Arduino Based Weather Station

[Gina]

For a start the Peltier TEC is not 20W at 5V it draws about 0.5A at 5v = 2.5W.  With this and the dew heater one of the IRLZ44Ns switched on at about 3.2v and it was totally off at 2v.  So a range of 1.2v goes from off to on.  3.2v is too close to RPi logic 1 to be reliable but a voltage divider to +5v or a resistor and diode to 5v would make a reliable circuit I reckon.

[Gina]

Not going to do that - I'm not going to input current into a GPIO pin even from the GPIO +5v line.  Been reading the book I have on interfacing the RPi and it seems advisable to use optocouplers so I've ordered 10 off EL817 Optocoupler from Amazon with Prime delivery Wednesday.  Best I can do with the dreaded bank holidays!  These can drive the MOSFETs.

[Gina]

Further closeup examination of the HAT has shown that several of the connections to the GPIO pins are broken.  The no.2 stepper driver will not work anyway.  The tops of the feed-through holes have come away.  It would seem that these HATs are unsuitable for modifications - if you don't get it right first time, throw it away!!   I think I have one or two multi-way cables that fit the GPIO pins so maybe I can separate the wires and solder them to a stripboard for I/O.

Edited December 30, 2017 by Gina

[Gina]

I think the HAT is usable, I've moved the connections on the second driver to good positions and cleared the area where the circuitry for the dew heater and camera cooler controls were located.

[Gina]

Dew heater control with optocoupler.

[MarkyD]

I was getting noise from the PWM powering my Dew Heaters.  I used these to Isolate the power side after the MOSFET drivers.  They are the square block on the pcb picture

50V 10A EMI Filters

[Gina]

I'm not using PWM, just ON/OFF control but when I used PWM with other imaging rigs I made up my own filters with a ferrite ring and several turns of wire plus a ceramic capacitor.

Edited December 31, 2017 by Gina

[Gina]

Took the HAT and inserted a new A4988 stepper motor driver then plugged the HAT onto the RPi and connected separate power supplies for the RPi and motor driver (13.8v) and commoned the Gnd connections.  Powered up and waited for everything to load then connected over LAN and re-compiled the Astroberry DIY driver for focuser.  Ran the indiserver command on the RPi followed by KStars/Ekos INDI startup on the client machine and everything worked   That includes the remote focussing   I do not yet have anything installed for the dew heater or camera cooler controls.

Edited December 31, 2017 by Gina

[Gina]

Now have the buck converter in use and single 13.8v supply from bench PSU.

[Stub Mandrel]

On 30/12/2017 at 14:10, Gina said:

Jeepers! Why all eth complications? You could replace all that with one logic-level FET.

[Gina]

I've tested some logic level power MOSFETs (IRLZ44N) and they needed +3.2v to turn on.  The RPi produces about this or slightly less - rejected on reliability grounds.

Edited January 1, 2018 by Gina

[Gina]

The datasheet for the IRLZ44N shows the transfer graph as Vgs being well under +3v for turning on with a few amps drain current but I have two samples and both show turning on fully at around +3.2v Vgs.  If you know of any that switch at a lower Vgs Neil, I'd like to know.  Obviously I would like to keep things as simple as possible.

[Stub Mandrel]

1 hour ago, Gina said:

I've tested some logic level power MOSFETs (IRLZ44N) and they needed +3.2v to turn on.  The RPi produces about this or slightly less - rejected on reliability grounds.

Sorry my brain assumed GPIO5 meant it was a 5V line...

There is another way using a voltage divider: 12V - [10K] - Gate - [1.5K] - GPIO

When GPIO is at 0V the gate is at 1.6V, with GPIO at 3V, the gate is at 4.2V.

[Gina]

Yes, I thought of that but the GPIO lines seem to be very sensitive to even very tiny currents pulling them up.  On power up the GPIO lines are floating and since the logic is CMOS it's very easily destroyed - as shown by my pile of dead RPies.  A 3.3v Zener diode could be used to prevent the line going above +3.3v but it's getting complicated again.  Seems the only really safe way of interfacing the GPIO lines with external circuits of higher voltage is to use opto-couplers.  It says this in the book I've got on interfacing the RPi.  Exploring Raspberry Pi: Interfacing to the Real World with Embedded Linux

[Stub Mandrel]

https://coolcomponents.co.uk/products/logic-level-converter-bi-directional?utm_medium=cpc&utm_source=googlepla&variant=45222911630&gclid=Cj0KCQiA1afSBRD2ARIsAEvBsNmtiBhYHF9tvN-U08cXi6OQS0KfxrKRVm0Fcf8jXgn6m8K2MNxD9zIaAqOdEALw_wcB

You might find these useful?

