A Wide Variety of free 1-wire slaves arriving soon. What do you need?

[ericw]

Wookey,

Data sheet for HIH5030 is here http://mouser.com/catalog/specsheets/Product_Sheet__HIH-5030_and_HIH-5031_Series_Humidity_Sensors.pdf

Accuracy plus interchangeability  +/- 3%RH. While there is a significant tempco, the effect will be quite small in a domestic environment with only a small range of room temperatures.

Their price £7.42  (£5.42 for a 10 room mansion) but it is a surface mount device
Wire ended HIH4010 is £11.15 but is only +/- 8%RH

Eric

[StBarnabas]

The cheaper humidity sensors tend to be capacitance based. This capacitance variation might be measurable using the PWM generator and a comparitor or ADC. The HCH-1000-001 for example is £3.26 from Farnell

Another possibility is the HIH-4000-002 stocked by Farnell which gives a linear voltage outout and is rated at+-3.5%

http://www.farnell.com/datasheets/315720.pdf

this however is £13.60 dropping to £10.09 for 10. (whearher ten rooms constitutes a mansion is open to debate!).
 

[mountaindude]

Extremely interesting work, congrats to the good progress so far.
I just started looking into a similar concept (multiple virtual 1-wire sensors within a PIC or Arduino) a couple of weeks back.

With the risk of asking something obvious: are you looking at open sourcing the design/code for this, or is the target a commercial product?
Short term I'd be most interested in having this as part of an Arduino based setup I am working on, and it would be great not having to start from scratch..

Keep up the good work,
-Goran

[ericw]

Goran,
Its not intended to be a comercial product, I suspect that Maxim would take a dim view of anyone selling almost clones of their products.
If you PM me I can send you a copy of the code used, so you can see how it was done.

I'm a little puzzled by the Arduino connection however. If you are building a master to read devices on the 1 wire bus, as the PIC's are specifically designed to look like DS18B20's, all you have to do is be able to read the standard devices.  There is quite a lot of Arduino code around on the web for doing this.

If you are wanting to read other types of sensors directly, then again there is Arduino code around for doing most things.

Given the price of PIC chips, I don't think the economics of using a Arduino for a slave make much sense, However if you are thinking of doing a 1 wire slave then as there are some very tight timings, you would have to program it in assembler rather than the the normal Arduino programming environment.  Assembler coding is very specific to the micro concerned and I'm not sure you could transfer much useful information.

[StBarnabas]

Hi Goran
At present there have been three people contributing to the PIC code EricW, Mike (Meliffera) and to a lesser extent myself. Wokey and Wyleu have been trying out the chips for robustness.
The current state of play is that the technology seems to be robust. Getting the PIC to work fast enough to mimic a one wire slave is the difficult bit. One wire masters are a doddle by comparison. However largely due to Eric’s tenacity we have succeeded.

I and I think all others involved so far are in it for love rather than money. I have no commercial interest – though would like to be able to buy a flexible data logger with temperature, humidity, irradiance, rms current, rms voltage, frequency, power factor, harmonic distortion, flow rate etc. with a mixture of wired and wireless options for a reasonable price. We have a bottom up KISS approach at present and the current phase is creating these sensors with interfaces such that they look like existing 1-wire slaves hence existing masters should be able to read them.

We have a very much open access approach I am happy to share what I develop and I’m fairly sure others feel the same. Do you have a programmer which will burn Pic12F683s? You can PM me with details if you wish. I assume you are also committed to open access?

KenB is looking at Arduino systems. It would be well worth looking at his threads also.

P.S.
I have seen Eric's reply and on the face of it KenB might indeed be the best person on this forum to get in contact with

[mountaindude]

Eric: My mistake - was thinking two thoughts at once and wasn't clear. Have worked on a 1-wire based sensor network lately, with Arduinos handling certain tasks (but then injecting the measurements directly to the small server running MySQL). There are effectively two separate networks here (1-wire based + Arduino with RF links and hooked into a small Linux server) - this will remain for now, but it would be nice to migrate towards a single network (probably 1-wire based given the cost of RF-equipped Arduinos..).

DS2423s counters have served well but are becoming hard to get and expensive.. this led to the idea of consolidating several sensors in a microcontroller, and with a bunch of Arduinos lying around I planned to use them initially. But long term cheaper microcontrollers are of course the way to go.

