Captains Log, Stardate Dot Dot Dot

[Madscientist267]

I've decided this board is the perfect place to keep track of what I'm doing, not just for individual projects, but in the long term.

Started this thread to keep that spirit alive.

My system is only 60W at this point, of which I regularly only see 45W of. Ok, so what. It's a start, and it's absolutely amazing what can be done with such a small amount of power.

It's all in how (and when) it's used.

Comments are of course welcome, but keep in mind this is where I'm keeping track of progress, so I may or may not agree with the replies.

And no, I'm not really a Star Trek fan, so the reference is for the eye-catch value only. It works, got you reading didn't it? 

All that being said, enjoy. I hope I provide some insight on the very basics required to get a system going. I'm not off grid by any stretch of the imagination, but I believe that anything worthwhile is worth thinking about, and this is no exception. It's a slow process, and my ultimate goal is to be at least carbon-neutral by the time I cease to consume oxygen for purposes of conversion into carbon dioxide.

Let the fun begin.

Steve

[Madscientist267]

With ~150PW (that's PetaWatts) of power reaching the earth from the sun at any given moment, I'm entitled to my share. As are any of you.

Astronomical number. Almost unimaginable. Even taking into account efficiency, this is enough for a 100 square mile area of the state of Nevada to power the entire united states.

Run Forrest, run!



In this diagram, the large Furby is the sun, I am the 'Millennium' furby, and the grid? That's the pesky little peckerneck that resembles Scrat. He's my hero under normal circumstances, which explains why I have him to start with, but squirrels trying to get a nut are not part of my master plan. I'm the squirrel in that case.

There's no reason we can't get away from the oil. I'm not a tree hugger; I'd rather strangle one than piss on one if they were on fire, but that's only because of what they've become. The truth is, the end of our greedy supply will eventually come to an end. And when (and as) it happens, energy will be so astronomically expensive, we might not be able to dig ourselves out of the hole. Do you want to be one of those people?

I don't.

Steve

[Madscientist267]

Ran some tests today to really get some hardcore info on the system.

Dug out every meter I have, and placed them in the system to gather a typical full sun day's collection of data. Days like today are rare around here (especially in April!); there wasn't a cloud in the sky all day. At all.  

The results coming soon, after I have the various genres of 'typical' days written down.

Today was ideal for getting an idea of the overall efficiency of the system, and as a result I found some hard evidence pertaining to a few pretty significant issues with the layout that I suspected were a problem.

About mid-day, I switched over to parallel (with the buck bypassed), and to my surprise, saw a very minor increase in current. But it's close to a wash; what's available at the panels and that which is available after the run to the house, are two very different animals when the panels are in parallel. I attribute this to the higher current (and therefore drop) in the lead-in cable. So what's lost in the buck, is roughly equivalent to what is lost in the cable at higher current.

I didn't do any Isc tests today; this will come as soon as the next clear day comes. Should paint a better picture in terms of where the losses are.

My rough actual calculated efficiency of the buck today averaged out at approximately 85%, although I didn't do nearly as much 'dialing in' as I usually do. I was primarily interested in seeing what the panels were doing. Normally throughout the day, I tweak it here and there, compensating for temperature and light conditions, and I can gain significant improvements with small adjustments.

I built two new boxes today...

The first is a parallel/series select box (at the panels) with an ammeter port to make life easier. When the meter isn't there, I have a shunt to plug in that simply jumps it out.

The other is a multi-function box that has no active electronics in it, it's just a switchboard to enable quick configuration changes without jumping through a bunch of hoops. It allows bypass of the buck converter, has a Schottky diode (which can also be bypassed with the flick of a switch), and a third switch that kills power going into the buck converter. The idea is to be able to quickly and easily dial in the highest throughput configuration with the least hassle. I used all switches in the design to avoid any parasitics (such as relay coils), and to keep reliability as high as possible. This box also has an ammeter port, for monitoring the current as the battery sees it. Again, complete with jumper for those 'I need that meter for something else' moments.

The pics and schematics are coming soon...

Tomorrow seemingly holds promise for patchy skies, and there's a particular cloud pattern that I'm looking for to get a lot of repetitive edge-effect action so I can compare notes with the system in various configurations.

