a case of drilling an extra hole for a connector and coax to where the xtal was.

I dont know why it wasn't done in the design.
as the damn meercat says "simple"

--
Best wishes / 73
Richard Bown

e-mail: richard@g8jvm.com or richard.bown@blueyonder.co.uk

nil carborundum a illegitemis
##################################################################################
Ham Call G8JVM . OS Debian Wheezy/Sid amd64 on a Dual core AMD Athlon 5200, 4 GB RAM
Maidenhead QRA: IO82SP38, LAT. 52 39.720' N LONG. 2 28.171 W ( degs mins )
QRV HF + VHF Microwave 23 cms:140W,13 cms:100W,6 cms:10W & 3 cms:5W
##################################################################################


Joe McElvenney <ximac@...>
 

Ed,

Don't worry about starting a thread because the more folks who see a
posting, the more likely it is there is someone who can help and,
anyway, it gives you a chance to strut your stuff. And, in any event,
just look at RSGB-Tech where everyone and his canine friend seems to
reply to whatever is put up there.

Now regrettably, the connector is a for a standard MCX (OSX) plug
either right-angled or straight. I cannot give you one of my spare
plugs/leads but can let you have a straight bulkhead, crimp MCX either
with a 4-inch tail or a new plug (better you than me when soldering
the pin).

I will post you the details later today. All I ask is that you pray
for either of the Sheffield teams to get promotion next year and to
refrain from eating white pudding or deep-fried Mars Bars over the
Festive Season.

I assume that you are QTHR?


Cheers - Joe G3LLV


Ed <ed.murphy31@...>
 

LOL ,
Yes Joe I am QTHR. But I must confess I did have a very nice helping of BLACK pudding this morning.
Thanks for all your help, I shall sort something out regarding the connector. I am about to embark on the HiQSDR project. Have a look, you will see how good I THINK I am to even consider the soldering required.
Best Regards,
Ed GM3SBC

--- In ukmicrowaves@yahoogroups.com, Joe McElvenney <ximac@...> wrote:

Ed,

Don't worry about starting a thread because the more folks who see a
posting, the more likely it is there is someone who can help and,
anyway, it gives you a chance to strut your stuff. And, in any event,
just look at RSGB-Tech where everyone and his canine friend seems to
reply to whatever is put up there.

Now regrettably, the connector is a for a standard MCX (OSX) plug
either right-angled or straight. I cannot give you one of my spare
plugs/leads but can let you have a straight bulkhead, crimp MCX either
with a 4-inch tail or a new plug (better you than me when soldering
the pin).

I will post you the details later today. All I ask is that you pray
for either of the Sheffield teams to get promotion next year and to
refrain from eating white pudding or deep-fried Mars Bars over the
Festive Season.

I assume that you are QTHR?


Cheers - Joe G3LLV