Article: Dew Heater Control and the Arduino

[mswhin63 (Malcolm)]

Cheers Hikerbob,

This was one project I was working on as well, although had to shelve it for the time being. I purchased some Nichrome wire and special silicon high temp flexible rubber strips about 1 year ago in anticipation. Completely forgot about the project after damaging my shoulder.

FYI I am considering using a different PCB design software and ordering package http://fritzing.org/ looks really great for PCB design and small production.

I have both PICAXE and Arduino and may consider one or the other for development. Still mulling over software flow design but this is a good start.

[hikerbob (Bob)]

Alistair, I don't have any science to support the choice but I've been working on only heating within 5 degree's of dew point.

Not sure if this makes sense but the relevant section of code is

if ( ChTempVal[TCount] <= dew_point + 5 ) {
ChPWMVal[TCount] = (dew_point + 5 - ChTempVal[TCount]) * 50;
if ( ChPWMVal[TCount] > 255 ) {
ChPWMVal[TCount] = 255;
}
}
else ChPWMVal[TCount] = 0;
analogWrite(ChPWMPin[TCount], ChPWMVal[TCount]); // set the PWM value

ChTempVal is an array storing recorded Temperatures
ChPWMVal is an array storing PWM settings. 255 is fully on, 0 is off.
ChPWMPin is an array storing the output pin number for a channel
TCount indicates the channel

I left some provision to use a switch and the variable resistor (which generally control's the display brightness) to be used to alter the profile but have not implemented that. I didn't allow for altering one channel and not the others. In theory the 5 in the above code would be replaced by a variable and if dew was being a problem with the standard range the variable resistor could be used to increase the temperature that the heating started.

Michael I purchased some nichrome wire as well and had a go at making a heater then went and bought some pre-made thick film heating strips. http://www.dew-not.com/Order_Form.htm

Love them. Home made might be viable in a fixed installation but it began to look risky and difficult for my observatory less equipment.

Bob

[hikerbob (Bob)]

BTW, I've ordered some bit's to try a Veroboard version of the remote controller. Not sure how much effort would be involved in getting the existing controller to talk to it but given I'm already using I2C maybe not too bad.

If it works as expected I'd have to look at some options for a surface mount version.

Since I wrote that I've been having another look at the pinout's for the 8M2 and am not sure that I can do both PWM and I2C, the one pin labelled with pwm is also the I2C SDA pin.

Should be able to do a version of PWM in code on another pin. Failing that look into the 14M2 chip's which do have PWM on pins seperate to I2C.

Bob

[alistairsam]

Hi bob, even if we can achieve independent duty cycles with a chip like the 28x2, we could have four channels off one.

I'll try and test this tonight by measuring frequency with an oscilloscope and a 28x2
You don't have to just use i2c, you could use serin as well but would be a bit more tricky

[Bassnut (Fred)]

Originally Posted by alistairsam

Hi,

I'm working on building a controller with independent pwm channels / distribted controller and as Bob mentioned, the picaxe's should be very cost effective even if individual mcu's are used.
I've used 4 picaxe's on the same I2C bus and they're easy to implement.

I'll carry out some tests and draw up a pcb schematic once I have the prototype working.
I'll try and make it modular so you can add on or remove channels as needed. only bit that will change is the code for the master mcu.

in terms of temp regulation, Bob what would the target temp range need to be at? 3 to 5 deg above dew point till 10 deg above dewpoint or a wider range?
the pwm freq can be linked to the sensor temp so they're not unnecessarily heated but still stay above dew point.

You don't need multiple picaxes or Pwm. The m2 chips are multitasking (multiple symultaneous programs) and the cycle rate on a dew heater is so low, simple timing is all that's required. With a single non-multitasking 28x2 you could easily control 6 or more heaters.

For one offs, vero board is a hassle. Picaxe proto boards or the kiwi board would be much easier, built in track work for coms and pwr regulator.

[alistairsam]

I was just reading up on the parallel tasking capability
Since it basically switches between tasks rapidly it might be a bit difficult to get a linear range of 4 independent duty cycles
Other option is just the pwm command that sends bursts
Ideal option is the Hpwm in single mode
Will test

[Bassnut (Fred)]

Why bother with hpmw?. Hi freq Pwm just creates noise. A 1 sec or so cycle time is fine and very easy to code with a timer.

The timing variation due to multitasking on an m2 is absolutely trivial.

[Bassnut (Fred)]

Just for example, I made a dew heater with a 08m2. The 8pin picaxe, a pot and a FET, just 3 components.
You tweak the pot and the heat changes.

8 lines of code in total. No PWM, no timer, works a treat .

