Robert Brown Arduino dew controller build, step by step

[wormix]

10 hours ago, fozzybear said:

 "mini ON-OFF 3pin SPDT" then it will be most likely the middle prong and one of the outer prongs depends on which way you throw the switch to "power on" use a volt meter to check which way round you want to place the switch (on the continuity setting) just read Roberts PDF

strange as in the PDF quoted SPST on page 25

Two mini-toggle switches (SPST ON-OFF) connected via a voltage divider network provide an option of overriding the power setting to each dew channel to maximum (100%) when activated. If dew is forming, a sudden boost to heat the telescope optics is available by using the override feature. Just remember to turn it off as it does not turn off by itself. The controller implements overriding the power to the dew straps and applying 100% to either channel. There are a number of ways to do this. The toggle switches (only for channel 1 and channel 2), when ON, overrides the power calculation and applies 100% power to the associated channel (this is referred to as Override Mode). When the toggle switch is OFF, power for that channel is set to 0%, and then on the next power calculation the correct power will be determined for that channel (this is referred to as Normal Mode). The override switches could be omitted if the intention is to use the dew controller via computer, as the Windows application provides this feature (as well as the ability to disable the override feature after a time elapsed period at 100% maximum of 2 minutes).

 

It did seem a bit weird - I have some spst already so was going to utilise them as couldn’t fathom why you would need on-off-on. The only thing I could think off was it’s a typo and they should be momentary, but then I believe the override is either on or off and not timed.
 

Will go ahead with what I have and see if it works

 

cheers 

[brown_rb]

The switches are ON OFF.

You can use the 3p ones by using the center pin and one of the outside buttons as already pointed out. They are not momentary switches. The 3p ones use are ON OFF. You can use 2p on-off ones if you prefer.

You really don't need them though as the software can do that already.

The confusion is more about the fact that they are often mislabeled on ebay or aliexpress.

[wormix]

Getting there now - forgot how much I enjoy soldering, very good for the soul. 
 

Special mention to @fozzybear who kindly sent me the resistors which saved me having to order in a bunch that I would never use again. 

[fozzybear]

6 hours ago, wormix said:

Getting there now - forgot how much I enjoy soldering, very good for the soul. 
 

Special mention to @fozzybear who kindly sent me the resistors which saved me having to order in a bunch that I would never use again. 

coming along nicely. Glad to have helped you out.

[brown_rb]

Looking good.

Something that is not clear in the PDF. J7 is where you want to solder the ds18b20 transistor [without long leads and metal case - just a transistor type package all by itself] - the reason why it mounted on the board is to measure the pcb temperature and thus control the fan by turning the fan ON to cool the inside board temperature down if it gets too hot.

I see you have already soldered J7 headers in place, I would suggest soldering the ds18b20 direct to these header pins.

 

Cheers

Robert

[wormix]

44 minutes ago, brown_rb said:

Looking good.

Something that is not clear in the PDF. J7 is where you want to solder the ds18b20 transistor [without long leads and metal case - just a transistor type package all by itself] - the reason why it mounted on the board is to measure the pcb temperature and thus control the fan by turning the fan ON to cool the inside board temperature down if it gets too hot.

I see you have already soldered J7 headers in place, I would suggest soldering the ds18b20 direct to these header pins.

 

Cheers

Robert

Thanks Robert - the other issue with the PDF was the arrangement of the capacitors - I have copied what James did earlier by studying his pictures - the pdf doesn’t differentiate which capacitor goes where

 

cheers

 

 

[fozzybear]

6 minutes ago, wormix said:

Thanks Robert - the other issue with the PDF was the arrangement of the capacitors - I have copied what James did earlier by studying his pictures - the pdf doesn’t differentiate which capacitor goes where

 

cheers

 

 

C1 0.1uF1

C2 0.33uF1

[wormix]

We had a bit of a chat around diodes in a separate thread last month so to keep everything in one place (as this is a really useful thread for someone putting this together) I’ll stick a link below.

