I have the power!
Lets talk about power amplifiers solutions for AO-100
The successful launch and commissioning of the AO-100 has been a real rule-change in the story of amateur satellites. Although foreseeing since the very beginning of Oscars story and classified as “phase 4”, we had to wait for more than five decades to have the very first geo stationary amateur satellite in place and fully available.
But why it’s referred as “phase 4”? Let step back to the origin, when the continuous technological development led the history of amateur radio satellites to be divided into "phases", each of which is characterized by clear objectives and technical or structural characteristics.
The following table includes the distinctive characteristics of each “phase”.
Phase | Features |
1 | Satellites powered only by batteries with short life dedicated to technology experiments |
2 | Satellites of long duration and capable of translating communications (transponder). LEO orbit powered with solar panels |
3 | Satellites of long duration with complex communication systems, control and telemetry. Mostly HEO and Molniya orbits |
4 | Satellites in geostationary orbit. The QO-100 is the very first actually been built and launched |
5 | Satellites or even space probes capable of interplanetary or lunar missions. Designed and partially built but never launched |
Since its commissioning, the QO-100 demonstrates of being even more sensitive than expected. I personally succeed to run some QSOs with only 400mW and a 10dB horn antenna and I believe this is probably one the most QRP way to use the bird. For a more reliable and comfortable traffic some more ERP is needed. As rule of thumb, it’s preferable to start investing on a good and big antennas, but not always space, cost and other constraint let us reach to the ERP wished in that way. Thus the only chance comes from an higher transmit power.
The market is now plenty of offer of smalls, compact and (often) cheap amplifier modules, mostly intended for WiFi, Cellular, WiMax, ISM and Bluetooth applications. But how do they perform? Which pro and cons? Let see my experience with some of the most popular ones.
An important note prior to moving on: I do not have a certified ISO 17025 radio frequency lab at home, but a “regular” ham one. So, all the figures you will find on the following pages have been measured with my best effort to minimize errors but of course, are not “official laboratory grade” ones.
SBB5089Z + SZA2044 
Figura 1: SBB5089 + SZA2044 amplifier
Figura 1: SBB5089 + SZA2044 amplifier
This is a very compact module, advertised for a 1W max output. The supply may range from 6 up to 30Vdc and this makes it very convenient also for portable operation. The idle current of the whole device is about 210mA @ 12Vdc; please note that the current drowned from the power supply decrease with the increase of the voltage due to the internal switching power supply that try to keep the input power constant across all the input voltage range.
The circuit use a couple of well known devices. The one at the input is theSBB5089Z, an high linearity gain block for applications such as PA Driver Amplifier, Cellular, PCS, GSM, UMTS, wideband instrumentation, wireless data, satellite terminals. It is internally matched to 50 Ω at input and output over a very broad range.
Following in the chain, we found the SZA-2044, a modern product specifically designed as a final stage for 802.11b/g and 802.16 equipment in the 2.0GHz to 2.7GHz bands.
The module shows an overall very high gain, in the order of 39-40dB, with a bandwidth a -3dB that ranges from about 400MHz up to 2.7GHz. This is something to concern for our application: in order to prevent a lot of spurious emissions over the air, it’s mandatory to feed it with a very clean signal and perhaps also to filter its output upon the cases.
Here its response curve at my bench:
“Out of the box” my specimen was not able to deliver more than 550mW @ 2400MHz with saturated input (-5dBm). At lower frequencies, let say 2380MHz and downward, I could see up to 29-30dBm with no problem and with -10dBm at the input. If you wanna squeeze up some half dBm you can just remove the component marked R7 and C17.
Surfing the web, you could find also a version with a pair of SZA2044, which I guess able to deliver some 1-1,5W.
Final advise: when operated at “high power”, the case of the final device may become pretty warm, so watch your fingers!
SZM2166
Figura 3: SZM2166 amplifier
Figura 3: SZM2166 amplifier
If the 1W module does not satisfy your needs, another very popular choice is the one based upon the SZM-2166 device, claimed for 2W. For my testing I bought two of those from different web supplier but I’ve always been unlucky. The first one arrived with several pin short circuited due to poor soldering process. The second one passed away after few seconds on my test bench due to some terrific self-oscillation that fed back into the onboard switching power supply. Definitely there are quality problems in the production of those cheap modules coming from far East, so be prepared in case.
Last but not least, pay attention to the fact that the input is DC coupled, therefore to insert a DC block could be a wise choice.