[Stub Mandrel]

Ah, this is what I was originally looking for. Note they say it gets confused by I2C = one wire.

https://coolcomponents.co.uk/products/8-channel-bi-directional-logic-level-converter-txb0108

[Gina]

The wind has finally destroyed my old weather station mast!

[JamesF]

18 minutes ago, Gina said:

The wind has finally destroyed my old weather station mast!

Oh dear!  It has been a bit fierce though, hasn't it?

We're having some work done in the cellar at the moment and have to empty it, to which ends I've bought a cheap prefab wooden shed that was delivered today.  Unloading that from the delivery truck was a bit more exciting than I anticipated

James

[Gina]

I can imagine!!

Doesn't matter much about the old mast as I'm going to replace it anyway - just means I no longer have a wind vane and anemometer to look at until I get the new ones up.

[Gina]

I've decided to erect just part of the mast for the weather station and ASC.  This will be much lower than the final arrangement but is something I can do in winter.  It's using the same mast as for the previous ASC version.  It's 2.5m high so rather low for the wind sensors but adequate for testing and setting things up.  Once I have everything working I can see about a bracket to take the larger aluminium tube that will form the lower part of the mast.  The smaller ASC mast (aluminium tube as well) will go in the top of the larger tube giving a height of around 6.5m.

[Gina]

3D printed a bracket for the new, larger diameter mast (5 metre high bottom section).  This is on the bed of my Titan printer with 300mm square print bed to give an idea of scale.  The height is required to space the mast away from the side of the observatory to clear the roof when it opens.  Two 10mm coach screws will go into the middle main upright (between scope room and warm room).  "Penny" washers will spread the load on the bracket.  The mast will go on the north side of the observatory where it will not obstruct the view for imaging.

 

Edited May 31, 2018 by Gina

[Gina]

I've removed the broken weather station mast and old vane and anemometer.

I have a 35mm x 5m aluminium pipe which I thought might be suitable to take the ASC and weather station wind instruments but having stood it up on end I think it's too whippy particularly with all that windage on the top.  So it's back to the 44mm diameter x 5m pipe.  I've been thinking about the height of the wind instruments and come to the conclusion that enormous height is not essential.  It's not as if this is going to be a Met Office site with a 10m high wind instrument standard.  All I need to know is approximate wind force and direction.

Currently, I have a combined ASC and wind instruments unit but I'm having second thoughts.  The wind instruments want height but having the ASC on a high mast, particularly with the extra windage of the wind units, means it's going to sway in the wind (and we get plenty of that at times) and mess with the image.  The main reason for the combination was to have the ASC above the wind instruments and avoid having them in the ASC view.  But I'm thinking that is less important than a steady mounting for the ASC.

Edited June 13, 2018 by Gina

[Gina]

Separating the weather station from the ASC means either a small or a large redesign of the wind measuring unit.  With a small change I could simply mount the unit on a new bracket from the side of the mast.  This will look decidedly odd I think.  The other idea is a redesign of the wind vane mechanism so that both sit atop the mast with the anemometer above the wind vane on the same axis, as I had on an earlier design.  Then again there's a common arrangement of a bar across the top of the mast with wind vane on one end and anemometer on the other.

[Gina]

I need to cut a new vane as the present one got broken - that's no problem (as long as I can get the old one out).  I think I shall try the 35mm diameter x 5m aluminium pipe as the mast.  If I find it's too whippy I can swap it for the 44mm diameter one.  There will be less windage with just the weather station stuff on top and some movement shouldn't affect the wind measurements. 

Think I'll just make a new bracket and put the present unit on the mast top. I could add a light measuring unit on the other side of the mast.  A photo-voltaic unit from an old garden light gives a volt or two which can be fed into an analogue input on the Arduino.  Full scale analogue 5v gives 1023 so about 200 counts per volt - crude but adequate I think.

[Gina]

Removed old bit of vane and cut a new one from 0.7mm aluminium sheet and fitted it.  This is actually a bit lighter than the piece of thin acrylic sheet I used before so moved it a couple of mm further out to balance.  The photo-voltaic cell I thought I had has disappeared from the old weather station so I'll have to look for it (or something else).

Edited June 14, 2018 by Gina