1-wire info and recommendations

Gina · **Posted:** Sun Jan 02, 2011 9:56 pm

I've been experimenting with software - running a loop continuously reading a 1-wire chip and reading data. This is just the simple DS18B20 thermometer chip, so no variation in output over a period of just seconds. Tests have shown data gathering to be very fast - a million reads in 5 seconds. But is this actually taking a new temperature reading every 200us? I think not! I think it more likely that it's just reading a buffer.

I increased the loop count in 10x steps until the process took several seconds to complete the count. I ended up doing a million loops reading the temperature and the computer clock in about 5 secs. At every 100,000 counts it prints the count, time and temperature and logs the latter two to a file. This is the terminal output :-

Quote:

gina@AMD64-Mint-9 /mnt $ sudo python ~/Weather-1-wire/ReadTemperature.py 28.6686E0020000 ~/Weather-1-wire/Temperature.txt
100000 2011-01-02 21:10:51 1.125
200000 2011-01-02 21:10:52 1.125
300000 2011-01-02 21:10:52 1.125
400000 2011-01-02 21:10:53 1.125
500000 2011-01-02 21:10:53 1.125
600000 2011-01-02 21:10:54 1.125
700000 2011-01-02 21:10:54 1.125
800000 2011-01-02 21:10:55 1.125
900000 2011-01-02 21:10:55 1.125
1000000 2011-01-02 21:10:56 1.125

ATM the only 1-wire chips I have are DS18B20s and a couple of DS2450 4 channel A/D chips. I have not tried a DS2450 as yet. I think that's the next step - get one working and reading an input signal. I have a waveform generator which covers a large frequency range with sine, square and triangular wave output at 0-5v. That should be ideal for testing this chip. I would think other chips that perform A/D would work at a similar speed.

This testing should tell me if it's feasible to sample the anemometer reed switch at 200 times a second. That gives 5ms to communicate with the chip and process the results. It only needs to increment a software counter for each positive (or negative) transition of the reed switch data. Then the counter can be read at the data logging interval chosen, be it say 10secs for real time dials or 5min for normal plotting.

Gina · **Posted:** Tue Jan 04, 2011 4:05 pm

malc-c wrote:

Gina, it might be that as it's Sunday and they need to give the exact cost of shipping overseas and can't do so that's why they are not processing international orders. It might be worth trying to place the order on Tuesday when (possibly) the offices are manned

Eureka!!! You're right!

HobbyBoards international sales is/are working

I have just ordered 5 DS2423 counter chips @ $8 each = $40 plus just over $5 shipping. PayPal converted the total to just under £30 so that works out at about six quid each. That's not too bad. Probably didn't need that many but better than not enough :lol:

And they could run out any time.

Gina · **Posted:** Thu Jan 06, 2011 12:34 pm

I now have a DS2450 four channel A/D chip wired into the 1-wire network and working. I'm feeding a 0-5v very low frequency sine wave into one input and an attenuated version of it into another. I've added it into the logger app which now logs date, time, temperature and four voltages every 10 seconds. In view of the frequent data collection I've added the date into the data file name so I get a new file for each day. So... records per day = 6x60x24 = 8640. Actually, ATM the logging is a little over 10secs since I'm just using a 10sec pause in the logging loop. I'll change the timing to log at exact time intervals later. The readings from the 5v sine wave input go from nearly zero to nearly 5 in a 3 minute cycle, exactly as expected

Gina · **Posted:** Mon Jan 17, 2011 5:45 pm

Still waiting fro delivery of counter chips from HobbyBoards (probably in Customs) so I've been thinking... Ouch! :lol:

Many decades ago I was involved in a project requiring a number of simple counters - ICs were in their infancy. I came up with an analogue solution to a digital problem - a diode pump circuit. Now I have another totally unrelated use for it! The idea is to transfer charge from a small capacitor to a larger one. The voltage reached in the larger one is proportional to the number of small charges applied. By adding a "current leak" resistor, the average voltage produced is proportional to the frequency of charges applied. Below is a circuit which takes the closing of a reed switch and pair of diodes to pump charge from one capacitor to another. The application I have in mind is the anemometer.

