Date   

Re: Just a thought ....................

Mike Willis
 

Thanks - enough information to work out what you are talking about next month!
--
Mike G0MJW


Re: Just a thought ....................

Neil Smith G4DBN
 

The billet gets a bit enormous at 3.4GHz. Rather wasteful. Now if I had a coring trepanning tool, that might make a difference. I made the 5.6GHz super VE4MA horns from tube then welded on the socket mounts and choke ring. Never again though. The current batch have a rear section up to the socket and probe which is machined from solid, then the tube is welded on to that.

The bigger choked feedhorns in my order book are for 13cm and 23cm EME, with 7-16 DINs and stepped septums, so I'll roll the tubes from sheet and TIG them along the seams. If I can then sweat on the mounting blocks and stiffening rings using ally solder, that would make things easier. If it is rubbish, then I will epoxy them on instead.

I do lap the faces of the WA6KBL feedhorns against the flat face of the w/g flange.

If the solder is no good for WR90 flanges, I'll give epoxy and lapping a try.

Neil G4DBN

-------- Original message --------
From: Mike Willis <willis.mj@...>
Date: 05/09/2019 13:56 (GMT+00:00)
To: UKMicrowaves@groups.io
Subject: Re: [UKMicrowaves] Just a thought ....................

The signals are in the wavguide, the idea is to have a continuous gap less guide. A choke flange is the exception with a deliberate tuned gap. As long as you have carefully machined the waveguide to mate properly at the transition you don't need to use a choke flange, and never use two together.. getting a perfect mating with no gap is more or less impossible (hence the choke) but close enough is possible and usually OK for us.

I am surprised, with your machining skills, that you don't machine that mounting block out of a solid billet.

--
Mike G0MJW


Re: Just a thought ....................

Chris Bartram G4DGU
 

The reason for using choke flanges derives from the particular problems of high-power radar systems, where MW (not mW!) are flowing down the pipe. The use of choke flanges very much reduces the risks of flashovers, and radiation. With amateur systems we're not in that power range ...

There are times when putting a small gap between the ends of waveguides is considered quite normal. One quite well known use is when making a 'thermal break' between a cryogenically cooled LNA and waveguide operating at ambient temperature. This is quite comparable in performance to the use of thinwall stainless steel waveguide for thermal isolation. FWIW, I've been looking at the WG gap approach for a 24GHz true WG CCLNA ie. one where the input match is done in the guide, not with a 50ohm coax transition and external matching circuit, probably using microstrip, as many seem to be.

I'll have a rant about the pros and cons of microstrip another time!!

73

Chris

G4DGU


Re: Just a thought ....................

Mike Willis
 

The signals are in the wavguide, the idea is to have a continuous gap less guide. A choke flange is the exception with a deliberate tuned gap. As long as you have carefully machined the waveguide to mate properly at the transition you don't need to use a choke flange, and never use two together.. getting a perfect mating with no gap is more or less impossible (hence the choke) but close enough is possible and usually OK for us.

I am surprised, with your machining skills, that you don't machine that mounting block out of a solid billet.

--
Mike G0MJW


Re: Just a thought ....................

Neil Smith G4DBN
 

My lack of education is showing... I thought it did matter when you used a choke flange on the thing you were bolting the flange on to.

Otherwise why not use a Delrin or PEEK flange?

My specific use case for the ally solder is fixing a mounting block for a coax socket on a circular feedhorn.

Neil G4DBN

-------- Original message --------
From: Mike Willis <willis.mj@...>
Date: 05/09/2019 10:51 (GMT+00:00)
To: UKMicrowaves@groups.io
Subject: Re: [UKMicrowaves] Just a thought ....................

Why not use areldite? The flange is not an electrical connection.
--
Mike G0MJW


Re: Réf. : Re: [UKMicrowaves] Modded LNB

Paul G8KFW
 

Hi I thought it was Time Nuts

 


From: UKMicrowaves@groups.io [mailto:UKMicrowaves@groups.io] On Behalf Of John Fell
Sent: 05 September 2019 12:29
To: UKMicrowaves@groups.io
Subject: Re: Réf. : Re: [UKMicrowaves] Modded LNB

 

Oh no - the Time Lords are out of their Tardis again !

 

Mumbles .......it's all relative ...departs ...

