04/05/2009 by peter.

The oscillating op amp was down to cable impedance I think. In any case the answer is to put the strain bridge amp in the little blue box thus turning said cuboid into a mast head amplifier (no pun intended). while thinking about this i wondered about putting a uC with an ADC in there too and then just have a master controller with a network(maybe I2C/SPI or a PAN radio net) that would allow any add on you like to be put onto the net and involved in input or output.

Further to all this i would also like to have some real intelligence (def:using past experience to evaluate current experience to predict future experience and modify actions appropriately) in the thing, a bit like the SaSi to allow the user to decide which actions they like and which ones they want to have less often.

That of course would be a total redesign.

However if I am to redesign it from scratch I will use a touch screen OLED display for user interface and write all the source code in C on an AVR.

Delays delays delays. Just as well I don’t design these things for a big company.

Posted in intelligent vibrator | Print | No Comments »

08/04/2009 by peter.

As can be seen in the picture below i have finished populating the first board.

(ignore the soldering blocks behind, my bench is currently covered with jewellery making stuff and vibrator making stuff)

faults so far found:

Power transistor not connected to the correct pin of uC
Tracks to voltage regulator end before they arrive
Opamps wired up back to front
pressure transducer acting as a theremin so something is wired up a bit wrong.

Always happens though, first PCB version never works. Looks quite professional though!

Posted in intelligent vibrator | Print | No Comments »

04/04/2009 by peter.

Apart from the fact that I would appear to have forgotten to order some of the bits (or I have an envelope of them somewhere safe) The population of boards is coming along. as can be seen in the photo below.
So far I have noticed 2 minor screw ups in the design and a few bits that can be optimised. That’s what happens when you make PCBs. You spend a day looking for faults and then you find them within three seconds of open the envelope of boards.

With a bit of luck it will all work once the rest of the bits arrive.

Posted in intelligent vibrator | Print | No Comments »

02/04/2009 by peter.

Finally the boards have arrived. now to populate them

Posted in intelligent vibrator, Stuff | Print | 1 Comment »

05/03/2009 by peter.

while I was looking at the PCB design I felt it was a bit wasteful of space so I started from scratch and redid it. it is now just 3.5″x1.85″.

After looking for a board house to make up PCBs for this little endeavour I found that crazily if i get them made in the UK there are few people with on-line shops for that sort of thing. if I go to mainland Europe the best I can hope for is still £50+ but if I use Sparkfun’s little side company, BatchPCB, the cost drops off enormously. They only charge $2.50 a square inch plus $10 and then postage. What is crazy though is that it is actually manufactured in china. so my design goes to America, where it is then sent to china for production, then the board goes back to America and then from there to me and that’s cheaper than going down the road to get it done!!!

anyway, boards are now ordered and should arrive sometime this month.

Posted in intelligent vibrator, Stuff | Print | 2 Comments »

01/03/2009 by peter.

Well after much faffing around I have finally finished the first draft of the PCB for this thing.

Its not pretty or well laid out but with a bit of luck it will work.

Getting there….

Posted in intelligent vibrator | Print | No Comments »

29/11/2008 by peter.

To machine the various square orifices required for the USB port and the LCD display I need to use my rebuilt Boxford. One problem, the software was still buggy and I haven’t written anything to generate control codes to give it. Simple solution, write a manual controller. That took all of 2 days of spare time to add all the features I wanted and with this I set about making the rectangular hole for the LCD. After much standing and watching the machine produce fine swarf as it slowly ploughed through the metal I had a nice rectangular hole. I de-clamped the box to admire my work when I realised I hadn’t drilled the display mounting holes which would be a lot easier under numeric control. Unfortunately while edge finding the rectangular cut-out, I crashed the machine into the top corner of said orifice (hit “End” instead of “Page up”) then I got the holes slightly in the wrong place but that’s one reason why I will make 2 of them. The first to screw up one and second so that I can test new code and features on my own one before sending the updates of to a beta tester.
The other minor annoyance is the lack of M2.5 nuts to fit the M2.5 LCD retaining bolts. For reasons of stupidity, I ordered M2 nuts. Well it will have to wait until I have a few more bits to order. There’s no point in paying postage on only one tiny bag of nuts.

Anyhoo. Piccies of the box so far

Posted in Boxford Duet, intelligent vibrator, Stuff | Print | 1 Comment »

08/11/2008 by peter.

With about 99% of the parts now arrived I have commenced the construction. The first part to be completed is the measurement head. “Why such a small part?” you may ask, well it is simple, I need to cut out a rectangular hole for the LCD and that requires me to find the clamping screws for the CNC mill which I have ever so carefully put somewhere safe. Once I find them then I can cut that out along with the hole for the USB port.

This is what I have so far. A box with a pressure transducer in one side and an M8 4 pin industrial connector in the other.

Onwards!!!

Posted in intelligent vibrator | Print | No Comments »

12/09/2008 by peter.

After much faffing around (about a year since i started this particular folly) I have actually got a Beta version finished. WOOHOHOO!!!

Some videos to show what it does.

http://www.vimeo.com/1718065 :showing the bootloader in action (only changing the version number in this example)
http://www.vimeo.com/1718090 :Biofeedback 1
http://www.vimeo.com/1718106 :Biofeedback 2

Now just to polish the code and make up the final beta hardware…

Always something else to do

Posted in intelligent vibrator | Print | No Comments »

12/09/2008 by peter.

So far there are 2 biofeedback algorithms in the this contraption although this is by no means all that there could be.
One is a straight feedback system that simply adds the input form the sensor to a pre-set in a way which will have no overflow control. This means that with zero pre-set the output will be positively proportional to input. With a little pre-set, less force is needed to reach high power and the motor can be kept ticking over. With a good bit of pre-set then to get max output muscle action must be maintained below a threshold or there will be an overflow and the thing will stop until the muscle relaxes again. finally with a lot of pre-set then the device will be at max at rest and the more force excreted the vibrations will drop off and then ramp up with pressure.
The second will have a time delay in it so it will ramp up slowly, then once a muscle pulse larger than a pre-set value comes along it will drop off, wait and then start ramping again. This can be tweaked in a number of ways e.g. the dead time can be modified based upon the frequency of pulses received or the pre-set could vary with time, either increasing or decreasing.
Of course data can be fed back and forth to the computer which will be able to carry out more complex algorithms which opens up amazing possibilities….

Posted in intelligent vibrator | Print | 2 Comments »