-
80GHz Radio Gigabeam WiFiber Teardown
Posted by
voltlog
on 28 Sep, 2015 07:16
-
This is me, doing a teardown of the Gigabeam Wifiber Radio which operates in the 80GHz range.
I start from disassembling the die-cast enclosure of the radio, then I continue with presenting the digital section and in the end I open the box where the high frequency RF is generated.
I have two of these radio's complete with antenna and one of them is broken showing an error about the "Costas loop".
This is a PLL circuit used for carrier phase recovery in this type of millimeter wave radio. So the main reason for the teardown is to have a base for a repair attempt and also to take a look at some cool gear.
I hope you will find this interesting:
-
#1 Reply
Posted by
lukier
on 28 Sep, 2015 08:02
-
Thanks for sharing, very informative teardown
This is some serious RF voodo
Good to know you can pass RF signals through a zebra strip
-
#2 Reply
Posted by
voltlog
on 28 Sep, 2015 08:05
-
I guess that's impedance controlled zebra strip
-
#3 Reply
Posted by
daqq
on 28 Sep, 2015 08:26
-
Thanks for the teardown!
Though I think that the middle RF board does not provide the IF frequency, but rather is the local oscillator - the waveguide on its output is far too small for for 4-8 GHz.
-
#4 Reply
Posted by
voltlog
on 28 Sep, 2015 09:03
-
Though I think that the middle RF board does not provide the IF frequency, but rather is the local oscillator - the waveguide on its output is far too small for for 4-8 GHz.
Good point! So it must be in the 80GHz range to travel that rather small waveguide.
I just couldn't figure out how their multiplying it that high on that board (Silkscreen "P3 SYNTHESIZER")
Maybe there is some multiplication happening on the small hybrid board/module right to the input of the waveguide.
-
#5 Reply
Posted by
daqq
on 28 Sep, 2015 09:36
-
Maybe there is some multiplication happening on the small hybrid board/module right to the input of the waveguide.
That is the case most likely. Intermediate frequencies for stuff like this is (as far as I know) generally in the hundreds of MHz regions.
See
https://youtu.be/TQWKbZo6UCY?t=1419 - at the bottom, the relatively low frequency enters, it exits in the really high frequency. The Rx Tx modules are likely "just" mixers and amplifiers.
-
#6 Reply
Posted by
voltlog
on 28 Sep, 2015 10:45
-
I have added a video annotation, with this observation.
Thanks for your insights.
-
#7 Reply
Posted by
T3sl4co1l
on 28 Sep, 2015 14:15
-
Hmm, looks like the LO board is ~8GHz, which should be roughly consistent with the size of the filter pattern on the PCB. The converter to ~80GHz must be on the little captive flex board on the housing.
Some crazy bits of hardware in there. Can't say I'd like to pay the MSRP on that thing!
Tim
-
#8 Reply
Posted by
orion242
on 29 Sep, 2015 02:05
-
Great vid, thanks for posting.
Can't think what the bit of kit costs, nothing I would be ever tearing down. Subscribed and look forward to your posts.
-
#9 Reply
Posted by
voltlog
on 29 Sep, 2015 06:48
-
I don't own these, I am just the repair guy, that is if I manage to find the fault
I believe the MSRP on a pair of these was around 10K USD so totally worth repairing
-
#10 Reply
Posted by
TiN
on 04 Oct, 2015 05:54
-
Nice RF magic there, thanks for video.
-
#11 Reply
Posted by
SNGLinks
on 05 Oct, 2015 09:59
-
Good video. Only thing missing is a view of the dish feed.
-
#12 Reply
Posted by
voltlog
on 05 Oct, 2015 13:59
-
Yes, I kinda skipped that one to avoid making the video too long. I didn't expect to find anything interesting in there.
Thanks for the encouragement.
-
#13 Reply
Posted by
GNU_Ninja
on 05 Oct, 2015 14:31
-
Bare feet. Check.
Black plastic pointer. Check.
Windowless lab. Check.
Looks like Dave has a competitor
Nice teardown
I've subscribed to your YouTube channel
-
#14 Reply
Posted by
voltlog
on 05 Oct, 2015 14:43
-
Nice teardown I've subscribed to your YouTube channel
Thanks! I am constantly trying to improve my english and the video quality.
-
#15 Reply
Posted by
AF6LJ
on 05 Oct, 2015 23:37
-
Nice RF Goodness, Thanks.
-
-
That was very interesting. Good video work and your English is fine.