MW7IC2725N
If you are not yet happy with 1 o 2 W of power you probably have to move to a different device. Luckily, today the catalogs of the main silicon producers are plenty of solutions for our purpose. In my humble opinion, one the most intriguing one is the MW7IC2725N made by NXP and Freescale, an 4W (digital modulation) wideband integrated circuit, designed with an on-chip matching network that makes it usable from 2300 up to 2700 MHz, thus just perfect for us. The pros that make it a good choice for our purpose are also its high gain (it’s a two stages with 26-28dB total gain), the pretty low cost, the ability to deal very well with digital signals and the easy of use. The only real cons that may concern us is the needs of 28V for its supply. In most of the cases this is not an issue, but it could be in some portable operation. The chip by itself is easily available from several sources as well you may find some “ready to go” boards based upon that chip. For my convenience I bought and tested the one featured and produced by SK9MTS.
The board is very compact and delivering a maximum of about 25-30W with an associated gain of about 26dB. Another good point: at full throttle (30W out) you need only 2,8A @ 28V from your power source, an input power of 78W which in turn means about 38% of overall power efficiency: pretty nice isn’t?
If you plan to use it at high output, you have to ensure a proper cooling, able to dissipate about 50W. Studying a little the data-sheet I believe that a convenient choice could be the use of an heat sink of no more than 1°C/W of thermal resistance in order to keep the junctions within 150°C.
The surplus solutions
If the 5-20W above solution are based upon devices of current production, we should not forget that there is margin and opportunities for some more QRO amplifier based upon ex-BTS (base terminal station) boards. So, why not to have a look to some surplus boards and try to bend them to our purpose? Herewith my experience with a couple of samples seek out at some local flea market for a bunch of bucks each.
Surplus board n°1
Figura 5: Surplus board with MRF21010, MRF21045 and a pair of SRF7068
Figura 5: Surplus board with MRF21010, MRF21045 and a pair of SRF7068
This board consist of a series of three devices: MRF21010, MRF21045 and a pair of SFR7068, suitable for FM, TDMA, CDMA and multicarrier amplifier applications. To be used in Class AB for PCN - PCS/cellular radio and WLL applications. At its design frequency (2.1-2.3GHz) this group is able to deliver more than 150W with a consistent gain in the range of 30dB. At 2.4GHz the devices (the actives ones and hybrid couplers too) are not specified and the matching should
probably be totally reviewed. Anyhow, with minor changes and some tuning you can squeeze out some 50W @ 28V and about 30W with 24Vdc.
Total gain is in the range of 20-23dB. Another possibility of experimenting is to strip out the devices from the board and redo all the matching on a new one: I may guess 70-100W could be achievable for AO-100 but this is far beyond my scope and skills.
Surplus board n°2
Figura 8: Surplus board n°2 overview
Figura 8: Surplus board n°2 overview
The second surplus board I’ve hacked is rather complicated because it’s composed by a number of sub modules. On the right end (see pic) there is a low level board, with filters, mixers, etc. The most interesting part is the one that comprises the medium level amplifier based upon the MRF2130 and MRF2190 devices placed about on the center-right of the whole assembly. After that, there is on the left, another stage of very high power devices, but usually they are both blow down (at
least I’ve found as such in all the ones I’ve checked). If you can catch a healthy pair, that could be your lucky day!
Both devices are intended for W-CDMA base station applications with frequencies from 2110 to 2170 MHz and are suitable for FM, TDMA, CDMA and multicarrier amplifier applications. From my tests, with some minors tuning you may get around 30W max @28V and 2400MHz. As above, our use is far out specification of the devices.
Probably those two boards could be of more interest for EMErs at 13cm (2304-2320MHz) especially the second one if it has the final pair working.
Nothing else?
The solutions above presented are just a part of the one available: for sure there are many more not investigated in this paper. If you are a fan of “old fashion” amplifiers, you may also figure out to reuse a former 3CX100/2C39 power amplifier for 13cm band. With a single tube, you may get 20 to 50W effortless. If you are for modern devices instead, you may consider giving a look or even a try to the following ones: NE5520279A (1,5W, 3,3 or 5V), MMZ25333B (2W, 35dB, 5V), MGFS45Hxxxx (hybrid modules, dozen watt), MHT1008NT1 (50W, 28V), surplus board based upon BLF8G22LS-160BV or BLC8G22LS-450AV (50-100W expected) and some more I surely forgot to mention.