I will also be shortening as many of the cables as possible, particularly those involving the buck converter. There are parasitic losses associated with the wiring that I cannot otherwise seem to resolve. Inductance is a biggie. I am designing a set of two ultra-low ESR capacitor bank boxes, one for the input, and one for the output at the buck. My experiments with heavy capacitance so far have shown some promise, but without shortening the leads going to the buck itself, the results have been less than uber-exciting.

So, that's today... More tomorrow.

Steve

[Madscientist267]

(Still) can't seem to upload pictures... Will try again with the updates later...

Steve

[Madscientist267]

EDIT - Shrug... not sure what the deal was, but appears to have been an issue with the machine I was trying to upload with. Will have to look into that, but in the mean time, here's yesterday's update:

Today brings not much new in terms of data collection, as the sky was rather similar to what it was yesterday, a general haze.

But, I have the pics and schematics for the new additions to the system.

Before I get to that, my wife said that I should probably explain how one of the boxes got it's name. It's a play on just what it sounds like.

Not being electronically savvy (she's a computer admin - they don't necessarily line up), she thought this whole time (honestly) that I was calling the buck converter, a butt converter. Who knows why or how, other than she's heard me mention butt splices and what not, maybe that was the connection... ?

So when trying to figure out a name for it, I jokingly mentioned that it was a "buck buddy", and it kinda stuck. So here it is, in all it's twisted glamour:


It kinda serves as the 'master' for the solar power routing. The panel attaches at bottom right, the battery connects at bottom left. Center right is input coming from the buck, center left is output to the buck, and upper left is the ammeter port. There's also a jumper (not in the pic) for when you don't care what a meter has to say (haven't figured out if I'll ever use it or not LOL). The 3 switches, left to right are Buck Disconnect, Buck Bypass, and Diode Bypass. Also in the picture is the inverter's remote receiver at the far right.



The guts inside the box. Nothing really super extravagant here; it's completely passive.



The schematic. Again, simple really. Just makes life easy.


But unfortunately, without the new box at the panels, the buck buddy is kinda pointless:


The series/parallel selector box mounted to the back of the panels. The jumper is plugged into the local ammeter port for easy monitoring if the panels are used by themselves, or it also comes in handy for finding the butter zone in an overcast sky. With it connected to the rest of the system in the house, it really doesn't serve much purpose, so it stays jumped out most of the time. The switch is a center off DPDT; Up is series, down is parallel.



The selector box schematic. The ammeter port wasn't drawn in this version, but just like the buck buddy, it is the last thing on the positive rail as the cable leaves the box.


As mentioned above, I recently made a mod to the inverter so that it could be controlled remotely. The basis for the remote is a chopped up "handy switch" (yep, 'as seen on TV'). I have two of these things, and have used them for a lot of different things, but finally gave in and chopped one of the receivers up to make it do a very specific job; allow me to kill the inverter from anywhere in the house:


The modification within the inverter is very simple - a PN2907 PNP transistor, a 470R resistor, and a 'stereo' 1/8" phone jack. The tranny jumps the existing power switch, turning the inverter on when it's base sees a ground via the 470R resistor.



The modified receiver is significantly simpler than the OEM handy switch design, but utilizes almost everything originally on the board (except for what was removed to allow the board to fit in the tiny box). The only changes I had to make to get basic functionality were to add a 1K resistor and change another to a 470R. The LED is a bicolor, and since the output of the receiver chip latches to either high or low, all I needed were two 390R resistors to put it into play. Green indicates on, red indicates off. The 390R were selected to minimize parasitic draw through the LED (they are connected to the 5V rail). There is a built in TO-92 7805 as part of the regulator, I simply removed the original preregulator (a zener that ran hot) and connected the 7805 straight to the 12V rail instead.

I don't have any pics of the inside of the receiver or inverter ATM, but may take some (along with a basic schematic) if anyone is interested in doing this simple mod to their cheap MSW inverter. Saves some watt-hours at night, as well as the legwork (which usually meant just leaving the inverter on with no load, at ~250mA).


The last picture is of the completed 'power zener' dumping regulator for the boom box's external battery:


I put it in a case and drilled a boatload of holes in one end to allow air in for cooling. It also has a few other components that the originally listed version did not have. It isolates from both the main battery, as well as another diode inserted strictly with the idea of exploiting the voltage drop across it. This was to reduce heat by increasing the difference between the radio's transformer output tendency and the clamping voltage of the regulator. Works fairly well, although it charges a little slower now. No big deal. The radio is not the primary charging source anyway. I'll probably post the updated schematic sometime soon.

