solar tracker circuit

[David HK]

An update.

I have improved the track layout on the PCB (shown in the above photograph) and now have all connecting wire points on the lower edge of the board.

The board is also set up in logical halves with VR 1 and its tell-tale LED on the left and VR 2 with its LED on the right.

99% of the tracks on the board are now 50 thousandth of an inch wide (no idea what that is in metric).

All parts on the PCB are now pin-through-the-hole type.

Next job is very careful checking of the track and component layout correctness.

Once this is done all the hard work is clear.

Considerable documentation also complete.

Whats the best way to provide this board with power from a 24 VDC system? An LM 7809 seems capable of taking the voltage but it's likely to become hot and need a good heat sink. Any comments or ideas?

Regards,

David in HK

[ruddycrazy]

Hi David,
            I noticed you haven't put up the updated schematic on this thread, if you prefer not to put it up would you mind pm'ing me a copy as I have a few ideas on setting it up for a 24 volt system by not even using a 780* linear reg. Also with the circuit boards you made doing 90 degree traces is a BIG no on with designing pcb's. I use sprint layout for designing my boards then just print the pcb image onto some glossy paper (inverted for the underside) then iron on the transfer. With a bit of playing around I reckon I could get the pcb to be single sided with minuim jumpers and makeup a test circuit here to test on my 24 volt array.

         As far as soldering multi pin smd chips might first foray was a 44 pin tqfp pic chip which works purfect. Also I have since done a few 64 pin tqfp's and the tip on my soldering iron ia 2mm one too......

Regards Bryan

[David HK]

Bryan,

Patience please. There is a bit of a story to this. I first saw the original circuit layout on the Internet about five years ago. I got in touch with the designer to clarify a few point regarding various parts. The designer then put out an updated circuit layout which contained some improvements and a change to various components.

This what resulted:-

Original drawing part R1 was 1K, on the revised drawing it became 4.7K.
Part R11 was 10K and became 4.7K. This slowed down the tracking process to avoid the irritation of motor whirring all the time.
In the revised drawing a manual override switch was added to motor the array East or West - most useful

In the original drawing VR 2 was a divider between the 9 volt rail and ground with a centre resistor tap off of 100K at R10. On the revised drawing The VR 2 was connected to two 15K resistors between the high and low rails respectively.

To cap all this I examined my working PCB this afternoon and I found I did not have 15K resistors in place but a 100K and a 1K instead. I therefore plan to fit 15 K resistors in place around VR 2 on Friday and see how it all works.

On the paperwork side I have drawn up the circuit using PCB Schematic and the track layout on PCB Express. The two can interface and circuit checking becomes less of a chore.

Yes I know right angle track line corners on PCB's are frowned on, but all I can do is work within the limits of PCB Express software.

Give me a couple of days to sort out the above problems and I should then be in a position to satisfy your request.

I have no experience in making PCB's using your method.

Regards,

David in Hong Kong

[SparWeb]

My boards are always cluttered perfboard, so I've nothing to criticize 

I've stumbled many times dealing with 24v dc conversion.  Stacking a 7815 and a 7809 together doesn't work, and a 7809 can't take 30v alone.

The solution I have found is a series of replacement VR's some with adjustable outputs that can churn out more than an amp.
At the time I came across them, I was very frustrated and got the 25W unit, to be SURE it wouldn't melt, when all I really needed was the 10W regulator.
I've tried my own tests and they work well - far better than 7809's.  Using an auto light bulb to load it, the voltage output was very well regulated between on and off I could barely see a differrence on the 3 decimal-place multimeter.  I cannot attest to longevity, resistance to extreme temperatures or vibration, but I actually look forward to giving them a try where I can.

Link    http://www.solarbotics.com/products/50220/resources/

...one step up on the webpage is a list of other voltage regulators.

[rossw]

I've stumbled many times dealing with 24v dc conversion.  Stacking a 7815 and a 7809 together doesn't work, and a 7809 can't take 30v alone.

Linear regulators are EVIL where more than a few volts drop are required.

