I received the AN758 AMP from CCI. I want this to be the genesis of our amp design.
The board and parts arrived and I got it together. The documentation was scattered and incomplete but I managed to piece it together by referring to several documents. Since I have built the AN762 AMP, I have some fore knowledge of what to expect in building this type amp. There are schematic errors, left out part identifiers for the circuit board, a wrong part specified in the app note. I am forgiving however because the important RF design is the important thing. I can see where new builders using this app note would have some pit falls.
The picture above is the AMP and the heat sink/fan I selected for the job. The heat sink is 3 inch square and attaches to the back of the circuit board. I plan on the fan to be attached to the rear panel having a 3 inch square cutout. A finger guard will be on the outside of the rear panel. So you might envision the rear view of the amp. You would see finger guard and square cutout for the fan. A passive heat sink is out of the question for a 100% duty-cycle amplifier. According to the engineer at Alpha Novatech, we will have a 0.33 C/W thermal quality with the components I selected. The temperature will rise 0.33 degrees C for each watt dissipated at the heat sink Assuming 100W RF and 100W heat (50% efficiency), the heat sink temperature will rise 33 degrees C. Add the ambient temperature of 30 C, we have a heat sink temperature of 63 degrees C. Our air flow design will pull in air from near the front (filtered slots). It will be drawn across internal components and the heat sink and exit out the rear. In the design of the air flow system, we have to be able allow the inflow of air to not create a back pressure. The specified fan has a max air flow of 1.0 m³/min with a rated noise level of 35 dB.
I measured to confirm that the bias supply is correct and discovered that a zener supplied with the kit (45V) exceeds the VIN max on the bias regulator chip. The spec calls for 40V maximum so I had to cease tests to not cause failure of the bias regulator chip. I ordered a 36V zener replacement and a 1W 1K to replace the original 1/2W part. I also ordered a multi-turn adjustment pot for the bias setting. I did not like the 1–turn pot supplied in the AN758 AMP parts bag. That would be too coarse of an adjustment.
The picture above is a test board that delivers 48V @4A from 12VDC. It is 3×2 inch. My manual test toggle switch operates it. The efficiency is about 95%. Pretty amazing stuff. This design is a dual phase which provides the load current with 180 degree phases. It is a pretty old design. I have had this board on hand for several years. I am planning a four phase design with 90 degrees per phase. That design will deliver 48V @ 8A! Now we are talking! The multi phases reduces the size of filter components It is 96–97% efficient. See the link: http://cds.linear.com/docs/Design%20Note/dn453f.pdf for the 4 phase 384W boost converter design. I like the features this design provides which are very ham friendly. The DC input voltage level will no longer be a factor as it is in 12V powered amp designs. Other 100W Amps powered by 12V require about 13.8V to get full power output. You can only get that with a charger in an automobile. Now you can get that from a camp site using a Lithium pack! Since the efficiency is about 97% the extra DC input current will be hardly noticeable. The design only supplies the RF transistors 48V during transmit so standby current will be nil.
This picture above shows the amp connected to the LPF with a manual control box to set the LPF band (160–10M). In the final design, this manual control box does not exist. In the low cost version, a band switch will control it. The advanced model will automatically select the proper LPF according to the transceiver setting.
In my tests, the transistors will not be installed until I have the bias supply working correctly. That is really important in bringing up a new amp. The bias has to be right or you pop the transistors right away. I should get the few parts I need to complete the proto amp this week. My immediate goal is to get the bias working to my satisfaction, mount the RF transistors and start testing. I built this proto amp with the optional resistive pi network parts installed. I installed parts for -9dB.
After I am operational, I want to characterize the gain, current and band response from 160–10M. I find it useful to record the data into an Excel spreadsheet and plot it. It gives you a big picture. I have a keen interest to analyze the linearity to find the linear part of the operating curve. Low IMD is very high on my list of specifications to achieve. At some point soon I will have to get appropriate test equipment for IMD measurements.
In a word about packaging, I am planning on all flat panel designs for minimal costs. I have still not come up with a low-cost frame work, working on it in my minds eye right now.
Thank you for your support! I am sorry this could not be done faster. There are so many financial pressures to consider.
73, K5OOR – Virgil HF Projects