That's it for now, been busy. 'Till next time...

Steve

[Madscientist267]

Today is showing interesting results.

I was starting to think that the buck converter really wasn't all it was cracked up to be, but this turns out to not be the case.

The temperature is up today, and the panels are struggling to keep cool, and the output is falling off somewhat.

I've been running in parallel for the last few days to get an idea of what I can regularly expect out of them in that configuration, but I was only sustaining about 2.5A at 14V. 35 watts. Come on, really? In full sun?

So I switched over to series, and engaged the buck, and gained an extra 100mA out of the deal. Up until this point, I have been seeing better output with the panels in parallel, but the heat is taking it's toll.

Currently, as I type this, the output is 2.616A @ 14.16V, a hair over 37W. Not much of a gain, but it is better.

A quick spray down (to get rid of pollen as well as provide some transient cooling) got me 2.735A @ 14.2V, almost 38.9W. Parallel is getting 2.780A @ 14.3V (39.75W) with the same treatment.

EDIT - Turns out if I am patient enough, and willing to stand there for a few minutes with the water bottle, I can make the difference between 2.4A and 2.8A, roughly a 15% increase in output.

Clearly, temperature is a factor here. I have a hairbrain scheme to lower the panel temps a bit further on a more permanent level, but haven't decided if it's worth all the work yet. It involves cutting up old soda cans to make heatsinks to attach to the back sides of the panels.

For posterity, according to weather.com, the temp here currently is 80F, 60% RH. Full insolation.

So, for now, when the heat is on, I'd say it's marginally better to use a buck converter to be able to dial in that sweet spot. Only means a couple watts for a small system, but could mean a few hundred for a large one.

Steve

[DanG]

Ah yes - someone believed the label ratings!  

60w with +/- 5% tolerances so 57w panels when new...

Actually leaving them out in the sun to overheat takes another 5% off for aging so you've now got 54~ w potential there...

And, umnn..  Temperature coefficient -0.53%/C (worst case) for typical (18% 24%) drop for a hot day with (60C  75C) panel temperatures..

Leaves a person with with (44.2w 41w) output before wire and connection losses; so letsee...

Six feet of 16awg (you may have larger but this takes in crimp and terminal losses) at 12V and 3 amps is 1.25% voltage drop

So there is another (.35 .40) watt loss for (43.85w 40.6w) total.

How accurate is your meter?

EDIt: I forgot the diode voltage drop...

[Madscientist267]

Ah yes - someone believed the label ratings!

I know right? It's too easy to get sucked in by that number...

Fortunately, I'm doing all the hard learning while it's still (reasonably) cheap.

I figure by the time it's said and done, for my location, I will only be able to count on a yearly average of sustainable production breaking a whopping 20% of the installation's value. Raw real estate on my roof says 12.5kW, so I'd be working with 2.5kW when it comes down to it. IF I go that far, which I'd like to. Sigh...

Funny thing is, I always figured that the panels would be the most expensive part of the system, and even though the ones I currently have were freebies, it's still cheaper to replace them than to replace the other equipment I've had to make by the time all the incidentals are included. Maybe I should just give in and buy more panels instead... but I'm so damned inquisitive! LOL

I've run the juice through all the meters, they agree within about 100mA of each other. The Radio Shack meter is the one in the middle, so I stick with that for the daily current measurements. But no, I haven't had it calibrated or anything...

The diode (there's only one right now) can be manually bypassed. I gain about 300mA in full sunlight by doing so, not as much in less light. It's a Schottky.

Steve

[Madscientist267]

A slightly off-RE related post (although it kinda ties in, since it was for use with the system)...

There's this little device called the "AirWick iMotion" (why everything has an 'i' in front of it in the 21st is beyond me).

I picked two of them up for $5 on clearance at my local pharmacy. Figured, why not.



I figured I'd probably see some ridiculously cheap implementation of a motion sensor inside, such as a CdS cell like in the halloween toys, but it turned out it was in fact an actual PIR sensor:



I went as far as to hack the board as far as I could, and got all the test points figured out:



... only to find out that there is no hacking the MDT10P55B1S; it's an EEPROM based microcontroller. I have no way of knowing exactly how to manipulate the innards. One version of the chip has a master reset pin, this one 'technically' doesn't, although I found that the same pin serves a similar purpose. It's labelled as '/RESET' in the pic above, but that isn't entirely accurate, as I later found out.