I've used a bucketload of these (or very similar) - ebay item 250840046710.
For about $5, you can't make anything like it

Brief specs:
Dimension:                            43 (L) * 20.5 (W) * 14 (H) mm (with potentiometer)
Input voltage:                        4.5-40V
Output Voltage:                     1.5-35V (adjustable)
Output current:                     Rated current 2A, maximum 3A (Additional heat sink is required),
Conversion efficiency:            Up to 92% (O/P voltage higher, the higher the efficiency)
Switching Frequency:            150KHz
Module Properties:                Non-isolated step-down module (buck)
Short circuit protection:         Current limiting
Operating temperature:         -40 ℃ to +85 ℃ (output power 10W or less)
Full load temperature rise:     40 ℃
Load regulation:                    ± 0.5%
Voltage regulation:                ± 0.5%

[ruddycrazy]

Hi David,
            Last night I got about 3/4's of the circuitboard design done in sprint layout and I will be able to make it single sided with a board size of 100x100 mm. I am thinking of tweaking the board size to suit a standard cheap UB box. I'll do the changes you suggested and mate don't get the wrong idea I'm just doing this for fun and to show you just how good sprint layout can be. Say if guy's wanted to route the board with a cnc I can upload the gerber and drill files no worries.


Anyway heres a peak at what I did last night, this is a snap shot view with the xray turned on so one can see the traces.



Cheers Bryan

[David HK]

Bryan,

This is not my thread, nor yours, but it's attracted over 1,500 reads so it must be of interest to a lot of people;

I am British and have lived and worked in Hong Kong for 42 years and I guess you are US American or Canadian.

Thanks for taking an interest because I am sure that between us we can offer the forum a useful piece of technology.

Today, Friday, I played around with changes to resistors around VR 2 and found that they did not work so I reverted to what I had before – 5 years plus – and my system tracked the sun until the end of the day. However, I kept one new value resistor in the circuit and want to give it a few days trial before I decide to adopt it or revert back to my past five year circuit experience without a problem.

Between us we obviously have a difference in software preferences – I like Express PCB and you like something else. May I suggest that you down load Express PCB software so that I can send you my files containing the circuit diagram and the PCB layout. This would be an advance in the process, but how you could convert them to your requirements is beyond me.

So far so good.

Next step.

I want to get in touch with anyone on this forum with a view to sending them a complete package of hardware for trial without obligation. Perhaps two people.

I can help out with most of the hardware (99%) and can include cables if they are not excessive in length.

The catch in this is that the end user has to write a fair and unbiased report back to this forum.

Any takers?

Regards,

David in Hong Kong

[rossw]

I want to get in touch with anyone on this forum with a view to sending them a complete package of hardware for trial without obligation. Perhaps two people.
I can help out with most of the hardware (99%) and can include cables if they are not excessive in length.
The catch in this is that the end user has to write a fair and unbiased report back to this forum.
Any takers?

David, I've got 4 tracking arrays at the moment. They currently track based on time, only.
I am happy to convert one or two to use your board.
I also have a pyranometer and measure "available" power, and two fixed (seasonal only) arrays.
This would give some nice, direct "side-by-side" comparisons of all 3 techniques.

Just an offer.

RossW

[Volvo farmer]

I have two tracking arrays and one tracking panel for the well, all on Redrok trackers.  I'd gladly put one of these side by side with the Redrok and compare.  Comparison would be visual only as I'm not measuring watts separately out of the arrays.

I need something that will run on 24V nominal and not blow up when I equalize batteries at 32V though.

[oztules]

Well DavidHK

looks like you have two excellent testers lined up.

For a 24 or 36v circuit, I think the simplest is the best. This circuit looks to only need a few milliamps to run 99% of the time. I would use a 3W resistor of around 1k in the positive power input line to the main board, and use a 12v 1watt zener  (actually i would use 18v as I have a thousand of them....)to pull the higher voltage to a lower voltage that will not bother the rest of the circuit. The 7809 will take care of the rest after that.

