Click to email me

July 2010 and I have been at last found some time to make some more progress.  In many respects the time off has been useful as it has required me to rethink my hardware design which has resulted in significant refinements and improvements.

I have built a new prototype board using a new processor and redesigned the analogue stages with considerable effort put into optimizing transient and frequency response and minimizing
noise. The new design includes improvements in component count and tolerances, circuit protection, and current consumption. The new board is also significantly thinner than its predecessors, trimming 5 mm off the case width.  See Figure 62.

Figure 62. My latest PCB. More than 100 components in less than 3 square inches!

The signal to noise discrimination is at least three times better than anything I have built to date as shown in Figure 63.  Transients, while still propagating through the analogue front end, have minimal ringing and are amplitude limited. The gain of the circuit is actually better than than the photos suggest because the transmitter used in the test was not optimally tuned to the resonant frequency of the transducers.  This is all good news.

Figure 63. Signal to Noise at Maximum Test Range. Vert. 50 mV/div. Horz. 20 ms/div.
Compare with Figures 44 and 56.
 

I have completed the hardware testing and my current focus is now on implementing, refining and documenting the software for the new microprocessor. Unfortunately this is not just a case of transporting my earlier code from one hardware platform to another because of the requirements of my new (and improved) DSP, the features of the new processor, and my desire to reduce power consumption.

13 August. Its taken a week but 80% of the code is written and I am well into the last 80% working on the critical receive DSP routines (these percentages are typical of how programming proceeds in my experience). The speed and additional memory in the new processor are letting me attempt all sorts stuff to make for better signal discrimination.  I my most recent simulations I have been able to reliably detect signals buried deep in noise (but recovering amplitude information has not been so easy).

I have also been busy updating the manual for the Buddy Locator. All good so far.

23 August and the initial version (V1.1) of software has been written and tested.  Aside from the effort in getting the DSP to run efficiently I have incorporated a number of refinements to aid with maintenance and field testing, improve usability and reduce power consumption. In the process I have also incorporated two improvements in the hardware but I am of the opinion that this is now as optimized as I am going to get it.  I need to get the hardware in a case (yip, I am going to build another two cases) and out in the water to test the effectiveness of the new DSP routines. Expect a report from the latest field trials soon.

29 Aug.  Boy have I been busy! I have been redesigning the case to suit the new, thinner, PCB; proving the noise reduction algorithm by simulation; completing my bench testing; and designing my own piezo switch because the Schurter device that I have been using for the prototypes is just too expensive and I haven’t been able to identify an economical alternative. Aside from the case, this is the most costly single component in the bill of materials. My only concern at this time is the potential cost of the case.  The pressures at recreational dive limits requires relatively thick walls which pretty much precludes injection molding. While I have made lots of cases by hand these are not suitable for a commercial device with O ring case seals and fillets. I have sent my first set of drawings off to Proto Labs to see what they can do with their fast CNC machining process.  Fingers crossed.

30 Aug and another day draws to a close. The new peizo switch is operating successfully and reliably on the test bench.  It draws no current when off, accommodates  gradual changes in pressure and is completely water tight with no moving parts or messy O rings.  It is also somewhat smaller than its commercial counterpart being just 12 mm in diameter and with a case depth of 7 mm and this better suits my application. And while the Schurter switches were costing around $70 each my switch comes in at less than $3 including the electronics. Because the new switch is in-essence another 40 kHz transducer I have gone to some lengths to ensure that there is no interaction between the switch and a transmitting transducer when they are mounted in the same case within an inch or two of each other.  Another problem solved.

In order to progress with the field trials I have started cutting some more Perspex for two new cases. While I could reuse the old cases I would rather retain these with there current assemblies for comparative noise measurements.  Hopefully I can get the cases finished tomorrow.

3 Sep and I’ve had to take a couple of days off the Buddy Locator to focus on my postgraduate studies. Still, there has been more progress. I’ve received and tested a new pulse transform core which is slightly smaller than the predecessor but with improved magnet properties.  This change was warranted to reduce both cost and the case size. And today I received some important components that have been on back order for some weeks which will allow me to make what I hope will be the final prototype circuit boards (subject to further field trials). The new cases are all but complete but I am waiting on some M2 thread inserts before fitting the bases.  I’m expecting these next week. Finally I have a number of new piezoelectric transducers on there way from China to enable field testing of my new switch design. FedEx has them in transit.

I have a further contract to complete to get some new PCBs made and machine populated but the manufacturing company is waiting on me to send them my nondisclosure agreement.  If someone had a mind to they could readily copy the hardware with minimal reverse engineering. Thankfully the software is critical to the function of the device and this is hardware locked from being read.  Further, programming is completed by me after assembly and prior to burn in, bench functional testing, pressure testing, packaging and quality control.  I’m not saying the software can’t be copied or its functions replicated but I figure that the time, effort and expense will be prohibitive.

The only significant sticking point at this time is the case cost. I can get beautiful CNC machined cases in moderate quantities in just a few days but the quotes to date have been prohibitive for commercialization. A former colleague suggested that I had another look at  underwater camera housings.  I picked up a disposable 15 m underwater camera from the supermarket tonight for $27 and set about disassembling it.  It has 18 injection molded parts (some of them extremely intricate  highly polished, and over-printed), two springs, a three-part pressed metal film housing with felt light seal, one metal and one plastic stamping, a roll of film, printed adhesive-backed light seals and some adhesive tape all packaged in a sealed plastic film bag in a cardboard carton.  The only thing it didn’t have was any electronics.  All this and a functioning camera for $27 with profit for the manufacturer, distributor  and retailer! This was an inspirational exercise.  I now know I can make a suitable housing within the target price and the molded parts have given me some new ideas.  It will just take some time to work through the design from my new concept sketches.

Once my studies are complete I shall set about making and populating two new PCBs (I up to Version 17d), make a couple of minor software changes (V1.2c - I started numbering again with the new processor) to suit the hardware revision, bench test them and get them in the water.

Anyone who has read this far will understand the difficulties I have had trying to find a reasonably priced commercial piezoelectric switch which has forced me to make one.  For anyone who is interested in making a reliable waterproof switch as good as any commercial device on the market but at a fraction of the price then click here to read/download my design.

Still lots to do and too little time.

More to follow ...

[Turtles are Here] [750Z Housing] [Construction] [A20 Housing] [Drawings] [Scuba Buddy Locator] [The Idea] [Proof of Concept] [Prototype] [Latest Update] [The Maths Behind] [Photo Galleries] [Galapagos] [Data Logger] [Making PCBs] [Links] [Contact Me]