The premise of OEM operation is something like the following:

- Power on
- Automatic power on 'boost' fragrance (turns on the extra 'heater', which isn't anything more than a 2W 8.2K resistor) for 10 minutes
- Cool back down to 'normal' temperature (only using the power supply 'heater', a 2W 4.7K resistor) for another 10 minutes
- Enable motion sensing
- When motion is detected, 'boost' for 10 minutes. Additional motion during the boost does not extend the time
- After boost, ignore PIR input for another 10 minutes
- GOTO 'Enable motion sensing'

... lather rinse repeat ...

So, I clearly needed to modify it's operation if I were going to use it as a motion detector for a light (which is plugged into the inverter, and how it relates to RE )...

I had already done the full wave mods to the power supply, as well as placed the 'heaters' in series to reduce idle power draw, since I had found that the board doesn't need as much juice as the original design had flowing through the primary resistor. I had to swap out the OEM half wave rectification (which then goes through an SCR) because the LED light I was testing with did not like the half wave and was flickering viciously.

Ok, so I had everything hacked up (in terms of connections) and realized I was not going to be able to manipulate the program on the CPU except by one method; a hard reset.

I found the way to do this fairly readily, and it only involved a single NPN. Emitter to ground, collector to +5V(A), and the base to the PIR trigger.

And it worked.

Kinda.

Problem was, there are two independent power 'supplies' on the PIR board, by virtue of a cute little trick with (Schottky?) diodes. The main regulating zener (on the heater board) is only 3.3V. I presume this was to minimize the power dissipation (and therefore cost) as much as possible. The two 5V rails are derived from this using some kind of overshoot exploit, by passing the sharply transitioning 3.3V through a set of two diodes, each pointed to their own filter cap (visible in the pic above). The result is a dual +5V supply capable of 7mA Isc at the caps.

Since the filter caps are so small, shorting one has a very similar effect (with a very short delay thanks to the isolation) on the other. One supply (A) runs the CPU chip, the other (B) runs the PIR and it's comparator (LM324).

I found out fairly early on that I could reset the CPU with a pretty quick shorting of the 'A' supply, causing the light to come back on (almost) immediately. There's a delay of about 1 second or so, presumably due to the CPU 'booting up'.

The result of this is that resetting the CPU when the light is on will cause it to go out briefly (while it reboots), and then come back on. I even had a way around this, by placing a small cap across the SCR gate and ground to hold the gate for that brief period.

So it was all set. Or so I thought. Remember that 'very similar effect' part? This causes the 5V 'B' rail to destabilize, which in turn causes the 324 to falsely trigger, again 'resetting' the power.

The end result is that the thing sits there and flickers fairly randomly every second or so.

To alleviate this, I tried to add capacitance to the 'B' supply, but never really got it stable enough to use it.

I threw the towel in when one of my tacked on resistors hit one side of the line on the 'heater' board, and the thing dropped dead. I think PART of the 324 might still be working, but can't really tell. The LED is stuck on solid now, and it doesn't do squat.

SCR still works... 

So why bother telling you all this? Because at 7mA Isc at either 5V rail, I viewed the board as 'indestructible' and it was -- until I temporarily boosted that up to several hundred (or more) mA!

I was pissed. You know how long it took me to get those boards out? They don't use screws on this thing, and the plastic is some kind of rubbery something...

I have one left, and now that I know where the components are, I can probably open it quicker. But it probably isn't worth the time to try it. Dammit.

Steve
 

[Madscientist267]

Today led to some progress, but started with a little head scratching.

The whole 'ideal diode' thing.

Makes sense, throw a MOSFET in series with the panel and turn it on when there's light, off when there's not.

Only one little problem with that.

A MOSFET has an integral diode that's just naturally there as part of the construction of the transistor. Problem is, it is forward biased when the panels are dark!  

Or is it...  

Turns out the secret to using a MOSFET as an ideal diode is to flip it around backwards, yet still drive it forwards. Interesting concept, wasn't sure if that was a possibility or not, 'till I finally found a schematic for the typical ideal diode design. There it was in black and white:


(snatched this from some site somewhere, I hope nobody minds... heh)

But forget all that. All I wanted to do was understand how the MOSFET connects between the panel and battery, as I had seen it done before but couldn't figure out how they got around the pesky diode problem. Now I know...  

