| On the right, you can see our new h-bridge boards, improved over our first custom boards of last year. They can now handle higher current (at the "expense" of restricting us to not greatly exceeding the 28V batteries we've been using since 2005), and the bidirectional current sense has been calibrated. The boards were designed to support active braking, although this has yet to be tested. |
|
| The first plot shows that, except at very low currents, the resistive load gives a very linear current (measured with an ammeter) with respect to the driven PWM; this is not surprising. Below it, we see how that trace deviates from linearity. These are likely consequences of the operation of the bridge itself, and, assuming the resistors are ideal, gives an idea of how closely the PWM relates to voltage. In contrast, below it, we see that the Hall sensors on the boards, when measured in-system (stated because it's possible other errors have crept in, aside from those inherent to the sensor), give nonlinearities deviating from the true value by as much as 0.4A. We have compensated this using a lookup table in the host PC software. A future iteration of the boards may instead use a current sense resistor, but the Hall effect sensor appeared to be a much less painful way of obtaining galvanically-isolated current sensing. |
|
| We've made a few updates to the Atmel AVR boards described in a previous post, but the most important changes were largely connectors. This replaces the old motor logic boards, for which the programming toolchain had difficult licensing terms. The software supports overcurrent shutoff on a one-shot, and has processing capacity to spare if we ever decide to implement current-control, rather than PWM control. |
 |
 |
| These are the new sonar AVR boards. The idea is that while these only run a few channels each, they can be daisy-chained together, and will synchronize at startup. Additionally, provision has been made to bring one of the digital inputs in that still has frequency information so that we can experiment with that method of distinguishing the pingers in the future, even though we currently use the 0.9/1.1s difference. |
 |
 |
| The new analog sonar boards, shown here, have been improved by fixing a number of design mistakes from the previous iteration, and the results have allowed us to avoid the need for the extra conditioning board up front that we needed in 2008. The frequency range has also been expanded to 22-40kHz, to allow for Dave's custom pingers (which have been promised for several years now). New LEDs now indicate when the hydrophones are disconnected by detection of the phantom power current drain, in addition to the legacy LEDs that indicated pings and power presence. This is all the more helpful, since we have more hydrophones attached to the faceplate than we did in 2008, and allows us to easily figure out which connector goes to which processing channel without unsealing the hull. Careful layout of the ground plane was perhaps one of the most time-consuming, if not most helpful, improvements. Additionally, we've found that layout improvements and some additional circuit changes have allowed us to detect the pinger when we're pointed straight at or away from it, so that, unlike 2008, we obtain good results over all pinger azimuths now. |
|
| Thus, if you'll recall, we always chose to "slide" in pointed 30 deg to the left of the pinger before. This, together with some controls improvements, will speed up our runs. At right is an end view showing the new connectors (standard 1/8" stereo jacks) and the analog/digital boards connected to one another. |
 |
 |
| Last, we've added a spare regulator board, currently configured for 12V, which can accomodate any of 6 TI/PowerTrends switching regulators that we frequently use. We plan to use this if we wish to switch to a 12V wireless buoy, since most models we've found are 12V, rather than 5. Connectors are available to chain the input 28V to additional boards, as needed, and screw terminals are available for last-minute connections, although we hope not to have to use them. A number of parts are shown unpopulated, including a place for a potentiometer for running an adjustable voltage. All boards shown on this page were manufactured by Advanced Circuits. Thanks also go to former team members: Collin for reviews, and Erik for quickly populating the boards. |
|
|
