This story began at a local hamfest. Who knew that bying a junk RF box was going to turn into a fantastic learning experience. Some time back browsing tables at a hamfest I came across a dealer who was selling RF items. I grabbed a heavy looking aluminum box with SMA connectors and power pins, with a "Solid State Amplifier" marking on the cover, which the dealer was selling as a "nice RF box for projects" or something. And it was a pretty nice milled heavy gage aluminum box with cover and RF connectors. Something that looked weird to me was what looked to be the power peg was marked "+90V". That was something unusual and I asked the dealer if he knows why so high voltage to power the circuit. The guy looked puzzled and said he never paid attention and would not know.
Anyways, I bought the box and went home where I opened it and what was inside looked pretty interesting. It had a FR4 board which was supposedly a voltage regulator and another board with[/img] a microwave circuit on a black glossy substrate. The third board had something that appeared to be a simple transmission line on a white ceramic substrate.
The power supply board had a IC and a transistor and the way the board was mounted was interesting. There were two snap-on small heatsinks glued to the box, and the board was held in them with the IC and transistor bodies snaped into the heatsinks.
On the microwave board I could immediately identify three microwave circulators, a few filters and a directional coupler. The board was stamped with the Raytheon logo and date of "14 Jan 72". Not only it was fascinating to see something that was stamped with actual day when the board was manufactured, but also it was interesting to see that such microwave technology existed in early 70-s.
The signal path looked as follows: from the RF Input SMA connector on the top right the signal goes into the first circulator, exits on the left side of the circulator and into the second circulator, exiting on the left side of it and going into the third circulator, exiting from the top side, and through the direcional coupler exists through the RF Output SMA connector on the top left. The directional coupler circuit had a 50 Ohm terminating resistor on one side of the coupled line and the other side of the coupled line connected to the ceramic transmisison line and exited the box through the third SMA connector marked "Power Monitor". Here is a close up of the directional coupler structure. The dark grey neck between the bottom side of the coupled line and a pad was measured as a 50 Ohm resistor, and the pad is apparently a microwave decoupling capacitor.
What else could be seen inside was something that looked a milled cavity with a tuning screw with some sort of a piece of polymer material on it that possibly was a part of the tuning structure. There was a milled slot under the tuning screw and a part of the polymer piece protruded into the slot, possibly interacting with whatever was inside the cavity.
On the side of the box was another big brass tuning screw, which supposedly was tuning something inside the cavity itself. It was moving very smoothly and almost with no friction when turned. The size of the brass screw and a smaller treaded hole in the center made me think it was designed to allow some sort of remote tuning attachment to be connected to it, i.e. a motor or some sort of extension hardware.
Looking back inside the box and the cavity end, it could be seen what looked as a microwave low pass filter structure connected to the power supply terminal, and one of the shoulders of the second circulator connected to whatever was inside the cavity via a ceramic capacitor. It was not a conventional SMD capacitor though but rather a one made using a piece of ceramic dielectric with plates.
So certainly some magic was hapenning inside the cavity, it looked to have some sort of a tuning structure that was essential part of this amplifier. And by the way, so far I could not locate any transistors or ICs on the RF board, and it was an amplifier! Fascinating... how the hell this thing worked? Hmmm... This is when I started vaguely remembering I read something about avalanche diode amplifiers. Could it be the thing? The +90V supply voltage also could be a hint. I had nothing to lose so I unsrewed and removed the big (br)ass screw that revealed an interesting structure:
It appeared that the act of adjusting the screw was changing the length of the structure consisting of a brass rod (measured 36mm length) and the pin on the other side of the cavity and going into the slotted side (12.5mm length) of the rod. And we saw previously that a signal from the second microwave circulator was connected to the pin on the interior side of the cavity, and thus, to the slotted end of the brass rod. Seemed to be a tuned circuit to me. But still was unclear how amplification action was achieved. Taking a closer look at the joint between the brass rod and the screw and measuring continuity eventually brough clarity to the matter. There was a diode at the joint:
The diode terminals were soldered to the screw and to the rod. So part of the RF signal mixed with DC bias of fairly high voltage was entering the cavity through a coupling pin, transferred to the slotted part of the brass rod and through it to one terminal of the diode, with the other diode terminal grounded through the big (br)ass screw on the other end of the cavity. Phew...! Found a non-linear component that may be the one producing amplification in this amplifier.