 

John

G0API

 

On Thu, 5 Sep 2019 at 12:26, Andy G4JNT <andy.g4jnt@...> wrote:

26MHz is "approximately  25.78846153846154MHz"

 

Andy

 

 

 

On Thu, 5 Sep 2019 at 11:42, Mike Willis <willis.mj@...> wrote:

On Thu, Sep 5, 2019 at 02:54 AM, Paul M0EYT wrote:

I'm running with a clock approximately 25.78846153846154MHz

Approximately? to 1 part in 10^16 !
 
--
Mike G0MJW

No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.8048 / Virus Database: 4793/15884 - Release Date: 08/14/18
Internal Virus Database is out of date.


--
Paul Bicknell G8KFW   South Coast UK


Re: Just a thought ....................

Stuart G1ZAR
 

"I believe Ken, G3YKI, published a design for a copper pipe antenna for 10GHz some years ago in Scatterpoint."

or the previous Microwave Newsletter. ( and maybe an RSGB publication too).


Re: Réf. : Re: [UKMicrowaves] Modded LNB

John Fell
 

Oh no - the Time Lords are out of their Tardis again !

Mumbles .......it's all relative ...departs ...

John
G0API

On Thu, 5 Sep 2019 at 12:26, Andy G4JNT <andy.g4jnt@...> wrote:
26MHz is "approximately  25.78846153846154MHz"

Andy



On Thu, 5 Sep 2019 at 11:42, Mike Willis <willis.mj@...> wrote:
On Thu, Sep 5, 2019 at 02:54 AM, Paul M0EYT wrote:
I'm running with a clock approximately 25.78846153846154MHz
Approximately? to 1 part in 10^16 !
 
--
Mike G0MJW


Re: Réf. : Re: [UKMicrowaves] Modded LNB

Andy G4JNT
 

26MHz is "approximately  25.78846153846154MHz"

Andy



On Thu, 5 Sep 2019 at 11:42, Mike Willis <willis.mj@...> wrote:
On Thu, Sep 5, 2019 at 02:54 AM, Paul M0EYT wrote:
I'm running with a clock approximately 25.78846153846154MHz
Approximately? to 1 part in 10^16 !
 
--
Mike G0MJW


Re: Réf. : Re: [UKMicrowaves] Modded LNB

Chris Bartram G4DGU
 

Mike,

Most, if not all, physical measurements are approximate to some extent! One of my ongoing fascinations with engineering is the need to balance approximations to make something which will do what's needed - usually meet an arbitrary specification ... :-)

73

Chris

G4DGU


On 05/09/2019 11:42, Mike Willis wrote:
On Thu, Sep 5, 2019 at 02:54 AM, Paul M0EYT wrote:
I'm running with a clock approximately 25.78846153846154MHz
Approximately? to 1 part in 10^16 !
 
--
Mike G0MJW


Re: Réf. : Re: [UKMicrowaves] Modded LNB

Mike Willis
 

On Thu, Sep 5, 2019 at 02:54 AM, Paul M0EYT wrote:
I'm running with a clock approximately 25.78846153846154MHz
Approximately? to 1 part in 10^16 !
 
--
Mike G0MJW


Re: Réf. : Re: [UKMicrowaves] Modded LNB

Paul M0EYT
 

Hi Frank, I have one of the Golden Interstar GI202 LNB's and am using that with a 70cm IF. I'm running with a clock approximately 25.78846153846154MHz which puts the QO100 PSK beacon down-link on 432.300MHz on the IF and I can confirm it works fine.

Regards,

Paul M0EYT.


Re: Réf. : Re: [UKMicrowaves] Modded LNB

Mike Willis
 

Both an octagon single and an octagon quad worked for me. YMMV
--
Mike G0MJW


Re: Just a thought ....................

Mike Willis
 

Why not use areldite? The flange is not an electrical connection.
--
Mike G0MJW


Re: Just a thought ....................

John Fell
 

I have made many slotted arrays from copper pipe and waveguides ( the GB3SC series on Bell Hill for instance) .

I agree slotting copper on a mill is an interesting exercise ....I used Ramin hardwood for the sacrificial mandrel for 2.3 and 3.4GHZ Alford Slots .

However if I were to make any more slotted assemblies , I would send a Cad file to a local CNC Laser or High pressure Water cutting company .

If  a group needs 10GHZ slotted omnis , the cost would be very reasonable compared to any individual machined alternatives .

Talking of 10GHZ antennas , is there any news about the West Wales beacon that Paul M0EYT , Adrian G4UVZ and I made the slotted array for some 2 years ago ?

