Cumulus Support

This may help - I haven't studied it, but the Peet website says no slip rings.
http://www.google.com/patents/about?id= ... dq=5231876

Gina,

Done some digging around and it seems Maxim have discontinued the DS2423 chip. http://www.maxim-ic.com/datasheet/index ... /2912/t/al with some interesting comments on its withdraw in 2009 http://www.maxim-ic.com/datasheet/index ... /2912/t/uc

Seems that alternatives are available but without the counter function, which is what you actually want

Also stumbled on some interesting sites for 1 wire weather stations, which you may or may not of come across
http://www2.buoy.com/weather/
http://benjamin-donnachie.blogspot.com/ ... ation.html

Just wondering if you could incorporate the counting and A/D in a small PIC - the 16F88 has I2C bus, 10bit 7ch A/D, 2 comparators, all in an 18 pin package http://www.microchip.com/wwwproducts/De ... e=en010243

mcrossley wrote:This may help - I haven't studied it, but the Peet website says no slip rings.
http://www.google.com/patents/about?id= ... dq=5231876

Ah that explains it perfectly, thank you Magnet in vane part, reed switch in centre so always equidistant from magnet and the anemometer part puts a ferromagnetic shunt in between as it rotates. Very clever! I might try and emulate that. It's my private unit so shouldn't infringe the patent - as long as I don't market it, which I have no intention of doing. I might produce a more conventional unit though which I can publish in the public domain - depends on spare time and inclination

I'll answer this in sections...

malc-c wrote:Done some digging around and it seems Maxim have discontinued the DS2423 chip. http://www.maxim-ic.com/datasheet/index ... /2912/t/al with some interesting comments on its withdraw in 2009 http://www.maxim-ic.com/datasheet/index ... /2912/t/uc

Thanks for those links - explains a lot - and produces yet another furious future Maxim customer!! It really beggars belief! They produce just the ONE counter chip in the whole 1-wire range, advertise and extol it's virtues for 10 years then suddenly (once everyone has designed their stuff round it) stop producing it. "How to win friends and influence people" - I think NOT!!

Disgusted of Devon!!

malc-c wrote:Also stumbled on some interesting sites for 1 wire weather stations, which you may or may not of come across
http://www2.buoy.com/weather/
http://benjamin-donnachie.blogspot.com/ ... ation.html

Those look interesting - I'll study them later. Thank you

Just wondering if you could incorporate the counting and A/D in a small PIC - the 16F88 has I2C bus, 10bit 7ch A/D, 2 comparators, all in an 18 pin package http://www.microchip.com/wwwproducts/De ... e=en010243

I'll look into that but I think if I want a counter I'll also look at the feasibility of a CMOS counter and maybe the DS2408 8 line addressable switch. I'll check the power consumption of CMOS (very low I think) and see if parasitic power would be suitable. Otherwise a 3v lithium cell might be adequate.

Gina wrote:They produce just the ONE counter chip in the whole 1-wire range, advertise and extol it's virtues for 10 years then suddenly (once everyone has designed their stuff round it) stop producing it. "How to win friends and influence people" - I think NOT!!

Disgusted of Devon!!

LOL -love the sign off

I quite agree... I noted that one poster had only just had a design published in a national magazine at the time it was withdrawn !

Given the comments it's amazing that no other competitor has stepped in and offered a replacement !

Gina wrote:Ah that explains it perfectly, thank you Magnet in vane part, reed switch in centre so always equidistant from magnet and the anemometer part puts a ferromagnetic shunt in between as it rotates. Very clever! I might try and emulate that.

Ingenious indeed, and purely digital, no A/D involved so potentially the head with no electronics could be on end of a long wire (reasonably) safely.

Gina wrote:

I think this has been mentioned before, but I have also seen it can done using the cup anemometer only by adding a small tab to one cup and using two switches (Derek Weston), the phase difference can again be used to calculate the wind direction - but no additional direction vane is required at all.

Yes, I've read about this method but don't remember the details (if they were given).

I found a couple of web pages on DIY 'Derek Weston' units (they seem to be accurate to about 5 degrees or better - 1.5 deg - with higher precision units).
This one shows the optical encoder disk:
http://homepages.ihug.co.nz/~chris-s/anemometer.html
Here is the original (DIY version):
http://home.alphalink.com.au/~derekw/an ... nemain.htm

mcrossley wrote:I found a couple of web pages on DIY 'Derek Weston' units (they seem to be accurate to about 5 degrees or better - 1.5 deg - with higher precision units).
This one shows the optical encoder disk:
http://homepages.ihug.co.nz/~chris-s/anemometer.html
Here is the original (DIY version):
http://home.alphalink.com.au/~derekw/an ... nemain.htm