I don't have the needed programmer right now, but this project just might be the reason to get one.
The 1-wire setup here is a HA7Net with 3 network legs from it, with about 12 sensors in total right now, a handful more to come in a couple of weeks. Using thermd for logging and graphing, looking at replacing/complementing this with Pachube and a suitable browser based charting tool.
So if you would like to test your work on this setup I'm happy to help, short term I cannot program the uc myself though.

-G

[StBarnabas]

Goran
sounds like a very interesting set up. I am happy to send you some PIC slaves just decide what you need and PM me your address. I have not had a chance to try out Eric's DS2423 counter so that might be an interesting one to look at. The disappearance of the DS2423 is one reason why I gou involved with Eric on this - apparently the DALLAS?MAXIM new chip-fab set up is unable to make them.

Eventually I am hoping to move to a setup more like yours and your skills look very valuable so in return for the chips your expertise may be called upon at a later date!

[ericw]

Goran,
We have two different designs for a DS2423 'clone' and would be very interesting to see if there is a difference in practice between the two methods used for the 1 wire interface, especially in a network environment in which it hasn't been tested before.

They only mimic the dual counter part of the functionality, dummy data being dumped from where the memory contents would be.

What sort frequency are you wanting to count?

[mountaindude]

5-8 Hz max given current plans.

-Goran

[StBarnabas]

Eric
if you mail me the code I'll be happy to blow a few chips and post to Sweden. Would also like to try counters myself. Should have guessed with the name "Goran". Now who was that football manager?

[mountaindude]

Would you mind sending the source, I'd be interested to have a look too. Are you working in c, asm or something else?

[StBarnabas]

Well
The past few weeks have been hectic so have had little time for Forum related activities. I have finally gotten around to testing Eric’s PIC DS2423 mimicking code. This mimics the 32 bit counter functionality but not the memory of the DS2423. This is particularly useful as the DS2423 has been discontinued and increasingly hard to get. There seems little out there that will replicate the counter functionality so this is very useful.
The test was positive with a LogTemp screen dump showing the counter working. At present the counter is volatile – power was removed from the PIC for a few seconds and the counter resets to zero. It may well be possible to create a non volatile version by writing to the data EEPROM on the PIC. Does that sound like a useful exercise?
StB

[russf]

I've been considering building a PIC implementation of 1-wire slave devices and was amazed to discover your fervent activities and evident success.

It strikes me that the project would benefit from its own page on one of the popular open source repositories, to encourage a larger number of people to help.  I would be happy set that up.

My energy saving applications are based on DS2406, DS2413P, DS2438, DS18S20, but several of those applications would integrate nicely into the PIC, cutting down on component and assembly costs.  I'm using honeywell humidity sensors at the moment, and would be able to move to the cheaper capacitative sensors using PIC.

My experience is in programming in python, C, assembler, on many platforms, including PIC10 and PIC12, and I do some hardware design. I'm British, based in California. How can I help?

[StBarnabas]

Hi Russf

sorry for the delay in replying . Been a bit busy. Your background certainly looks useful! I'm sure there is a lot you could help with but we remain busy on sensors and sadly in my case even busier on the day job! There is a LOT going on and things may accelerate again later this year.

My vision is sensors for all - high quality at bargain basement prices. A decent solar irradiance sensor should appear soon, and a flow sensor not long afterwards. I am toying with a capacitative humidity sensor  also and some input would be very useful.

Dallas/Maxim frown on slave clones so we would have to be careful re repositories - but EricW who has written the 1-wire slave interface needs to be kept in the loop.

Thanks again for you interest  and PM me so you can get a copy of the current DS18B20 clone.

Sean


 

[russf]

Just following up with a public post after my PM to Sean.

I'm very interested to see the slave code, and if possible add some device families to it.  I'm interested in the DS2406, DS2438, DS2413, and DS18S20. Apparently Eric is implementing the DS18S20 at the moment. I'm interested to see the slave code, as well as the schematic for the sensor part. Eric?

My goal would be to add a conditional compilation switch to enable a functionality like the DS2413 first. The 2406 has EEPROM, in addition to the 2 bits of GPIO that the 2413 has.  The 2438 is more complex, and I would not duplicate it. What I need for that task is 2x10bit A/D and say, 2GPIO. So there will be a simple progression and addition of a couple more features for each device.

I think that it's probably quite safe to proceed with these implementations, from the legal POV. I would suggest setting up a public repository, and licensing the code under the MIT license, which gives anyone right to use it in any reasonable way, and will encourage further contribution of expertise, bug fixes, etc.

My email is russf@topia.com and I will only look and test what I receive. Won't publish or do anything else until we have spoken more about it.