Attachment:

00-diode-pump-cct-a.png

How it works :-
Parasitic power is obtained from the 1-wire bus in the usual way with Schottky diodes and smoothing capacitor. The anemometer reed switch causes the small capacitor to be charged and discharged alternately. During the charging, current is fed into the reservoir capacitor through a diode from the parasitic 5v supply. The discharge current flows through the other diode and resistor to earth. The resulting voltage across the reservoir capacitor may be measured with a DS2438 1-wire chip (called a Smart Battery Monitor) using the current sense A/D converter. This has a full scale input of 250mV. This relatively small voltage compared with the 5v supply maintains adequate linearity.

Component values may be calculated to give a full scale reading of 100mph or whatever is wanted. Resolution is limited to one charge unit (pulse) = half a revolution or the A/D resolution, which I think is 8 bits (1 in 256). EDIT Checked up - signed 10bit so 1 in 512.

A variation of the circuit could involve removing the leak resistor and discharging the reservoir capacitor after each reading but this would require an extra chip.

Of course, this is a bit more complicated that using a DS2423 counter chip with more components but it beats using a CMOS counter and DS2408 8 bit I/O chip.

Gina · **Posted:** Tue Jan 18, 2011 1:33 pm

Yippee!! My counter chips have been delivered

I guessed as soon as I'd sorted out an alternative, they'd arrive :lol:

nking · **Posted:** Tue Jan 18, 2011 1:40 pm

Gina wrote:

Yippee!! My counter chips have been delivered

I guessed as soon as I'd sorted out an alternative, they'd arrive :lol:

Always the case :roll:

Hopefully you didn't have anything extra to pay

Gina · **Posted:** Tue Jan 18, 2011 4:00 pm

No, it would seem not

The package came airmail. The chips are so small and thin that the package was no more than a slightly thick letter. They're about 3mm square and just over a mm thick. Just been "having fun" soldering fine wires onto the pads and then soldering the wires onto 0.1" stripboard to join the DS2450 A/D and other components.

This is the complete wind sensor circuit.

Attachment:

00-0-new-wind-sensor-cct-final.png

How it works :-
Parasitic power is obtained from the 1-wire bus using the 2 Schottky diodes and electrolytic capacitor. This is used both by the 1-wire chips and the sensor circuit. The anemometer reed switch feeds the DS2423 counter chip with a resistor to aid pull down. The reed switch signal is de-bounced in the chip and the operations counted in a 32bit counter. The wind vane operates 8 reed switches and the 47K resistors encode the result onto the 4 analogue inputs of the DS2450 A/D chip. The resistor arrangement is such that not only are the 8 primary points exclusively encoded but the other 8 in-between points are too, with two adjacent reed switches on at the same time.

nking · **Posted:** Tue Jan 18, 2011 4:44 pm

Gina wrote:

How it works :- Parasitic power is obtained from the 1-wire bus using the 2 Schottky diodes and electrolytic capacitor. This is used both by the 1-wire chips and the sensor circuit. The anemometer reed switch feeds the DS2423 counter chip with a resistor to aid pull down. The reed switch signal is de-bounced in the chip and the operations counted in a 32bit counter. The wind vane operates 8 reed switches and the 47K resistors encode the result onto the 4 analogue inputs of the DS2450 A/D chip. The resistor arrangement is such that not only are the 8 primary points exclusively encoded but the other 8 in-between points are too, with two adjacent reed switches on at the same time.

Well that has cleared any confusion I may have had :ugeek:

.................. If only I understood it :lol:

Do you think you would be able to feed your wind data into the FO Transmitter or is that just not feasible.

Gina · **Posted:** Tue Jan 18, 2011 5:56 pm

nking wrote:

Do you think you would be able to feed your wind data into the FO Transmitter or is that just not feasible.

Yes, you could do. A different resistor network, one for each reed switch could be used and then connected to the FO transmitter. In other words a copy of the circuit used by FO. Some of the resistors FO use are unusual values and of close tolerance. 1% metal oxide resistors should be alright with the odd values made up from two standard values. If you want the values, I can find it in the FO sub-forum (I think) and copy it here.