I think the part you have labelled as a combiner may actually be a duplexer.
-
#17 Reply
Posted by
1design
on 23 Dec, 2015 20:02
-
From the size of the wave guides and topology of the mmW section it might be a sub-harmonic mixer design with the x4 multiplier in the end section of the synthesizer. Old designs used to have a 5GHz IF so they could cover both 71 to 76 and 81 to 86 GHz with the same design just changing the side band they use. This means that the LO would be in the range of 38 and 40.5 GHz. If we divide this by 4 the Synthesizer frequency should be between 8.xx and 10.xx GHz which seems to fit the amplifier range. If they want to fine tune the TX to RX they can do it in the IF section. Most newer designs use a heterodyne or direct conversion architecture. I might be work, just an assumption form what I could catch from the video. The only part that baffles me is the mmW coupler at the RX side. it splits the signal into two branches and thus increasing the noise figure. Any clue why they would decide for such a solution?
BR.
-
#18 Reply
Posted by
AF6LJ
on 23 Dec, 2015 20:21
-
From the size of the wave guides and topology of the mmW section it might be a sub-harmonic mixer design with the x4 multiplier in the end section of the synthesizer. Old designs used to have a 5GHz IF so they could cover both 71 to 76 and 81 to 86 GHz with the same design just changing the side band they use. This means that the LO would be in the range of 38 and 40.5 GHz. If we divide this by 4 the Synthesizer frequency should be between 8.xx and 10.xx GHz which seems to fit the amplifier range. If they want to fine tune the TX to RX they can do it in the IF section. Most newer designs use a heterodyne or direct conversion architecture. I might be work, just an assumption form what I could catch from the video. The only part that baffles me is the mmW coupler at the RX side. it splits the signal into two branches and thus increasing the noise figure. Any clue why they would decide for such a solution?
BR.
The only thing I could think of falls in ether of two possibilities.
1. two mixers quadrature LO fed to both producing an output that is only the sum or difference frequency, not both.
2. That splitter looked more to me like a rat race. (circulator) which would provide protection for the receiver.
The view of that section isn't all that good unfortunately.
-
#19 Reply
Posted by
1design
on 23 Dec, 2015 21:34
-
I don't see nay mixer at the input, only 2 LNAs. It would make sense to be a hybrid coupler as they have a hybrid also at the exit of the LNAs. This would suggest the IP3 was too low and they needed to back off the power? It still doesn't make sense why to use this method at the receiver, again you loose on the noise figure. It might be possible that this is a universal board and each of the units cover one band as they are older parts?
After taking a better look I noticed all three ICs are the same type of LNA/VGA. This means they are all wide-band and there must be a reason why they decided to use that configuration at the input and it isn't the IP3 as this would affect the second stage even more.
I checked the RX IF, it seems to be around 10GHz, so my frequency plan goes out the window. It seems like the TX is directly upconverted and the RX has a heterodyne architecture with the IF being at 10 GHz and all the VGAs are there. The mmW parts are all fixed gain LNA ICs. This also makes sense.
-
#20 Reply
Posted by
voltlog
on 26 Dec, 2015 19:07
-
It's nice to read your theories on the architecture that might be used in the microwave section. Myself I know nothing about this stuff and in the end I wasn't able to fix the broken radio. I couldn't find anything wrong in the power section, the cpu/digital section seems to work fine as the OS fully boots and responds to commands but it was showing an error about the "Costas loop".
-
#21 Reply
Posted by
daqq
on 26 Dec, 2015 21:48
-
Please post the whole message/screenshot, might be interesting.
-
#22 Reply
Posted by
voltlog
on 27 Dec, 2015 11:09
-
I don't have the radio anymore to get a screenshot from the GUI but it was just an alarm flag "Costas Loop Lock Alarm Flag". Also the RSSI value, shown in the GUI was all over the place for this radio(oscillating), while for the second one it was staying at a constant level.
-
#23 Reply
Posted by
1design
on 27 Dec, 2015 20:52
-
This could mean that something was wrong with one of the ICs in the RX path. Have you checked the spectrum on the receive chain, or the bias/VGA voltages of the chain?
BR
-
#24 Reply
Posted by
voltlog
on 27 Dec, 2015 21:43
-
I checked every voltage rail that I could find and they were all ok. No signal measurements though since I don't have a spectrum analyzer. That's why I kind off gave up on the repair, I was lacking an important tool and this was a rather complicated radio to start with..