Since I had no interest (nor need for that matter) in putting so many parts together to accomplish this simple task, I found another way that exploited another part of the design for the (eventual) tracker that will be added to my layout:



It uses the aux panel to turn the 'diode' on, sending it into full conduction. A MOSFET conducts just as well backward as it does forward (but the gate signal must remain positive) and has the added benefit of using the integral diode in absence of a gate signal when wired this way. This diode still happily conducts, albeit at a slight loss (0.7V). When turned on, the drop is between 0.022 and 0.05V, with no sign of any heat when touching the heatsink-less transistor. I'm using a SUB85N03 (N-channel), but any decent power MOSFET (such as an IRFZ44N) will do just fine. The voltage divider is to keep the gate voltage well away from breakdown. About 12V reaches the gate in full sunlight with my configuration; breakdown is 20V (for the two trannies mentioned), and they both have thresholds in the 2-4V range. Either one will completely turn on.

The 1N4001 and 1uF combo act as a 'just-in-case' to keep spikes and other unwanted irritants away from the gate (primarily thinking about the switching action of the dump controller).

The dump controller runs the two fans that cool the main panel when the tracker battery is full, providing (as it turns out, marginal) cooling for the panels, and acting as a sufficient dump load to keep the battery from overcharging.

Steve

[DamonHD]

Interesting stuff, and yes, I'm annoyed by throwing away at least 0.4V for each blocking diode and would like to do the 'perfect' version especially in winter when losses hurt worse.  I'd like a simple solution that doesn't require an aux source.  I have P and N MOSFETs to hand as it happens...

(BTW, I think that you should in consideration of the 'all rights reserved' notice on the first image convert it to a link to the owner's page, and not either embedded here not a hotlink to the owner's image.)

Rgds

Damon

[dave ames]

Comments are of course welcome,

Hi steve,

we can confirm the solar switch as an ideal diode idea. been playing with a small 1.5 volt nominal system..no room for any diode drop at these silly low voltages.

having good luck with a tiny pv array from a broken calculator as our gate voltage source..doubt that this logic level fet is ever into saturation but on enough to pass all the current this 1.5v pv has to offer (isc is the same before and after the switch)

some pics:




we will be following along..you got some good stuff cooking!

cheers, dave

[Madscientist267]

Damon -

I haven't built it, but the circuit in that schematic should work for anything. Just a matter of tweaking it to fit the voltages and currents you need. An IRFZ44N can handle almost anything that any individual panel could throw at it, but simplicity is king, I'm learning.

Dave -

I like the idea of using such a small source. I had actually considered a much smaller panel for the job, from some solar yard lights. They cook up 4.8V or so in full sunlight, theoretically enough to drive the gate into saturation for a logic level MOSFET. I left enough overhead that so that the FET stays in saturation even at reduced light. It's not a big deal if it doesn't, because the heat is proportional to the voltage drop x current (obviously); the voltage drop is limited to about .7V or so, even if the source is insufficient.

Anyone -

Remember that one panel can drive multiple 'diodes' - the circuit is pretty much cookie cutter for common ground systems. Also, don't forget that the driving panel only needs to be able to source voltage, not current. A very small panel can switch a virtually infinite number of MOSFETs with ease!

I put a very small heatsink on mine so that it can run in 'passive' mode, even though testing has proven to me that in my circumstances, it clearly doesn't need it. It only gets lukewarm at 3A continuous even without a heatsink. The tranny's 'diode' (for the one I'm using) is rated at 85A continuous. I'm sure you'd need a heatsink for THAT, but I'm nowhere close to needing one. Added it purely for peace of mind. When fully saturated, there is no heat, period.

Speaking of which, I did the permanent installation of mine today:


That's it there with the small copper tab on it.

Also did a little 'yard art' today to help with tracking until I get a more automatic system:


I call it 'Solar Henge'... I didn't intend originally for it to have this look, but that's just how it kinda worked out. I do however confess that I tweaked the functional part of it to help it out aesthetically.  

It's not immediately obvious how this works just by looking at the pic. The handle of the array lays on the blocks as the day progresses, providing the change in elevation. This is just a temporary workaround until I get a real tracker built.

