|
|
|
[<< Home](/home#4-signal-transport-panel-charges-3-and-4)
|
|
|
|
|
|
|
|
[<< Section 4.1](./4.1)
|
|
|
|
## 4.2 Transmitter Design
|
|
|
|
The RF over Fiber Transmitter was first designed as a single channel module for testing, followed by a four channel card for the full production module that incorporates additional circuitry to provide cryo-LNA bias and monitoring, and calibration tone injection.
|
|
|
|
|
|
|
|
### 4.2.1 Single-Channel Transmitter Module Design
|
|
|
|
The transmitter module is designed to provide adequate gain and noise figure to successfully modulate a 5mW, 1320 nm laser and keep system noise temperature contribution to a minimum. The transmitter uses three low noise amplifiers, all from Mini-Circuits, instrument bandpass defining filtering in the 1.30-1.72 GHz band provided by a custom KR Electronics filter design, calibration signal injection coupling provided by a Mini-Circuits directional coupler, and a 6-bit variable attenuator to set gain and noise floor levels relative to adjacent channels.
|
|
|
|
|
|
|
|
Care is taken to ensure all amplifier stages are appropriately DC blocked and have enough RF isolation to prevent oscillations. Since this board has ~50 dB of gain, two chip attenuators, as well as the variable attenuator, can be used as pads between amplifier stages to improve stability if needed.
|
|
|
|
|
|
|
|
<div align="center">
|
|
|
|
<img src="../uploads/c4c08bbda8e95b28e485bfea087ac079/oshPark_4layer_stackup.PNG" width="500"/>
|
|
|
|
|
|
|
|
Figure 1: OshPark four layer stackup
|
|
|
|
</div>
|
|
|
|
|
|
|
|
OshPark four layer PCB material was chosen as the PCB material because of its fabrication with FR408-HR substrate and prepeg dielectric (stackup in Figure 1), which works reasonably well at L-band microwave frequencies. However, the laser matching network did not perform as expected on this PCB choice, likely due to the small 6.7 mil prepeg thickness separating the top copper layer and layer 2 (1st internal layer). This necessitated a redesign as discussed below.
|
|
|
|
|
|
|
|
<div align="center">
|
|
|
|
<img src="../uploads/c8edd5a59a3e86efeee5d5684f27b053/tx_r3-3_schematic.PNG" width="600"/>
|
|
|
|
|
|
|
|
|
|
|
|
Figure 2: TX r3-3 single channel module schematic
|
|
|
|
</div>
|
|
|
|
|
|
|
|
The laser is mounted in an end launch configuration within a cut-out in the PCB so the pigtail faces the same direction as the rest of the I/O connections on the FEB. This allows for less strain on the fragile fiber pigtail and reduces the length of the housing of the four channel card, since bend radius does not need to be accounted for on the far edge. Bolt-on Southwest SMA connectors are used for testing the single channel test and prototype module because they eliminate the need for any coplanar waveguide "neck downs" that affect trace impedance. On the final four channel boards these bolt-on SMA connectors are eliminated in favor of a single SMP connection for calibration signal injection and a custom ganged 4-channel SMP connector that interfaces with 4 LNA signals on a flexible stripline connector exiting the cryostat.
|
|
|
|
|
|
|
|
<div align="center">
|
|
|
|
<img src="../uploads/1cacf75c5ea187533a39aae57c984610/tx_r3-3_rendering.PNG" width="300"/>
|
|
|
|
|
|
|
|
Figure 3: TX r3-3 single channel development and test prototype module
|
|
|
|
</div>
|
|
|
|
|
|
|
|
### 4.2.2 Four-Channel Transmitter Card Design
|
|
|
|
The four channel production design expands on the single channel module used for testing by incorporating four channels on one PCB, along with the following additions:
|
|
|
|
- Bias tees for each channel to power the cryo-LNAs
|
|
|
|
- Sense lines from each bias tee to monitor cryo-LNA voltage and current
|
|
|
|
- A single calibration signal injection SMP connector with onboard 1:4 splitter
|
|
|
|
- A single 25 pin micro D-sub connector for +/- 5v power, ground, and cryo-LNA power and sense lines
|
|
|
|
- A ganged 4 way SMP connector that interfaces with the flexible stripline
|
|
|
|
- A fixture to hold the four SC/APC fiber connections to mate with a coupler on the I/O plate
|
|
|
|
|
|
|
|
<div align="center">
|
|
|
|
<img src="../uploads/d5258598cefab9cfd796c1d78836ad0a/4ch_top_level_schematic.PNG" width="600"/>
|
|
|
|
|
|
|
|
Figure 4: Four channel top level schematic
|
|
|
|
</div>
|
|
|
|
|
|
|
|
The first 4-channel TX card PCBs shown in Figure 5 have been delivered to BYU, are being fabricated, and will soon be under test.
|
|
|
|
|
|
|
|
<div align="center">
|
|
|
|
<img src="../uploads/357348f0582791b4ea451a3cce4b31bb/tx_4ch_r3-4_render.PNG" width="600"/>
|
|
|
|
|
|
|
|
Figure 5: Four Channel TX Card
|
|
|
|
</div>
|
|
|
|
|
|
|
|
### 4.2.3 RF Over Fiber Performance Analysis
|
|
|
|
#### Single Channel Analysis
|
|
|
|
|
|
|
|
The single channel transmit module was originally designed with a matching network at the laser, with the intent of improving gain and in-band spectral flatness. However, it was later removed from the design as it was found to be unnessary with the four layer board stackup. While the matching network was working to improve gain and flatness in early tests with a different substrate, it reduced both gain and flatness on the OshPark 4-layer boards. With an extensive campaign of redesign and testing a variety of approaches, it was found that best gain and flatness was achieved using no matching network and the shortest possible coplanar waveguide traces between final TX amplifier and the laser diode. This design is now implemented on the new 4 channel PCBs.
|
|
|
|
|
|
|
|
The "TX r3-3" transmit module was also tested at different attenuation levels and input power levels to ensure linearity. As seen below, the single channel transmit module maintains its spectral response shape over a wide range of variable attenuator levels and input power levels up to about -40 dBm.
|
|
|
|
|
|
|
|
<div align="center">
|
|
|
|
<img src="../uploads/2e233fa91f9e84fb46493100274513da/tx_r3-3_attenuation_test.png" width="500"/>
|
|
|
|
|
|
|
|
Figure 8: TX r3-3 tested at different attenuation levels to show the spectral shape remains the same
|
|
|
|
|
|
|
|
<img src="../uploads/0ebf782652409c70c4d94af7154a732d/tx_r3-3_linearity_test.png" width="500"/>
|
|
|
|
|
|
|
|
Figure 9: TX r3-3 tested at different drive power levels
|
|
|
|
</div>
|
|
|
|
|
|
|
|
[Section 4.3 >>](./4.3) |
|
|
|
\ No newline at end of file |