Optional but useful ancillary devices
Limiter
Figura 10: Example of limiter for instrument use, with N connectors
Figura 10: Example of limiter for instrument use, with N connectors
The use of this component on our up link chain is to prevent or protect from overloading the input stage (of a power amplifier, of a mixer and so on) from unwanted high-power signals.
The Limiter's function of protecting from high level signals while maintaining very low loss requires the use of a non-linear device. A shunt diode is an ideal because it’s virtually transparent when biased "off" and highly reflective when biased "on". PIN diodes are commonly used because they can handle also high power signals and provide reasonable leakage. In the presence of small signals, the shunt diode does not conduct and therefore represents high impedance, or a low loss condition on the through path. In the presence of high power signals, the RF level lead the diode to conductive state, driving the impedance to drop well below 50Ω and the entire circuit becomes reflective, preventing the bulk of the energy from getting to following component.
Please note that while acting, clipping and limiting the input signal, its output contains many harmonics of the applied input signal, thus a proper filtering could be advisable of even required.
Limiters are specified by a number of key parameters:
Operating Frequency Range: defines the range over which the limiter must protect the succeeding components. Lower limit is usually set by the recovery time of the diode, while the upper limit by the parasitic parameters (inductance and capacity) of the components.
Insertion loss: defines the small signal throughput loss (S21) of the limiter. Insertion loss is defined over a dynamic range up to the input limiting range.
Leakage: when high input power is applied to a limiter, only a small portion of the power can pass through the limiter. The small portion is called leakage and is typically specified as an absolute level (dBm or Watt).
Input Power: defines the limiting range. It is specified at both Min and Max levels. Over this range, the Limiter will follow a specified Δoutput /Δinput characterization. For input signals above the maximum input power, the output will begin to increase again at a greater slope.
Please be careful to not confuse this parameter with the Max Rating which is higher and defines the maximum power that the unit can handle without being damaged.Limiting: is the ratio of difference of output power to difference of input power over the limiting input power ranger. (Δoutput /Δinput)
Recovery Time: it’s the time it takes for the limiter to recover from a high-power signal clipping. This is a pulsed condition and is defined as the time between the 50% point of the trailing edge of the high-power pulse to the time where the output reaches 90% of the final small signal level.
Circulator
Figura 11: Example of a three port circulator with N connectors
Figura 11: Example of a three port circulator with N connectors
This component is one of the less known and appreciated among hams but its useful is out of doubt. A circulator could be seen as a sort of RF “roundabout” where you can leave the road only at the very next exit. In more technical words, it’s a passive, non-reciprocal three or four-port device that only allows a radio-frequency signal to exit through the port directly after the one it entered. For instance, on a three-port circulator, a signal applied to port 1 only comes out of port 2; a
signal applied to port 2 only comes out of port 3; a signal applied to port 3 only comes out of port 1, and so on. Said that, how it can help us in our application? Well, when a three-port circulator is terminated in a matched load, it can be used as an insulator. This configuration is used to shield an equipment on its input side from the effects of conditions on its output side. A very appreciable use for us is preventing the final stage being damaged in case of highly mismatched load. For this application you will have to connect the output power stage to the port 1, the antenna to port 2 and a suitable dummy load to port 3. Under normal condition, the power will smoothly flow from 1 to 2 (antenna), with (almost) nothing popping our from port 3. In case of severe mismatch on the output 2 (i.e. connector or coax fail, damaged antenna, icing of the dish feeder, and so on), the reflected power, instead to hit back the output stage will be directed to dissipate its energy on the dummy connected to the port 3. And all that with no delay! While choosing this kind of device, three are the most important parameters to look after:
- frequency range: it has of course to include our working frequency
- maximum power handling: is shall be high enough
- insulation and loss: for our purposes the most important one is having a low loss device to not waste output signal
DC-block
As its name may suggest this device is intended to stop any direct current flowing on the line. Basically it’s composed by a series capacitor canned into a suitable case with a proper connectors pair. Usually radio frequency amplifier are not DC coupled, but sometime they are so! Thus to prevent and DC to leaks out or get into another module an inter-stage DC block may be the solution.
Conclusions
The matter is very interesting and I’m sure many of you may seek out and propose at least another solution and device. However, I hope these pages could be of help to most of the passionate user of QO-100 and stimulate radio hamming via satellite!
Bibliography
www.richardsonrfpd.com/docs/rfpd/PIN_Limiter_Design_Guide.pdf
en.wikipedia.org/wiki/Circulator