As promised, Marc N2UO has done some work with a new GaN device from CREE. In new files please find:
http://www.amsat.org/amsat-new/eagle/EaglePedia/uploads/1/13/VHF_SiC_amplifi...
This looks REALLY good for HELAPS.
Bob
Apologies. I know CREE is a leading company in GaN and I leaped without being careful. This is SiC technology.
Bob
Robert McGwier wrote:
As promised, Marc N2UO has done some work with a new GaN device from CREE. In new files please find:
http://www.amsat.org/amsat-new/eagle/EaglePedia/uploads/1/13/VHF_SiC_amplifi...
This looks REALLY good for HELAPS.
Bob
Robert McGwier wrote:
Apologies. I know CREE is a leading company in GaN and I leaped without being careful. This is SiC technology.
Bob
We shall forever remember that Bob's GAN was [sic]'d (see http://en.wikipedia.org/wiki/Sic) :-D .
Seriously -- this looks great. I hope Marc can get a hardware version running soon. I am very impressed with the nearly parabolic Po vs Vd curve (fig 5). This sez to me that the "plate" source impedance is quite constant vs Vd, which in turn makes the HELAPS "plate modulation" application even easier.
From Fig 3, I gather that the predicted 3rd harmonic is more than 70 dB
below the fundamental. This is good because the 3rd harmonic of 145.x downlink is pretty close to the 435.y MHz uplink, and the fewer filters that are needed, the better! I wonder if perhaps some of the C5,5,7,8 capacitors could be open circuit ¼-wave @ 70 cm coax stubs to improve the isolation even more?
I'd like to suggest that the EAGLE Class-E "phase" PA be 3 "parallel" separate stages differing in phase by 120º, each feeding a separate dipole (which are physically oriented at 120º on the +Z antenna farm), running at Vd in the 15-25V range. This would afford 3x redundancy (remember how the AO-40 2M amplifier death was a single-point failure), and also puts the total PA DC efficiency > 90% (see Table I), and also saves the power losses in a 3-way high power divider. My gut feeling is that, for the relatively narrow signal bandwidth, all the amplifiers can share the same "plate" modulator.
Nice work, Marc!
Tom has an intesting suggestion of the 3x amplifiers/antennas. I have been suggesting this antenna configuration as I just take the P3C hardware and use it! Please help my poor brain understand how this gets the desired RHCP? Is is just the three rotationally oriented antennas?
Dick Jansson --------------------------- [email protected] ---------------------------
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Tom Clark, K3IO Sent: Friday, 18 August, 2006 2253 To: Robert McGwier Cc: 'EAGLE' Subject: [eagle] Re: 2 Meter amplifier
Robert McGwier wrote:
Apologies. I know CREE is a leading company in GaN and I leaped without being careful. This is SiC technology.
Bob
We shall forever remember that Bob's GAN was [sic]'d (see http://en.wikipedia.org/wiki/Sic) :-D .
Seriously -- this looks great. I hope Marc can get a hardware version running soon. I am very impressed with the nearly parabolic Po vs Vd curve (fig 5). This sez to me that the "plate" source impedance is quite constant vs Vd, which in turn makes the HELAPS "plate modulation" application even easier.
From Fig 3, I gather that the predicted 3rd harmonic is more than 70 dB
below the fundamental. This is good because the 3rd harmonic of 145.x downlink is pretty close to the 435.y MHz uplink, and the fewer filters that are needed, the better! I wonder if perhaps some of the C5,5,7,8 capacitors could be open circuit ¼-wave @ 70 cm coax stubs to improve the isolation even more?
I'd like to suggest that the EAGLE Class-E "phase" PA be 3 "parallel" separate stages differing in phase by 120º, each feeding a separate dipole (which are physically oriented at 120º on the +Z antenna farm), running at Vd in the 15-25V range. This would afford 3x redundancy (remember how the AO-40 2M amplifier death was a single-point failure), and also puts the total PA DC efficiency > 90% (see Table I), and also saves the power losses in a 3-way high power divider. My gut feeling is that, for the relatively narrow signal bandwidth, all the amplifiers can share the same "plate" modulator.
Nice work, Marc!