The only time it will use a few more milliamps is when the relays are pulled in.... and I would use a different/separate  5w 500R resistor to drive the relay primary... or just use 24v coils in the relays... but a resistor does this fine..... ie run the relays on a separate line to the regulated 339 system

It is brutally simple, and has 20 years use in our battery charger boards we used to build ( in the good ole days)
Those boards were running 24/7, even when they switched off the load... in this trackers case, the relays are used for very short times, so the rest of the time only the 339 is running... and it is a  mostly turned off high impedance.

Whilst silly simple, it also protects from spikes that will terminate voltage regulators and buck chips. Suggest you try this first.

Here it is in a similar setup, you can see two big resistors, the 270R drives the black relay from 24v, the 470R comes down from the 24v to 12v (see tiny zener) and then goes through a 1205 regulator to a voltage divider and trimmer for a comparator.... very similar to your needs.... ie a stable voltage , and a "wild" input of a higher value than you need. and this circuit is also used up to 50v with different resistors of course.... but has stood the test of time.





Ruddycrazy as you can see on this board, I have no problems with right angles on low voltage DC boards... frequency and voltage changes this to avoid corners, but for this and the tracker will make zero difference.

Also losses in the resistor system will be small if you get the R correct for the A that the board uses, so the zener is only barely used.


Volvo "
I need something that will run on 24V nominal and not blow up when I equalize batteries at 32V though."
This is exactly what I'm talking about with using this system. One of these boards is used in my EV to control my mill. It is resistored for a 36v system, but regularly sees 48v from the mill, and has done so for years now.  (if it sees > than 45v for more than 5 mins, it shuts the mill down)....

observations from the side





...................oztules

[David HK]

Oztules,

Your a good man! Thank you for the previous post which I have implemented. The PCB allows the resistor and Zener to be accommodated and to change from 12 to 24 volts all that is required is a track to be cut and a jumper installed.

You are correct in assuming that relay coils will be driven from the system supply - either 12 or 24 volts.

Everything going well - documentation - hardware - PCB layout.

Regards,

David in Hong Kong

[David HK]

Still making good progress.

Am doing treble checks of the PCB schematic for 100% compliance before I order it - hopefully Wednesday or Thursday this week.

Much hardware has been assembled and is being put together.

Ross W and Volvo farmer you are very kind to offer services and I hope to take them up. When I have finished the kit assembly I shall make further contact with you. In the meantime could you work out where you would place the control box for the tracker and the sensor unit on the photo-voltaic panels and measure the distance plus a bit of slippage. I want to include wiring to make life and installation easy. Please PM me with details.

For readers who picked up on my comments about some satellite actuators not having the correct type of limit switches for this kind of operation I have attached a photograph that shows a limit switch found in some satellite dish actuators that is Normally Closed.

For use with this circuit limits switches in actuators possessing Common and Normally Open and Normally Closed will be essential albeit for one function - Normally open.

Another thread entry in a few days time.



Regards,

David in Hong Kong

[rossw]

Ross W and Volvo farmer you are very kind to offer services and I hope to take them up. When I have finished the kit assembly I shall make further contact with you. In the meantime could you work out where you would place the control box for the tracker and the sensor unit on the photo-voltaic panels and measure the distance plus a bit of slippage. I want to include wiring to make life and installation easy. Please PM me with details.

Sending shortly.

For readers who picked up on my comments about some satellite actuators not having the correct type of limit switches for this kind of operation I have attached a photograph that shows a limit switch found in some satellite dish actuators that is Normally Closed.

For use with this circuit limits switches in actuators possessing Common and Normally Open and Normally Closed will be essential albeit for one function - Normally open.

David, I looked again at the schematic, and have a couple of questions on the limit-switches.

If someone has only normally-closed switches, I believe it can still work quite reasonably, and could still be done with only 3 wires, if the NC switches went between the +12V rail and the  relay coils. They'd still be switching only modest current (relay coil) - which might actually be better for them - enough to actually whet the contacts.

The second part of my question is: my actuators already have two microswitches and the cams and diodes, to disconnect the motor in when it reaches full travel in one direction or the other. (The diode allows it to drive back the other way). Looking again at the circuit, I can see no operational problem with using the actuators internal "disconnect" feature and simply not installing the limit switches in the rest of the circuit, can you? I mean, apart from the relays operating even though the actuator won't be driving - hence additional current draw (but how much really?).