Some googling produced results confirming that what I had could be a IMPATT avalanche diode microwave amplifier. The US patent 4050032 came pretty close to it, with a difference in the method of connecting the diode.
The action of amplification was provided by the IMPATT diode. They operate in reverse breakdown mode and have negative resistance which implies in operating mode they have no loss but amplification. Having a same port for input and output is possible by using a microwave circulator which is a device that forwards RF energy in one direction between its ports in a circular manner. The signal path looked the following:
RF energy from the first circulator on the left (not shown) enters the second circulator which forwards it to the port connected to the tuning cavity with the IMPATT diode. Amplified energy exits through the same port but in clockwise direction and not back to input port due to the black magic property of the circulator. The amplified signal then passes the third circulator which isolates the previous stage from mismatched load. Any reflected energy from a mismatched load would go back through the third circulator but via a different path and be absorbed by the 50 Ohm termination load as shown by the skinny green line. Brilliant - no matching structure is needed beacuse of using a output circulator that plays a role of a directional RF isolator.
Part of the RF Out signal is diverted to the Power Monitor output connector by using a directional coupler structure as shown by the skinny red line.
Now, it is not exactly clear to me how the first circulator works, specifically what is the circuit on its left port as shown in the next picture. My guess is the RF input signal goes out of the right port towards the second circulator, and possibly when some condition exists that too much reflected energy gets back to that port , i.e. because of mismatch in the tuning cavity or when the signal is outside of the working range of the circulators, the reflected energy would travel to the left port and into the circuitry that terminates with a SMA connector marked Alarm.
The circuit seems to be some sort of a low pass filter with a component that may be a RF rectifier, a close up picture of which is provided next. If that is the case, the Alarm signal may be a DC voltage representing mismatch, overload or out of band confition of the amplifier. If someone recognizes how the Alarm circuit works please let me know.
I took a stab at reverse engineering the schematic, this is what it may look like.
To estimate the amplifier's working frequency range, I tried to measure approximate sizes of the transmission lines width, the circulators ports length, and the substrate thickness and plug the numbers into well known formulas to come up with the working frequency this design was for. The substrate was of a shiny deep black color, I have not worked with exotic microwave substrates, so some googling I made a guess this may be GaAs material with er=12.8...13. The substrate board thickness was difficult to measure but it looked approximately 30mil thick. So the input data for calculation of the design frequency was:
- Substrate material: GaAs, er=12.8...13
- Substrate thickness: 30 mil
- Transmission line width: ~20 mil
- Circulator port length: 2.5mm (assuming it is Lambda/4 electric length)
With that I came up with a possible working frequency range of 10...10.5GHz, which fits in a frequency range allocated for radiolocation, and that makes perfect sense considering the technology of the amplifier and the manufacturer (Raytheon). This same frequency range seems to be available for Amateur radio use, so it seems I may hold on to the amplifier and may be even use it when I migrate to 10GHz or perhaps to play with the low cost radar modules. I have a 20GHz spectrum analyzer but do not have a signal source to verify the amplifier frequency range. Maybe some indirect indications like elevated noise level within a certain frequency range observed with the spectrum analyzer can give clues, or maybe I will try to use the 10GHz low cost radar module as a signal source, will see.
So here we go, it was time well spent and I learned something in the process. Hope you enjoyed the post. Please give your comments if you can say more about the amp or point me to any errors that I may have made. At least one thing remain unclear to me: the purpose of the inside tuning screw with a blob of a polymer protruding into the cavity slot and what it does. Oh yes, and one more thing that puzzled me was this:
Am I a happy owner of the first unit in the batch or perhaps the "Only" unit ever made by Ratheon on Jan 14 of 1972 ??
Who knows. Anyway, it was fun.
Home
Help
Search
Login
Register