 

 

[JamesF]

2 hours ago, brown_rb said:

I see you have already soldered J7 headers in place, I would suggest soldering the ds18b20 direct to these header pins

I crimped mine into a 3-pin female header that I then slipped onto the pins to make it easier to replace should it fail.

James

[wormix]

3 hours ago, fozzybear said:

C1 0.1uF1

C2 0.33uF1

Just checked - that is how I have soldered it up as - thanks for confirming. 

[fozzybear]

@wormix how's the build coming along? You nearly there yet?

[wormix]

6 hours ago, fozzybear said:

@wormix how's the build coming along? You nearly there yet?

Currently on hold till I get the enclosure printed so I can be sure of length of wires, etc. Will start that this weekend but will need to relocate my printer as it’s currently in the shed and the cold / drafts are creating pretty severe warping. 
 

No huge rush now that some %{€€~*{! <impolite word> stole my camera whilst it was on its way to be modified 

[AlexK]

Awesome project! Thank you for sharing the progress in such great details!

I just had a thought to share about the possibility of unifying and simplifying it by splitting everything to multiple self-contained functional components (vs integrating everything in a single box) by utilizing a bunch of ESP32s.
I.o.w. each probe and each heater would have their own controller box attached communicating with the CPU over BT or WiFi.
You can think of it as of "Internet of Astronomy Things" 

That would in a snap

1. reduce the need of wiring to just the common powerline (OTOH you can easily opt for probes being completely wireless, as ESP32 can do just fine on a single CR18650 or even less for weeks);
2. remove a lot of connectors, possibly even the custom PCB, and allow to move related components closer together reducing the power consumption and heat buildup (and heat loss if you move mosfet with its sink directly into the heater module)
3. make the system modular (thus extensible, maintainable in the field, tunable, and upgradeable);
4. and could allow replacing the primitive Arduino CPU with a common smartphone with much more robust and easy to tune on the go software with more convenient user interface possibilities.

What do you think? Doable?

Edited November 5, 2020 by AlexK

[brown_rb]

The dew controller supports Bluetooth so no need for USB cable in this mode.

And yes you can use a phone with the DewController BT app for android phones.

 

Using separate ESP32's

You would need 12V with mosfets to driver the output. So you end up with little esp32

- 12V power needed - 3A capable wire needed

- dew strap output - 2p RCA socket and wires needed

- temp sensor input - 3p lead required

So you wouldnt actuall save any cables or wires in my option.

The dew controller as is now using mosfets really does not get very hot.

 

Edited November 7, 2020 by brown_rb

[AlexK]

14 hours ago, brown_rb said:

The dew controller supports Bluetooth so no need for USB cable in this mode.

And yes you can use a phone with the DewController BT app for android phones.

Using separate ESP32's

You would need 12V with mosfets to driver the output. So you end up with little esp32

- 12V power needed - 3A capable wire needed

- dew strap output - 2p RCA socket and wires needed

- temp sensor input - 3p lead required

So you wouldnt actuall save any cables or wires in my option.

The dew controller as is now using mosfets really does not get very hot.

Gotcha. Still, I can see benefits of independent components control vs all eggs in one box.

An advice for the app: Make it WebView+CSS+JS driven. That way you could port the UI/API to almost anything (even to the ESP32 and make it nearly headless) and would allow users to customize it to their liking (in CSS, or you can help them with that). E.g. the "Night mode" UI switcher can be implemented in 15 minutes.

Following!

Edited November 8, 2020 by AlexK

[fozzybear]

dismantled my stripboard version I was not happy with it and ordered some boards from JLCPCB well 10 to be exact all for the price of 22 euros, Ordered 1st Nov and arrived today from China fast delivery. Now to get this little project under way nice looking board. Only problem is with the diodes with the larger dia legs.

[wormix]

That’s a great price - when I looked originally I think I was looking at around £30 for 3!

[fozzybear]

Just now, wormix said:

That’s a great price - when I looked originally I think I was looking at around £30 for 3!

yep I looked on aisler and yes 3 for 30 quid/euros delivered. so from this company in china I was amazed at the difference from ordering 5 to 10 was only a couple of euros extra though the delivery is via DHL and that ate up most of the total price but overall could no grumble can afford to make a couple of mistakes...  