[oztules]

I'm not Dave, but yes and yes and about .5A for headlight style relays.




...............oztules

[rossw]

about .5A for headlight style relays.

The ones I was basing it on are 12V DPDT 5A rated, which have a coil resistance of 285 ohms - or 40mA @ 12V.
The larger 12V DPDT 10A cradle relays claim 180 ohm coil, which would be nearly 70mA @ 12V

500mA sounds like a LOT - would the little TO92 transistors even manage that long-term?
(I could live with 50mA odd waste, but 500mA is pushing the friendship!)

[David HK]

Ross,

Thanks for the note.

I too have an actuator you describe with built in diodes and its coupled to a small gearbox for winding my washing line up and down. Wife and daughter love it!

I have given this matter some thought and, like you, can see no reason why your description of a diode equipped actuator would not work in the circuit.

The above post was to illustrate that not all limit switches are the same, and anticipates a question someone may have wanted to ask in the future.

Regards,

Dave in Hong Kong

[oztules]

I seem to use overly large relays...
I have 50A headlight relays that read 35R, some 20A ones reading 92R and some 30A reading 62R. (and I can't find any 40A ones lying around.)

For heavy inductive loads..... I gravitate towards the heavy rather than the lighter.....particularly as this circuit does a lot of tiny start stops during a full cycle. The linear actuators I have state on the label 12vdc 80A max ( out of old floor scrubbers), which if starting under load will have an instantaneous current worthy of a decent relay.



That said, the 50A used in  13.7v (14v) systems (12v nominal) uses around 400ma, so a generalisation of .5A is perhaps heavy handed.... but what I would use as a guide to my loads, and really is still under done for loaded startup by the looks..... at least it is only 15A running, but gee it's mostly starting in this application.

I admit not everyone uses junk like I do, but I have 2 of them  in my greedy little hands.... and out here.... thats what counts.

The TO92 is probably ok in this instance, but I would use 2n4401 ( buy them (and the 4403's) by the hundred as I use them for totem pole drivers for fets) They max up to 600ma. (625mw max) The 3904 is not a starter for me for this as it is a 100ma unit. Good for your relays perhaps, not for auto ones. (they would not be happy driving even the 30A ones.) If I was to leave it turned on... I would perhaps go to a bigger package.

So I guess I might be completely off base whith what you use.... and obviously works for you.... but thats what I have to work with.

Yes I am just an agricultural animal .... its true....

I think this is where I say ... YMMV.....


...................oztules

[rossw]

I seem to use overly large relays...
I have 50A headlight relays that read 35R, some 20A ones reading 92R and some 30A reading 62R. (and I can't find any 40A ones lying around.)

For heavy inductive loads..... I gravitate towards the heavy rather than the lighter.....particularly as this circuit does a lot of tiny start stops during a full cycle. The linear actuators I have state on the label 12vdc 80A max ( out of old floor scrubbers), which if starting under load will have an instantaneous current worthy of a decent relay.

Fair enough too.... I guess we all use what we can get, or have.

The actuators I got are nominally 36V but run fine on 12 or 24, and take around 1.5A peak current, well under 1A for most of their travel even loaded with the array as previously photographed.

*IF* I were designing the system from scratch (I'm not), I'd use a low-side FET to actually switch the current on and off, and the DPDT relay purely as a reversing relay. This way, the contacts would never have to make or break the current flow and short movements "forward" would not require any relay drive at all.

[Volvo farmer]

My trackers also use satellite actuators with limit switches and built-in diodes. The Redrok trackers need only 4 wires. Power+ and - and motor + and - .  I have these 4 leads already coming to the corner of the array to the Redrok in a peanut butter jar.  If you are designing the sensor and the board to be mounted separately, I could hang the board out of the weather, under the solar panel, about 8" from the sensor.  I don't see myself needing more than 12" of wiring harness from there.