_______________________________________________ Via the Eagle mailing list courtesy of AMSAT-NA [email protected] http://amsat.org/mailman/listinfo/eagle
I am hoping that CREE makes a 50w output device and we build one amplifier to make the thing so we can run at 50w PEP and smaller ones for the N amplifier/antenna version. I am wondering about the SDX/HELAPS on the multiple amplifier version. It will be easier for Dick to cool but harder for our HELAPS so that is the note of caution I sound. I like everything about it except contemplating that. Our HELAPS will be sensor based and heavily dependent on predistortion to achieve the best performance. We might be pushing the SDX/HPSDR Sasquatch to the limit controlling three amplifiers. If we really are going to try to get total DC efficiency well over 50% for the transponder and get good IMD, we will need the polyphase DSP HELAPS to work.
These E/I curves are looking really good if the simulation approaches reality. I would like comments ASAP for Marc.
We are getting well down the road on the HPSDR developments. When Lyle gets back from Oz, that project should take a leap forward. A bunch of our SDX hardware will be from that project (in prototype form).
Bob
Dick Jansson-rr wrote:
Tom has an intesting suggestion of the 3x amplifiers/antennas. I have been suggesting this antenna configuration as I just take the P3C hardware and use it! Please help my poor brain understand how this gets the desired RHCP? Is is just the three rotationally oriented antennas?
Dick Jansson
[email protected]
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Tom Clark, K3IO Sent: Friday, 18 August, 2006 2253 To: Robert McGwier Cc: 'EAGLE' Subject: [eagle] Re: 2 Meter amplifier
Robert McGwier wrote:
Apologies. I know CREE is a leading company in GaN and I leaped without being careful. This is SiC technology.
Bob
Thanks all for the comments. I am waiting for samples from Cree. Once I get them, I will build a unit. I just got the Modelithics models for all the capacitors and inductors today. I will run another simulation with those models that seem to be better than the ones I used.
In any case, SiC devices seem to perform just fine down on 2 meters. GaN technology from Cree so far is only one internally matched device good for 3 to 3.5 GHz, and it is only 15 W. Eudyna has the devices we need, but they will not give away samples under any circumstance. Nitronex devices are also narrowband. And, more important, Cree is the only company so far that has a bunch of talented engineers who developed some reliable models. If GaN devices become available, we can always try them out in a similar circuit.
Regarding this statement:
From Fig 3, I gather that the predicted 3rd
harmonic is more than 70 dB below the fundamental.
The harmonics have been effectively attenuated by the 5 element Chebyshev filter and the matching network. The output of the amplifier without the filter has the harmonics at very high levels since it is a switching amplifier.
Usually, if you can simulate somthing good, you have a good chance of getting it to work close enough. The opposite could be true, too, but it is much less common.
I'll be out in the field this weekend for the 10 GHz contest, and then off to the EME conference in Germany. Will do more work on this after that. I am sure there will be more question and suggestions for improvement in the meantime.
73, Marc N2UO
__________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
Dick Jansson-rr wrote:
Tom has an intesting suggestion of the 3x amplifiers/antennas. I have been suggesting this antenna configuration as I just take the P3C hardware and use it! Please help my poor brain understand how this gets the desired RHCP? Is is just the three rotationally oriented antennas?
Dick Jansson
[email protected]
Dick -- if you have 3 antennas at 0º,120º and 240º, you can make them be circularly polarized by feeding them the same RF signal, but at relate phases of 0º,120º and 240º (or 0º,-120º and -240º which is equivalent to 0º, 240º and 120º, i.e. swapping any two of the three, to get the opposite sense of circular). This is completely analogous to taking two elements 90º apart and feeding them in phase quadrature.
Of course, since the elements do not have their centers at the same physical point, the polarization purity degrades off-axis.
FYI -- P3A, P3B=AO-10 and P3C=AO-13 all had 2M antennas sticking out the three tips of the star and were fed in just this way. Of course, AO-10's polarization was screwed up when the launcher came up and "bumped" the satellite just after separation. [I did the mathematical simulation of the "half ZL-special" for Karl as we were building P3A. That was back in the minicomputer fortrash-4 era. Because interactive graphics were unknown, the simulation was an inch-thick pile of computer paper.]
My suggestion was that we basically that, if we had 3 amplifiers we could do the 0º/120º/240º phase split at low level, avoiding the "cost" of the inevitable losses in the 3-way power splitter that would be necessary if we did the split at the amplifier output. And we would have simple built-in redundancy too. As Bob pointed out, this might be too complicated for the DSP HELAPS to deal with.
73, Tom
participants (4)
-
Dick Jansson-rr -
Marc Franco -
Robert McGwier -
Tom Clark, K3IO