John
G0API

On Thu, 5 Sep 2019 at 09:49, Neil Smith G4DBN <neil@...> wrote:
There are loads of designs published using alternating diagonal slots and other shapes cut into round tube, but they are mostly the product of PhD students playing around with EM solvers. Lots of them are behind academic paywalls of course. Simple to machine if you happen to have a five-axis CNC machining centre in the Uni engineering workshop.  There is a standard sized aluminium extrusion that is just about right for reduced-height guide on 10GHz, and is easy to weld with a TIG, and so much easier to machine than copper w/g, which is made of a nasty gummy, grabby copper. Very easy to weld on some wings to improve the pattern as well. 

22mm copper pipe is a swine to hold and machine over any significant length, it is just waiting for a chance to rip itself apart. I tend to use brass tube, or bored solid brass bar, as it machines well and doesn’t self-destruct.  It is probably possible to use a sacrificial mandrel and collet but you’d need to find a way to step it, or use a good solid mandrel between centres. Then you have to work out how to remove the mandrel as the burrs on the copper will lock it solid. It isn’t even a simple shape to make with 3D sintered metal printing. 

I’m trying out some of the new aluminium solder pastes to see how easy it is to join an aluminium flange to an aluminium waveguide without using TIG welding.  There are now pastes which allow soldering of aluminium to brass or copper, but we shall see how they work in practice…

Neil G4DBN

On 5 Sep 2019, at 09:25, Mike Willis <willis.mj@...> wrote:

You could try a few designs in an electromagnetics simulator. CST is easy to use.

The problem with circular waveguide is keeping the right mode, that's why we use elliptical guide or square guide. As square guide has always been so cheap, it's not really worth using copper water pipe. Surplus waveguide is not so easy to find these days but it does not evaporate or rust, so it must be out there somewhere.
--
Mike G0MJW


Re: Just a thought ....................

 

Peter, my unconventional antenna used WG16 but instead of slots had round holes, and so could be made without using machine tools.
Ken

I believe Ken, G3YKI, published a design for a copper pipe antenna for 10GHz some years ago in Scatterpoint. I can't lay my hands on it at the moment but Ken might be reading this and reply.
 


Re: Réf. : Re: [UKMicrowaves] Modded LNB

Paul G8KFW
 

Hi Mike

I am told by others that moving the local oscillator down enough to get the IF for oscar 100 at 1240 mhz  is not that easy

as apparently most PLL will not  work  so could any one com beck that has actually achieved this without degrading the phase noise

 

My solution is to build an up converter to 1240

 

PS I have found on ebay  GNSS Filter GPS L1-L5 GLONASS BeiDou Navic 1100-1700 MHz band Filter

That seams ideal for the project

 

Anny comments welcome   Paul


From: UKMicrowaves@groups.io [mailto:UKMicrowaves@groups.io] On Behalf Of Mike Willis
Sent: 05 September 2019 08:48
To: UKMicrowaves@groups.io
Subject: Re: Réf. : Re: [UKMicrowaves] Modded LNB

 

Not all PLLs will lock, it depends on the chip and it's circuits. Some will lock over 1 MHz either side of 25 MHz, others much less. It's not just the lock but also the phase noise which you need to pay attention to.

I don't understand why you would want to inject 26 MHz into a 25MHz LNB. The multiple is 390 and that would give you and LO of 10140 MHz. The transponder would appear at 350 MHz, well outside the normal IF range and most, not all LNBs, if they did lock, would be well down on gain at that point. The older Octagon LNBs used a 27 MHz crystal and maybe this is why you are trying 26 MHz. This would give an LO of 9388.8888.. and IF around 1.1 GHz.

A much more common choice for nominally 25 MHz reference LNBs is 24 MHz giving a 9360 MHz LO and the transponder appearing at 1129MHz, within the normal IF range. Not all LNBs lock to 24 MHz.

Some people have had success with an LO at 10050 MHz that puts the IF in 70cm. It depends on the PLL and IF gain. Personally I use  9500 MHz (24.35897436 MHz in)
which is within the range of my Octagon LNB and  puts the IF at 989MHz which is within the normal IF range allowing a standard satellite receiver to tune as well as SDRs etc.

Mike G0MJW

No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.8048 / Virus Database: 4793/15884 - Release Date: 08/14/18
Internal Virus Database is out of date.


--
Paul Bicknell G8KFW   South Coast UK


Re: Just a thought ....................