[Len1257]

26 minutes ago, fozzybear said:

 Only problem is with the diodes with the larger dia legs.

 

I just drilled the holes correct size. Where the track had been thinned too much I scraped the track resist locally to reveal the track copper and added suitable copper wire along the track, around the diode leg and along to the next piece of track. After soldering along the wire, as good as original.

[fozzybear]

@Jonk what diodes did you use on your build? they are a lot smaller than mine 1N5822

[brown_rb]

10 hours ago, fozzybear said:

dismantled my stripboard version I was not happy with it and ordered some boards from JLCPCB well 10 to be exact all for the price of 22 euros, Ordered 1st Nov and arrived today from China fast delivery. Now to get this little project under way nice looking board. Only problem is with the diodes with the larger dia legs.

Whatever you do, do not drill the PCB. This is a double sided board with plated through holes - drilling is a no no.

The 10A10 is a biggie - 10A rated, I suggest filing down the leads with a small file to get to fit or first solder a 2p header in place then solder the 10A10 to the 2p header. Do not drill the PCB.

The 1N5822 are schottky fast acting diodes. These help shunt the back emf produced during switching off the dew strap. Long leads act as inductors and try to maintain current flow when connected - so are very important. It is true that the MOSFET's do have this protection on chip but the 1N8552 IMHO are still needed. If they do not fit they will after a small amount of filing the leads

 

 

 

Edited November 13, 2020 by brown_rb
spelling

[MrFreeze]

Drilling is only a no-no for multi layer boards (3 or more track layers) due to the risk of shorting the internal layers. For a simple double sided board it's perfectly OK to drill it as long as you solder BOTH sides of the board (you may have to remove some of the solder resist to do this).

David

[skybadger]

On 05/11/2020 at 08:12, AlexK said:

Awesome project! Thank you for sharing the progress in such great details!

I just had a thought to share about the possibility of unifying and simplifying it by splitting everything to multiple self-contained functional components (vs integrating everything in a single box) by utilizing a bunch of ESP32s.
I.o.w. each probe and each heater would have their own controller box attached communicating with the CPU over BT or WiFi.
You can think of it as of "Internet of Astronomy Things" 

That would in a snap

1. reduce the need of wiring to just the common powerline (OTOH you can easily opt for probes being completely wireless, as ESP32 can do just fine on a single CR18650 or even less for weeks);
2. remove a lot of connectors, possibly even the custom PCB, and allow to move related components closer together reducing the power consumption and heat buildup (and heat loss if you move mosfet with its sink directly into the heater module)
3. make the system modular (thus extensible, maintainable in the field, tunable, and upgradeable);
4. and could allow replacing the primitive Arduino CPU with a common smartphone with much more robust and easy to tune on the go software with more convenient user interface possibilities.

What do you think? Doable?

Sounds like mine. Using esp8266-01 wireless and node-red. 

[fozzybear]

9 hours ago, brown_rb said:

Whatever you do, do not drill the PCB. This is a double sided board with plated through holes - drilling is a no no.

The 10A10 is a biggie - 10A rated, I suggest filing down the leads with a small file to get to fit or first solder a 2p header in place then solder the 10A10 to the 2p header. Do not drill the PCB.

The 1N5822 are schottky fast acting diodes. These help shunt the back emf produced during switching off the dew strap. Long leads act as inductors and try to maintain current flow when connected - so are very important. It is true that the MOSFET's do have this protection on chip but the 1N8552 IMHO are still needed. If they do not fit they will after a small amount of filing the leads

 

 

 

Many thanks Robert for the heads up was going to get my Dremel and perform some dentistry...

[fozzybear]

Well I thought I would try and file down one of those Diodes what a faff there must be an easier method maybe a wire stripping tool would do it, but that is with my son at the moment? I thought life was complicated enough back in the day when wiring like this on some old IBM machine upgrades.