[hikerbob (Bob)]

Some interesting input. I went PWM in the first place because it was easy with the Arduino and because that's what other dew controllers were doing.

Didn't really think about other options available driving the thing from a microprocessor.

The thermal inertia involved and relatively low power of the heating strips should make fast PWM unnecessary although massively longer heating cycles may alter what the temp sensors see (hot strap rather than the temp being applied to the surface below). As was suggested earlier turning the heater off while temp measurements are taken might be useful there when the measured temp is far enough above dew point.

Bob

[alistairsam]

Originally Posted by Bassnut

Just for example, I made a dew heater with a 08m2. The 8pin picaxe, a pot and a FET, just 3 components.
You tweak the pot and the heat changes.

8 lines of code in total. No PWM, no timer, works a treat .

Hi Fred

Would you have a circuit diagram for the fet and a part number
I was looking at pwm as that seemed to be the norm but if the frequency is low, just the high and low or pulsout with a variable pause in between is easy enough.
My objective is to keep heating to a minimum to conserve power when running on a car battery, hence bob's idea of calculating dew point and regulating heat.

[Bassnut (Fred)]

Originally Posted by alistairsam

Hi Fred

Would you have a circuit diagram for the fet and a part number
I was looking at pwm as that seemed to be the norm but if the frequency is low, just the high and low or pulsout with a variable pause in between is easy enough.
My objective is to keep heating to a minimum to conserve power when running on a car battery, hence bob's idea of calculating dew point and regulating heat.

I lied, 4 components .

The FET is a MTP3055V, good for 12 amps !!. But any digital N channel FET would do.

Dew point calculation is very clever to save power, would also need an ambient temp sensor. Could be easily handled by even the 08m2 and a bit of code.

Bobs article is a work of art though, a whole different ball game and very comprehensive.

[hikerbob (Bob)]

Fred thanks for your kind words. I'm somewhat proud of my heater (although I look forward to it being improved on or superseded).

"My objective is to keep heating to a minimum to conserve power when running on a car battery", when I'm not on main power I often run off one of the smaller SLA batteries used in burglar alarms etc. Minimising power usage goes up in importance.

I've been using a cheaper temp/humidity sensor in more recent play, I think I mentioned it elsewhere but just in case http://www.techtoys.com.hk/Components/SHT1x/SHT1x.htm

Somewhat cheaper than the SHT15 (http://www.sparkfun.com/products/8257), a bit less accuracy but for what we are doing I can't see that matters. There are other alternatives around for the humidity sensing but I've not seen anything quite as neat for the purpose as the SHT sensors.

Bob

[mswhin63 (Malcolm)]

Originally Posted by hikerbob

Fred thanks for your kind words. I'm somewhat proud of my heater (although I look forward to it being improved on or superseded).

"My objective is to keep heating to a minimum to conserve power when running on a car battery", when I'm not on main power I often run off one of the smaller SLA batteries used in burglar alarms etc. Minimising power usage goes up in importance.

I've been using a cheaper temp/humidity sensor in more recent play, I think I mentioned it elsewhere but just in case http://www.techtoys.com.hk/Components/SHT1x/SHT1x.htm

Somewhat cheaper than the SHT15 (http://www.sparkfun.com/products/8257), a bit less accuracy but for what we are doing I can't see that matters. There are other alternatives around for the humidity sensing but I've not seen anything quite as neat for the purpose as the SHT sensors.

Bob

Hi Bob,

Using a Temp Humidity Sensor is a better play as I was considering this for fully automatic dew heater. I feel there is no requirement for ASCOM compatibility as it should find the sweet spot for dehumidification, Unless someone can tell me otherwise, maybe something I am missing.

[hikerbob (Bob)]

Malcolm the only credible use I can think of for ASCOM (or similar) is the ability to tweak individual channels without making the panel on the controller more complex. Overall I like having the controller sitting there doing it's thing without the need for me to bother with it once it's turned on.

If you got really keen maybe the ability to use the temp sensor already on the scope to tweak focus as temperature changed but from the little I saw of that I suspect that it might create more problems than it solved.

I've also thought about logging humidity and temperature during a session but suspect that I wouldn't bother looking at the results so have not done anything about that.

Bob

[mswhin63 (Malcolm)]

Originally Posted by hikerbob

Malcolm the only credible use I can think of for ASCOM (or similar) is the ability to tweak individual channels without making the panel on the controller more complex. Overall I like having the controller sitting there doing it's thing without the need for me to bother with it once it's turned on.

If you got really keen maybe the ability to use the temp sensor already on the scope to tweak focus as temperature changed but from the little I saw of that I suspect that it might create more problems than it solved.