Which leads me to the next part of today's post...

My step-dad is going to see about getting me a satellite tracker system for use with this. It will require some modifications for sure, but comes with a few interesting possibilities as it is not just a C-band dish tracking device.

The unit in question is for marine use, and is designed to find satellites and lock onto them, and keep track of the beam no matter what the boat does. They're kinda fascinating to watch, I helped him install one once.

Its going to take some engineering to get it working as a solar tracker, but one cool aspect of these things is that you can nail them with a computer and tell the dish exactly where you want it to point. A table with azimuth/elevation angles for given dates and times comes to mind... heh

BUT, and it's a big one - it uses steppers to do it's magic, and of course this means that it will draw power all of the time, if it is not implemented with a little bit of electronic modification (pronounced 'power kill'). That's something to work on, but isn't completely hopeless. He says that the elevation uses a worm gear (so that means the elevation motor drive can be dropped when it is in position. Thats half of it. The not so easy one is the azimuth; it is a ring gear type thing. The non-powered 'self control' is likely nil.

Another problem is, the panels collectively weigh quite a bit more than the gimbals on the antenna unit are designed to support. There's a chance that they can support it just fine, as he says they are fairly heavily built, but balance may be an issue. A counterweight is highly likely a must.

We will see. There's a possibility I will have it in my hands as soon as tomorrow, as I have to make a trip in his general direction anyway for something else.

One thing is for sure - I'm going to have some fun with it with the computer before I go full bore into it and tear it apart to make it 'my own'.  

More as it comes,

Steve

Oh and as for the pic? Mods, if anybody's uncomfortable with it, I'm perfectly cool with someone nuking it. It's not all that key to the story. Just kill the side references to it as well so people aren't scratching their heads. If reading this statement confuses someone, don't be. You didn't miss much. Might I just add a quick note for Damon - download it first before it go bye bye!

[Madscientist267]

Hasn't been much to see the last few days, although today I 'slammed' together a project thats been cooking in my mind forever, just haven't really gotten ambitious enough to do it.

I have some pics, but I'm going to hold off on them until after I have all of them. The goo is drying right now, so it will be at least tomorrow night before I post them.

I've had the need to put a volt and ammeter in the same box for ages now, and honestly, I'll be glad to get my Fluke and RS meter back! 

I wanted to go analog, but it just doesn't seem to be in the cards, even though I built one several years back that served the purpose. Unfortunately, it really wasn't geared toward use in RE. It was really more of a distribution box designed for use with the UPS that held the servers, and just happened to measure current and voltage in the process. A lot of connectors that I don't even use anymore (coax DC for example), so when I couldn't find any more d'arsonval movements, I bailed and decided to look for some chinsy digitals that I could twist up into my needs.

Turns out, I found a couple of them at the local auto parts store, for $15 a pop, and they had 'cut me up with a dremel and make us yours' written all over them. Accuracy isn't all that bad either considering the price.

The overall goal with this particular box is to provide an all-in-one 'brains' box for when I want to take the panels out and show them off in their full glory. The idea hit me that I needed to finally build this when my parents came over for easter dinner and my step-dad and I were looking at the panels and the accessories I had made for it. It hit me that there was no simple means of getting the basic vitals for the charge current and battery voltage without hooking a bunch of crap up.

No more. This has both, and measures either panel or net current, positive or negative, is selectable for use with anything from 6-24V by flipping a switch, has an additional fuse in the box above and beyond whatever might be in series with the attached battery, and a flashing LED wired through a bridge so that it blinks if the fuse is blown, regardless of current direction.

I was going to put a few white LEDs in it to illuminate the meters (for easy reading at night), but ended up leaving the 'option' for a later date. The switch is present, but when I played with shooting light into the displays, my efforts were discarded by layers of (apparently) opaque plastic. I'll figure it out someday, and went ahead and provided myself the wires inside so that whenever I finally get around to it, all I have to do is figure out how to make the light effectively reach the displays.

One of the biggest surprises for me is that given the amount of time I put into it, it looks better than almost anything else I've made. Somehow, it all just fell together.

Can't wait till it's done so I can bring it online and post the pics, from beginning to end. I took pictures just about every step of the way in case someone wants to build something like this. I don't have much more than about $60 in it, but to me, it's going to be worth much more than that just because I don't have to 'borrow' my real meters from the system when I want to use them anymore! 