I can't really decipher how circuits work like Ross can, but if you can make it work with 4 leads and the existing limits/diodes found on all satellite actuators, it would be plug and play for anyone using a Redrok, and would make my retrofit really easy.

[David HK]

Volvo Farmer,

I am pretty sure all you will require is the Motor + and - wires to be connected to the control box. I fully understand what you are writing about with regard to diodes etc.

Some years back I did come across a horizon to horizon satellite dish which used extensive gearing, a 12 Volt DC motor, and completely separate limit switches which are common in many industrial applications.

With my package there will only be two units - the sensor, and the control box.  I am fairly sure you can set it up for use in less than one hour.

Still making good progress.


Regards,

David in Hong Kong

[David HK]

Interim progress

Two control boxes made. Each fitted with cable glands, toggle switch for manual activation of an actuator, main PCB with cable connector and supports for PCB control board, plus a stainless steel universal mounting strip.

Two sensor units made. Cable gland connectors fitted, plus PCB for holding the light dependent resistors (LDR's) . Also copper pots drilled with a 1.5mm light pinhole to sit over the LDR's.

PCB's received back from Express PCB.

One unit will be equipped for 12 VDC and the other for 24VDC.

Next step, go shopping for resistors and commence populating the boards. May be complete and ready for testing at the end of this week





Regards,

David in HK

[rossw]

Looking pretty slick, David.

[fabricator]

Looks like excellent progress to me.

[David HK]

Ross W and Oztules,

I have finished assembling the first 12VDC unit and sensor assembly and now need some decent sunny weather to test both items. Hong Kong has just been skirted by a tropical storm so the place is still soaking wet and affected by cloudy skies. Another tropical storm seems to be on the way so relentless sunshine for the next week or so remains doubtful.

Ross W i think a few posts back you mentioned the use of diode controlled actuators which would obviate the need for limit switches. Good idea.

However whilst assembling the main PCB yesterday (Thursday) I realised that the purpose of the Limit switch is to short to ground the outputs of the LM 339 - Pin 1 and Pin 10 to prevent the transistors firing the relay coils into the ON state. If a diode switched actuator is used and the system is without limits switches to perform the above function  then surely it will mean that, at the end of the day when the system is parked in the East, the relevant relay will remain switched ON all through the night until the following daylight reactivates the system.

That's about the best way I can explain it, so could you mull over the thought and let me know if I am up the creek without paddle and canoe.

Regards,

Dave

[rossw]

However whilst assembling the main PCB yesterday (Thursday) I realised that the purpose of the Limit switch is to short to ground the outputs of the LM 339 - Pin 1 and Pin 10 to prevent the transistors firing the relay coils into the ON state. If a diode switched actuator is used and the system is without limits switches to perform the above function  then surely it will mean that, at the end of the day when the system is parked in the East, the relevant relay will remain switched ON all through the night until the following daylight reactivates the system.

Yup, that's why in my post, I'd said that the downside would be that the relay would remain pulled in - and how much current did they take. I figured that if its under 50mA it'd be no great problem.

HOWEVER: I was just looking at the circuit again wondering about "fixes". I had some cunning plans, but they would all require fairly substantial changes to the circuit - not something I wanted to do.

Referring to http://www.phoenixnavigation.com/ptbc/images/suntracker1.jpg - how's this for a simple idea?
Lift the left hand end of R18 and R19 from the connection to IC1/14 and insert a large value cap in series, and a modest pull-down resistor.
When the output of IC1c goes high (fairly quickly, its being used as a comparator), the large cap will pass enough current to turn on Q2 and Q3.
After a while, it'll be charged enough that the transistors will no longer be able to keep the relay pulled in - making a crude timer. It only needs to run for maybe 2 minutes, depending on the actuators and voltage (mine take 90 seconds end-to-end).
It's crude, but it doesn't require significant changes to the boards.
It could also be done by just replacing R18 with the cap, and putting a small pulldown across C5. The LED would remain on then (but then, it does with the limit switch anyway).

Thoughts/comments?

[oztules]

That was the discussion Ross and I had last page...... and I think the get around suggested by Ross is elegant.