I've also thought about logging humidity and temperature during a session but suspect that I wouldn't bother looking at the results so have not done anything about that.

Bob

Very sound logic, I think we can overload ASCOM is bit too much sometime. I have always though some ASCOM drivers like Dome could be done with so much more simplicity instead by using discrete component or at most Arduino.

[alistairsam]

Originally Posted by Bassnut

I lied, 4 components .

The FET is a MTP3055V, good for 12 amps !!. But any digital N channel FET would do.

Hi,

I went to jaycar and they mentioned the MTP3055V is discontinued.
I got the P16NF06 instead. datasheet here.
http://www.st.com/internet/com/TECHN...CD00002501.pdf
the test circuit for resistive loads is attached, but should it be wired as per Fred's diagram or as per the test circuit?
I don't get what Vd is. Would the mcu output be connected to input of Rg?

As for the issue of temp sensors being under dew heater straps, what if we used a piece of stryofoam as a thermal insulator between the strap and the sensor so the sensor contacts only with the metal of the scope?
you'd need temp sensing to keep heating regulated and save power.

I also got some nichrome wire. Its this one http://jaycar.com.au/productView.asp...e&form=KEYWORD

I wanted to make a strap for a 2" eyepiece barrel for my Qhy8 and guidescope.

for a power of 2W, the current would be 180ma.
for 180ma, Resistance required would be 66ohms.
is this correct?
if so, the wire I got is 13ohm per metre, would I need to use 4metres to get close to 60ohms?
I might just get a dew strap but thought I'd try the nichrome as well.

should get a unit based on Bob's concept built soon. I love the idea of an automated and efficient controller that conserves power intelligently based on real time feedback
and going by Fred's comments, if pwm is not required we could easily build a 1 chip 28x2/20M2 controller with 4 or more channels just using pulsout as you can define the width of the pulse in the command and each channel can be independent based on its sensor.

[hikerbob (Bob)]

Alistair if it's any help my 240mm long dew not strap draws 0.29A at 12V.

The longest strap I've got (for a 12" tube) draws a bit over 1.2A.

Bob

[DancesWithWords]

Hello everyone. I've been thinking about building a dew controller of my own since I first saw the Iceman's. I designed this and my friend Sungjune did the build. Let me know what you think or any questions you may have.

Follow this link to technical project page: http://goo.gl/cEY0z

==
DancesWithWords

[rcheshire (Rowland)]

Reading an earlier post about temperature margins. It depends on how close you are to freezing. If the dew point temperature is 0 then dew point +5 is probably ideal. At 4C ice begins to form and the volume of dew droplets increases. In any case you want to avoid freezing.

At other temperatures dew point +2 is usually adequate. I found with DSLR cooling that a set point temperature of dew point +2 assured no dew formation.

There are some interesting aviation papers on ice prevention, which deals with dew formation.

My Arduino sketch just calculates dew point and adds 2. It would be easy enough to ramp that up automatically when nearing freezing - if you've ever spent an all night session at Snake Valley, you'll know what I mean.

The other thing I found with telephotos, is that dew formation inside the lens can catch you out. So I always run the dew heater if there is any chance of dew.

The native Arduino 490hz seems adequate for PWM. I run it direct to the FET gate through a 100R resistor. MOSFET is a NXP - PSMN1R1-30PL - MOSFET, N CH. You need a 750R pull down resistor from Gate - GND for top performance.

[DancesWithWords]

Originally Posted by rcheshire

Reading an earlier post about temperature margins. It depends on how close you are to freezing. If the dew point temperature is 0 then dew point +5 is probably ideal. At 4C ice begins to form and the volume of dew droplets increases. In any case you want to avoid freezing.

At other temperatures dew point +2 is usually adequate. I found with DSLR cooling that a set point temperature of dew point +2 assured no dew formation.

There are some interesting aviation papers on ice prevention, which deals with dew formation.

My Arduino sketch just calculates dew point and adds 2. It would be easy enough to ramp that up automatically when nearing freezing - if you've ever spent an all night session at Snake Valley, you'll know what I mean.

The other thing I found with telephotos, is that dew formation inside the lens can catch you out. So I always run the dew heater if there is any chance of dew.

The native Arduino 490hz seems adequate for PWM. I run it direct to the FET gate through a 100R resistor. MOSFET is a NXP - PSMN1R1-30PL - MOSFET, N CH. You need a 750R pull down resistor from Gate - GND for top performance.

I've forwarded your comments on to the builder as well. I've ran into what looks like my first bug and that is when the temp goes belong 0, the program spits out way too many digits after the zero.

Current tempatures are a bit extreme for out door testing as it is presently -32 with windchill and the night of course are even colder.

===
DancesWithWords