EDIT :- This is the thread :- http://sandaysoft.com/forum/viewtopic.php?f=13&t=2685

Gina · **Posted:** Tue Jan 18, 2011 8:06 pm

Although not really belonging in the "1-wire" thread, here is a diagram showing the resistors needed to feed into the FO transmitter.

Attachment:

0-new-wind-sensor-cct-FO.png

You could get 16K from 15K and 1k in series. Likewise 64K9 from 62K and 2K7 in series to give 64K7 which would be near enough. If you want to use resistors in parallel, you'd use higher values - formula is 1/R = 1/Ra + 1/Rb. Or R = Ra*Rb/(Ra+Rb). Obviously the series arrangement is easier to calculate!

nking · **Posted:** Tue Jan 18, 2011 10:36 pm

Thanks Gina - yet another project to look forward to.....

Gina · **Posted:** Sat Jan 22, 2011 1:36 pm

One thing I've discovered in setting up my 1-wire network is that although you can get RJ11 plugs to work with RJ45 sockets, they are too loose a fit for a reliable connection. Think I'll post this in the General Forum too as it applies more widely than just 1-wire.

Gina · **Posted:** Tue Jan 25, 2011 9:35 pm

Having got my 1-wire network and sensors working and software logging the results, I've found a niggling problem with the DS18B20 temperature sensor. They are prone to an occasional error resulting in a reading of 85C. I guess I shall have to detect this and re-take the reading when it occurs. I'm rather disappointed - I din't expect this sort of thing from a professional component. This is not just a bad sample - I've Googled the problem and found several references to it. There are other 1-wire temperature sensors but they are not as convenient to use and I have bought several DS18B20 chips with the intention of measuring several different temperatures.

So it's not just Fine Offset stations that suffer from temperature spikes!!

Gina · **Posted:** Mon Feb 21, 2011 7:38 pm

I now have a working 1-wire circuit for my new wind sensors. It uses the usual reed switch and magnet for the anemometer but optical encoding for the wind vane (I was unable to produce an effective read switch version). I have overcome the need for more current than available parasitically from the 1-wire network by switching the LEDs on only during the reading process. A fairly large electrolytic capacitor supplies the LED current for the short period they are on, being recharged in the off period.

Attachment:

1-wire_vane_optical_gray-11-02-21.png

The anemometer reed switch is connected to a DS2423 counter chip which accumulates counts - 2 per revolution of the anemometer rotor. The count is then read periodically to determine the wind speed (and wind run). Circuitry to remove switch contact bounce is provided within the DS2423 counter chip.

For the wind direction :- A DS2413 dual PIO chip is used to switch the LEDs and the current in the photo-transistors produces a voltage across individual resistors and the 4 voltages are fed into a DS2450 4 channel analogue to digital converter chip. Using analogue sensing allows variation in light levels, sensitivity and opacity of the encoding disc to be allowed for in the software.

Power to the ancillary components, DS2423 counter and DS2450 A/D is provided parasitically from the 1-wire network by a Schottky diode and reservoir capacitor. The latter is made big enough so that the voltage only drops by a couple of tenths of a volt when the LEDs are on. Another Schottky diode is connected across the 1-wire network to trap possible negative going spikes on the line. This is standard procedure to protect the chips.

Gina · **Posted:** Mon Feb 21, 2011 9:49 pm

Here's a better diagram...

Attachment:

1-wire_wind_optical_11-02-21.png

	It is currently Sun May 26, 2013 7:23 am
Please click here before posting. Help me to help you!
Useful Links: Cumulus FAQ • Enhancement requests • Wiki (documentation)
Please put your approximate location into your profile
Add your web site to the Cumulus user map
Vantage Pro2 users with firmware 3.00 should upgrade to fw 3.12 and Cumulus 1.9.4
Latest Cumulus release v1.9.3 (build 1059) - Feb 09 2013 Latest Cumulus beta v1.9.4 build 1078 - May 24 2013 --> Latest Cumulus Realtime v1.0.0.5 - Sep 05 2009

Sandaysoft

1-wire info and recommendations

Who is online