mydan
Member
- My Satellite Setup
- Historically, D-mac (Ferguson) with D2-mac mod, followed by Echostar DVB with various CAMs.
- My Location
- Helmshore, England
I'd like to share progress on a Squarial revival project I've been faffing with, on and off, for several weeks. The aim is to get the old Matsushita (Panasonic) version picking up the main-beam Freesat channels at 28.2°E (the 100-odd FTA channels) and, as a bonus, anything it can manage from the European beams. I've hit a brick wall near the prototype finish line and would really appreciate thoughts from anyone in the community who's got an interest or expertise.
A bit of my background - Lancashire lad, born in 1982, I grew up playing with BSB kit after it went bust. My dad and I did the film-flip for LHCP (CQ-TV issue 159) and we even bought the D2-MAC EEPROM conversion from Trac for the Ferguson receiver we were butchering. (Remember the back page of What Satellite?) The Squarial made a big impression on me as a kid and I'm sure it did for all of us who were around about the time. It'd be nice to see some out there delivering satellite signals as they themselves, for direct-to-home TV use anyway, enter their twilight years.
Fast-forward 35 years and there are Squarials on eBay. Nostalgia won over, I bought a small collection, and, as is typical for BSB it seems, even 35 years after its demise, a bloody expensive and energy sapping project was born. I have a background in software development with some mechanical and electrical engineering thrown in, but antenna design and microwave behaviour are very much "learned from reading and AI advice" territory for me.
The original film-flip mod was pointless (it only gives LHCP) and the factory fit single-probe LNB (vertical, noisy 2.5 dB, 10 GHz LO) had to go. In its place I wanted a modern 0.1–0.2 dB universal LNB. I did some searching and came across CQ-TV issue 234 which detailed a Linear-H polarity mod (which is part of this build) which changes the Squarial so it can receive horizontal polarisation from satellites, which brings a modern universal LNB somewhat within reach.
One important point: the vertical meander line and horizontal waveguide in the Squarial deliver the horizontal satellite signal to the probe in a vertical orientation. As far as I understand it, that is a fixed physical constraint, due to the shape of the waveguide and the fact there's only one vertical meander line. So the LNB <-> Receiver side needs to be clever. It can be confusing though to think no matter the polarity from the satellite, it's always delivered vertically.
We already know a universal LNB has two probes: horizontal (selected by 18 V) and vertical (selected by 13 V). In the Squarial the horizontal probe is useless because nothing gets through the waveguide. The plan is to keep the vertical probe permanently active by forcing 13 V to the LNB regardless of what the receiver sends. I'll describe the simple voltage-regulator circuit later if anyone wants it (or the project moves forward); it splits the coax so the IF passes through DC-blocking caps while a separate path feeds a regulator (with chokes and caps) to give a steady ~12–13 V. I also tap off a voltage divided 1–3 V signal before the regulator so a microcontroller can tell whether the receiver is asking for H or V.
That leads to the core idea: use an ESP32 and a small weatherproof servo to rotate the entire Squarial 90 degrees on its boresight axis. 18 V = "Home" position (BSB diamond at the bottom, natural orientation) for horizontal. 13 V = "Offset" position (diamond on its side) for vertical. Because the vertical probe is always live and the dish stays pointed at the satellite, vertical signals should appear once the rotation is correct.
I've grouped my project as it has grown into the following steps:
Job 1: Linear-H Mod
I started with a bog standard Ferguson Squarial and followed the Linear-H mod from CQ-TV Issue 234. The original pictures and instructions are a bit vague, so it took a while to get my head around it, but the process is actually straightforward once it clicks.
Remove the 12 screws from the back and lift off the front cover. Place the dish face-down on the table in its Home orientation (BSB diamond at the bottom).
First layer is the metal aperture plate – up-arrow and "1" on that. I also put an up arrow on the waveguide cover. Remove its four screws and set it aside. Next is the foam layer (up-arrow + "2"). Then the printed patch-array layer (up-arrow + "3"). Finally the rear foam (up-arrow + "4").
The metal aperture plate has 256 apertures feeding a 16×16 patch array. The original design splits the 256 patches into two 128-element groups that are deliberately 90 degrees out of phase to produce RHCP. These two groups are power-combined at the central junction while preserving the phase offset.
The mod itself: on the printed patch layer you’ll see four small metal blue strips in the corners (presumably manufacturing tabs for guillotining). Cut one out, complete with several millimetres of clear film to the side.
Look at the vertical meander line that feeds the waveguide. The factory slit is offset "south" – roughly centred 16 mm from the bottom of the 42 mm (outer to outer) meander line. Cut a new slit exactly in the middle (21 mm centre) with a sharp Stanley knife. Glue the "quarter-wave stub" you cut earlier directly over the original (southern) slit using clear nail polish – it doesn’t need to be electrically bonded; it's a microwave antenna. Following John's guidance, I used my wife's clear nail polish too
The following pictures should show the process:
Original meander line with offset slit
New slit at 21mm (centre of meander) and stub glued in place:
With the new central slit, the 90 degree phase offset disappears and horizontal signals are delivered to the vertical probe. Because the new slit must sit dead-centre in the waveguide, the whole film has to then be pulled down about 5 mm. I heated a bolt with a blowtorch and punched new mounting holes through the film.
Holes marked and punched 5mm lower:
Positioned on waveguide centrally:
Mounting holes equally punched:
Then everything is reassembled layer by layer in the correct orientation. That's the Linear-H mod done.
Due to picture limits, I'll need to post progress over several posts. Bear with me!