Thanks for that Very interesting. Yet another design to play with

I like the idea of making the anemometer provide wind direction as well as speed but the software could be a bit complicated. Also, we're now looking at sampling the data output at something like 16 times the maximum rotation rate (minimum) if we are to capture the timing of the chopper disc and evaluate direction into 16 compass points. We've gone from logging data at something like once a minute through a second or two to gather gust info and now... well, goodness knows, I'd have to experiment to check revs versus wind speed. But as a "ball park figure" if 1 mph gives 1 rps then 100 mph gives 100 rps or a sampling rate of 1600 per second. I forget what the 1-wire time scale is and haven't been that much concerned when the sampling was of the order of a second. The other consideration is the software reading the 1-wire devices and speed of calculations following. Again I have been basing my choice of programming language on there being plenty of time to do things - therefore going for ease of programming rather than speed.

In view of all the above, I don't think I'll be using the tabbed anemometer method of measuring wind direction.

I like the Peet Brothers idea - although slightly complicating the hardware (in the sense of tubes etc.) it reduces the circuitry considerably. Using the principle of mixing analogue and digital to reduce the part count, I could get away with just one 1-wire component for anemometer, wind vane and light sensor.

If I were to use the Peet Brothers invention, the circuit could be reduced as shown below.

The 2 reed switches in conjunction with the 3 resistors will produce a digital value of 0, 1, 2 or 3 which means that one data bit represents one reed and another bit the other reed. The photo-voltaic cell now feeds the 10v full scale A/D which is adequate as the cell in question produces just over 2v in sunlight. If reed contact bounce is evident then a capacitor across the lower (sense) resistor will smooth them out.

Been thinking about this (Peet Bros system) and realise there's a snag. This will also mean sampling the reed switches at least 16 times per revolution in order to provide sufficient timing accuracy. That's assuming the simple 1-wire circuit above. If the reed switches were to be connected to an interrupt on a microprocessor or to a counter plus timer chip the timings could be measured directly. Again, this is getting away from the 1-wire principle but a 1-wire chip could be used to send the data to the computer.

Been looking into providing a counter for 1-wire. A CMOS 4040 chip is a 12bit ripple counter and available in the UK from Maplin. The data outputs could be fed into one or more DS2408 1-wire chips. Messy but do-able and it still comes nowhere near the two 32 bit counters in the DS2423 (actually 4 counters, 2 with hardware clock inputs). However, the DS2408 data sheet is marked "Not recommended for new designs" so I guess they're thinking of withdrawing that too!

This is getting silly!!!! No!! This is getting beyond a joke!! Are they trying to kill off 1-wire?? It makes you wonder.

Looks like the mixed digital/analogue may be the thing yet again using the DS2450 4 channel A/D and resistors. What a bodge-up!

One other protocol you might want to consider is SPI (Serial Peripheral Interface). There may be more options, and some may include counters, but I've not checked.

malc-c wrote:One other protocol you might want to consider is SPI (Serial Peripheral Interface). There may be more options, and some may include counters, but I've not checked.

Had a quick look at that and it seems rather complicated. I'm persevering with 1-wire for now.

I've just made up a 1-wire temperature sensor with cable long enough to reach from the computer in the lounge and out via the kitchen and back porch, across the grass and up into my Stevenson screen box, next to the F.O. unit. The F.O. is reading 5.0C and the 1-wire 4.625C. The 1-wire is specified as accurate to 0.5C and they are within 0.5C of each other The resolution of the F.O. is 0.1C and the 1-wire is 0.0625C ie. 1/16 of a degree. I estimate the distance to be something like 18-20 metres. I plan to thread the cable through plastic water pipe and bury that under the grass to protect it from the goats.

Has anyone in the UK managed to order anything from Hobby Boards? Just been on their site and they have DS2423 counter chips for $8. They're also selling counter board modules at $28 but have run out of the kits. I created an account with them - that went fine, it accepted my UK address without any problem (a drop drown list to choose country from). Then tried to buy some stuff. I got as far as the delivery options and it said USPS (United States Postal Service) was the only available delivery service available on this order and that an error had occurred. I tried again with the same result. I had ordered 5 DS2423s and a humidity sensor. Another dead end it would seem!

HomeChip UK have a counter box which also uses the DS2423 counter chip for £28 (28 UK pounds) (ie. £28 + postage +VAT). Bit pricey for the simple thing I want to do. Do I spend three dozen quid to avoid the hassle of CMOS and all the extras? I dunno. I don't think software + 1-wire sampling at 200 times a second is going to be practical. The anemometer reed switch would need checking at least twice a revolution and allow for the maximum wind speed.

Maybe I should crawl back into my shell and hibernate until spring and perhaps then the I'll have forgotten the whole idea! What would have been easy-peasy and dead simple a couple of years ago has now become a real headache! I guess I'm lucky that this is only a hobby and that my business does not depend on the DS2423 as I'm sure many do.