Neil Smith G4DBN
 

There are loads of designs published using alternating diagonal slots and other shapes cut into round tube, but they are mostly the product of PhD students playing around with EM solvers. Lots of them are behind academic paywalls of course. Simple to machine if you happen to have a five-axis CNC machining centre in the Uni engineering workshop.  There is a standard sized aluminium extrusion that is just about right for reduced-height guide on 10GHz, and is easy to weld with a TIG, and so much easier to machine than copper w/g, which is made of a nasty gummy, grabby copper. Very easy to weld on some wings to improve the pattern as well. 

22mm copper pipe is a swine to hold and machine over any significant length, it is just waiting for a chance to rip itself apart. I tend to use brass tube, or bored solid brass bar, as it machines well and doesn’t self-destruct.  It is probably possible to use a sacrificial mandrel and collet but you’d need to find a way to step it, or use a good solid mandrel between centres. Then you have to work out how to remove the mandrel as the burrs on the copper will lock it solid. It isn’t even a simple shape to make with 3D sintered metal printing. 

I’m trying out some of the new aluminium solder pastes to see how easy it is to join an aluminium flange to an aluminium waveguide without using TIG welding.  There are now pastes which allow soldering of aluminium to brass or copper, but we shall see how they work in practice…

Neil G4DBN

On 5 Sep 2019, at 09:25, Mike Willis <willis.mj@...> wrote:

You could try a few designs in an electromagnetics simulator. CST is easy to use.

The problem with circular waveguide is keeping the right mode, that's why we use elliptical guide or square guide. As square guide has always been so cheap, it's not really worth using copper water pipe. Surplus waveguide is not so easy to find these days but it does not evaporate or rust, so it must be out there somewhere.
--
Mike G0MJW


Re: Just a thought ....................

John Fell
 

22mm Copper pipe is a good material to work with for 10GHZ , but I would stick with WG16/17 .
Getting a phase array of slots to radiate well is not something you can do easily but there are good designs available that will work "off the peg" for standard rectangular guide - G8AGN etc .

The 22mmm pipe is best used to make the transition , if required , to match the rectangular guide - needs a vice and a pair of Swan neck pliers .

Back in 1994 when testing feeds for 10GHZ EME , I made a Swan neck feed for a 0.43 f/D prime dish (3.4m diameter). 

The insertion loss was high - my Friend Julian GYGF demonstrated that losses reduced by conversion to rectangular guide , using a Lump Hammer ...... I still have the item in the garage .EW 90 eliptic guide works well at 10GHZ .....

I am still using copper heating pipe , suitably (carefully) rectangularised , for my 24GHZ RX system and this seems to still have good low loss characteristics 

73
John
G0API

On Thu, 5 Sep 2019 at 09:25, Mike Willis <willis.mj@...> wrote:
You could try a few designs in an electromagnetics simulator. CST is easy to use.

The problem with circular waveguide is keeping the right mode, that's why we use elliptical guide or square guide. As square guide has always been so cheap, it's not really worth using copper water pipe. Surplus waveguide is not so easy to find these days but it does not evaporate or rust, so it must be out there somewhere.
--
Mike G0MJW


Re: Modded LNB

John Fell
 

Frank,
I think Paul M0EYT uses a 24MHZ reference on the Interstar LNB , with an i.f. output in the 23cm area .

If you have a reasonable sig gen try coupling the 0dBm ish output across 50 Ohms and take a 1nF ceramic Cap to the pin of the wire ended Xtal closest to the centre of the LNB .
It should lock up when within a few Hz at 25MHZ and you can then experiment by slowly moving the generator frequency .

On an SDR or Spec An you will see the capture points - it will go into dither mode and beyond loose lock .Also test the frequency offset from nominal that lock up will readily occur from a cold start.

The good thing with the Interstar is the ease of de-soldering the leaded Xtal and the space available to add in stuf - all with the primary lid removed and the RF stages operating under separate covers , so no stability of frequency pulling issues .

For the availability and cost , this LNB is my current best buy for 10GHZ operation , including EME .

73
John
G0API

On Wed, 4 Sep 2019 at 22:49, Frank <frankg3rmd@...> wrote:
John,
I have a Golden Interstar dual LNB which works very well in it's unmodified state with usual PLL drift characteristics. I have a spare 26MHz VCXO and have tried removing one of the 25MHz crystals and injecting my 26MHz source to no avail. Has anyone modified this LNB for external or internal VCXO and experimented with it's ref lock range?
My 26MHz source can provide 0dbm to the PLL input. 
73
Frank
G3RMD

12441 - 12460 of 64011