A bit of my background - Lancashire lad, born in 1982, I grew up playing with BSB kit after it went bust. My dad and I did the film-flip for LHCP (CQ-TV issue 159) and we even bought the D2-MAC EEPROM conversion from Trac for the Ferguson receiver we were butchering. (Remember the back page of What Satellite?) The Squarial made a big impression on me as a kid and I'm sure it did for all of us who were around about the time. It'd be nice to see some out there delivering satellite signals as they themselves, for direct-to-home TV use anyway, enter their twilight years.
Fast-forward 35 years and there are Squarials on eBay. Nostalgia won over, I bought a small collection, and, as is typical for BSB it seems, even 35 years after its demise, a bloody expensive and energy sapping project was born. I have a background in software development with some mechanical and electrical engineering thrown in, but antenna design and microwave behaviour are very much "learned from reading and AI advice" territory for me.
The original film-flip mod was pointless (it only gives LHCP) and the factory fit single-probe LNB (vertical, noisy 2.5 dB, 10 GHz LO) had to go. In its place I wanted a modern 0.1–0.2 dB universal LNB. I did some searching and came across CQ-TV issue 234 which detailed a Linear-H polarity mod (which is part of this build) which changes the Squarial so it can receive horizontal polarisation from satellites, which brings a modern universal LNB somewhat within reach.
One important point: the vertical meander line and horizontal waveguide in the Squarial deliver the horizontal satellite signal to the probe in a vertical orientation. As far as I understand it, that is a fixed physical constraint, due to the shape of the waveguide and the fact there's only one vertical meander line. So the LNB <-> Receiver side needs to be clever. It can be confusing though to think no matter the polarity from the satellite, it's always delivered vertically.
We already know a universal LNB has two probes: horizontal (selected by 18 V) and vertical (selected by 13 V). In the Squarial the horizontal probe is useless because nothing gets through the waveguide. The plan is to keep the vertical probe permanently active by forcing 13 V to the LNB regardless of what the receiver sends. I'll describe the simple voltage-regulator circuit later if anyone wants it (or the project moves forward); it splits the coax so the IF passes through DC-blocking caps while a separate path feeds a regulator (with chokes and caps) to give a steady ~12–13 V. I also tap off a voltage divided 1–3 V signal before the regulator so a microcontroller can tell whether the receiver is asking for H or V.
That leads to the core idea: use an ESP32 and a small weatherproof servo to rotate the entire Squarial 90 degrees on its boresight axis. 18 V = "Home" position (BSB diamond at the bottom, natural orientation) for horizontal. 13 V = "Offset" position (diamond on its side) for vertical. Because the vertical probe is always live and the dish stays pointed at the satellite, vertical signals should appear once the rotation is correct.
I've grouped my project as it has grown into the following steps:
- Perform the linear-H mod
- Replace the factory LNB with a modern unit (modifying the casing as needed)
- Fabricate a new central bracket with bearings so a pole passes exactly through the geometric centre / boresight axis (think "spinning a plate on a stick")
- Mount the pole on a simple adjustable yoke for testing
- Test Home (H) and Offset (V) by hand
- If it works, add a second pole and servo with GT2 belt (2:1 or 3:1 ratio)
- Build the voltage-sensing circuit and write the microcontroller code
Job 1: Linear-H Mod
I started with a bog standard Ferguson Squarial and followed the Linear-H mod from CQ-TV Issue 234. The original pictures and instructions are a bit vague, so it took a while to get my head around it, but the process is actually straightforward once it clicks.
Remove the 12 screws from the back and lift off the front cover. Place the dish face-down on the table in its Home orientation (BSB diamond at the bottom).
First layer is the metal aperture plate – up-arrow and "1" on that. I also put an up arrow on the waveguide cover. Remove its four screws and set it aside. Next is the foam layer (up-arrow + "2"). Then the printed patch-array layer (up-arrow + "3"). Finally the rear foam (up-arrow + "4").
The metal aperture plate has 256 apertures feeding a 16×16 patch array. The original design splits the 256 patches into two 128-element groups that are deliberately 90 degrees out of phase to produce RHCP. These two groups are power-combined at the central junction while preserving the phase offset.
The mod itself: on the printed patch layer you’ll see four small metal blue strips in the corners (presumably manufacturing tabs for guillotining). Cut one out, complete with several millimetres of clear film to the side.
Look at the vertical meander line that feeds the waveguide. The factory slit is offset "south" – roughly centred 16 mm from the bottom of the 42 mm (outer to outer) meander line. Cut a new slit exactly in the middle (21 mm centre) with a sharp Stanley knife. Glue the "quarter-wave stub" you cut earlier directly over the original (southern) slit using clear nail polish – it doesn’t need to be electrically bonded; it's a microwave antenna. Following John's guidance, I used my wife's clear nail polish too
The following pictures should show the process:
Original meander line with offset slit
New slit at 21mm (centre of meander) and stub glued in place:
With the new central slit, the 90 degree phase offset disappears and horizontal signals are delivered to the vertical probe. Because the new slit must sit dead-centre in the waveguide, the whole film has to then be pulled down about 5 mm. I heated a bolt with a blowtorch and punched new mounting holes through the film.
Holes marked and punched 5mm lower:
Positioned on waveguide centrally:
Mounting holes equally punched:
Then everything is reassembled layer by layer in the correct orientation. That's the Linear-H mod done.
Due to picture limits, I'll need to post progress over several posts. Bear with me!
. It's single-output (perfect, because the servo rotation demands means only one receiver can ever be connected) and its body is at 90 degrees to the flange, so it fits the Squarial with almost no case butchery. I got mine for £27 on eBay, the prices fluctuate by a few quid.
.
00MHz either side of 12GHz and something the Astra 28 East bunch of birds far exceeds (both under and over) .
. It may be that 11386 on both polarities at - 8PSK - 27500 is simply too much, but I'll keep persevering a bit longer.