"It could also be done by just replacing R18 with the cap, and putting a small pulldown across C5. The LED would remain on then (but then, it does with the limit switch anyway)."




............oztules

[David HK]

Ross w and Oztules,

Thank you for the notes. I plan to stay with the existing circuit and identify it for use with actuators fitted with limit switches. I don't want to carve up the three existing  PCB's.

I would however like to develop/modify the circuit for use with actuators equipped with diodes for voltage control. In other words a new circuit with a different PCB layout. Could you please suggest values for the Cap and the pulldown - I will then PM you a copy of the circuit schematic so that you can check it for correctness (if you don't mind).

Having got past these two stages, move on again and see how the circuit would work with FET's.

In other words, create a family of options based on a common circuit plan.

Should keep me going until winter or summer for you.

Regards,


David in HK

[rossw]

Thank you for the notes. I plan to stay with the existing circuit and identify it for use with actuators fitted with limit switches. I don't want to carve up the three existing  PCB's.

David, it's not so much as case of actuators "fitted with limit switches" - mine are for example. Its more like "fitted with EXTERNAL limit switches".
I can add a couple of microswitches to my array to satisfy the switch issue if that's a concern - as long as the external switches activate before the internal limits of the actuator, it won't make any difference.

Whilst I had a number of thoughts for other ways to achieve the aim, I settled on the cap option specifically because it doesn't require any chopping up of your existing boards.
Pulling out a resistor and replacing it with a cap should be straightforward and doesn't require any surgery. Similarly, adding a resistor across the cap (or indeed, on thinking about it - replacing it with a resistor) won't require surgery.

I would however like to develop/modify the circuit for use with actuators equipped with diodes for voltage control. In other words a new circuit with a different PCB layout. Could you please suggest values for the Cap and the pulldown - I will then PM you a copy of the circuit schematic so that you can check it for correctness (if you don't mind).

If you're going to do a new board, there are other ways I'd consider far better than the capacitor I offered (for the above reasons of simplicity).

[David HK]

Ross,

Thanks. I am tending to think with a focused mind having run my own satellite television installation business in HK. In those days I came across a variety of actuators hence my interest.

Some shops in HK have an endless choice of limit switches for countless applications so the options for playing around with them for solar array applications is useful to know.

Will carry on tests when the sun decides to appear and contact you in a few days.

Dave in HK

[rossw]

at the end of the day when the system is parked in the East, the relevant relay will remain switched ON all through the night until the following daylight reactivates the system.

Just realised an easy fix in *MY* situation anyway....

I already have the rest of my arrays controlled from/by my home automation system.
I have spare relay channels in the battery room, it's a trivial change for me to use one more of them and simply turn off power to the tracker circuit say, 30 mins after sunset and back on 30 mins before sunrise.

I already do similar (30 mins after sunset I drive all the arrays half-way home to an overnight "park" position, then 30 mins before sunrise drive them fully east) - so killing power won't be a drama.

R.

[defed]

slightly off the topic, but what kind of wires are you guys running from the batts to the array?  i presume that most have the bank away from the array.

for my old c-band, the actuator power wires are only 16ga, with 22ga wires for the reed sensor.  i'd like to throw some cable in the trench before i bury the conduit for my power wires.

[rossw]

slightly off the topic, but what kind of wires are you guys running from the batts to the array?  i presume that most have the bank away from the array.

for my old c-band, the actuator power wires are only 16ga, with 22ga wires for the reed sensor.  i'd like to throw some cable in the trench before i bury the conduit for my power wires.

I have each array (6 of them) returning via their own cable runs.
4 arrays I used 4 sq mm cable (10 AWG), 2 arrays I've used 6 sq mm (8 AWG), but I'm also running at 100V.
So my "worst" run sees about 0.25 ohms over the 2 strands (out and back), for a loss of about 9 watts (98.6% or thereabouts transmission efficiency at full output, better as output drops)

[defed]

i always wondered how the c-band actuators worked w/ such small wire at long distances.  they are typically 24v or 36v.  i never checked the output at the receiver, maybe it is higher than 36v to compensate for voltage drop.  i suppose they just work slower at lower voltage.