Steve

[Madscientist267]

Here they are as promised... Yesterday wasn't all that great in terms of sun, so I waited for today instead. Got home and the dump was running full out. First time in a few days that they had seen a good charge.

The project started with a couple of cheap digital meters as mentioned...


These things are $15 a piece at my local auto parts place. They're not even remotely high end, but tell me that YOU would cut up a Fluke... 


Popping one open didn't reveal anything super exciting either. They don't even have 'real' chips in them, just the epoxy blobs:




First thing to do is to determine where on the board to make the permanent connections for the selector switch. After removing the 3 screws that hold the board to the face, the selector is revealed:


The disk isn't held in by anything, and comes out readily. Also, take the LCD out of the case at this point by very carefully pressing on the display's window while holding the meter face down. Put the entire display somewhere safe that nothing will bother it. Don't disassemble it any further than removing it from the case; LCD connections of this variety can be finnicky and don't like to be bothered very much.


To determine the contacts that need to be bridged, the easiest way is to hold the board in place on the now empty case and mark a line from one end to the other that corresponds to the selection you want to make permanent:


In this picture, I was setting up for the voltmeter. Once you have the contacts marked, take the selector, place it on the board as it would be inside the original design, rotate it to the selection you chose, and 'scrub' it back and forth while putting pressure on it against the board. The contacts on the selector will rub away the sharpie marks on the board that need to be bridged.


For the voltmeter, I decided to make a provision to be able to switch between 12 and 24V readings. This was necessary because the meter's limits are based on twos. For voltage, all 6 contacts are used, and 4 of them are common to both voltage ranges. The other two are used to select the range and move the decimal point:


The 6 wires running to the contacts for the range and decimal, with the other 4 bridged permanently.


The same was done for the ammeter, in the respective position for the selector:


Since there were no ranging requirements, the '10A' selection was permanently selected. There are only 4 contacts in use for this mode.


Once the meter selections have been made, the bottom of each case was cut off and the probe connectors removed from the main boards. One of these was set aside for use with the ammeter (contains the shunt), the other was discarded. The connections are then made directly to the boards where needed for the appropriate function. The LCD displays are then put back in the original cases, and the board fastened in place with the 3 screws.


The meter box's cover is cut out for the switches and displays, and the parts arranged and soldered.


It was very difficult to get a good shot of the most critical part of the entire project, the shunt. Extreme care must be taken as to not circumvent this simple device's function. Connecting a wire in the wrong location will lead to very off readings, and will make the whole thing pointless. The original interface between the shunt board and main board was of the 'jumper pin' variety, and was replaced by pigtails. They are twisted together to help reduce interference from noisy peripherals/loads, as the voltages measured across the shunt are very small indeed.


Note that the 'input' for the shunt is actually going through the original probe connection holes, with only one slight modification. I reinforced the traces on the board with the heavy wire coming from the switch. These wires make contact in the original solder loops, but are also attached to the very end of the shunt. It is important that the shunt not be 'shortened' (or lengthened for that matter) electrically!

Also visible in the picture above is the load carry diode, which is only there to carry the load through the net/panel current selection process when the switch is 'broken'. It is a standard silicon 6A diode with a 0.7V drop. It doesn't need to be rated at the full capacity of the box (10A by fuse) because the amount of time it sees a load is so small. It has no effect on the shunt because the voltages across the shunt are minuscule compared to the Vf drop of the diode.



A peek inside the 'container' part of the box right before final assembly. Note that there are two 9V batteries, one for each meter as they need to be isolated both from the system, as well as each other. The common power switch is a DPDT (which I am using as a DPST) to control them.

NOTE: This isolation is very important! Don't try to connect them to a common power source (ie single 9V battery), or to the main (system) battery, as tempting as it might be! It will have unpredictable results at the very least, and likely will cook the meters (particularly the ammeter)!!!


TADA!


Up and running, it took all of 10 hours or so from beginning to end, in the price range of about $60. Maybe a little more depending on what you do or don't have laying around.

The calibration is reasonable, and more than adequate for portable use at these low power levels. I made some minor adjustments just to bring them in line with the averages of the other meters that I've been using.

Not bad considering the time and money put into it.

I do have one minor mod I want to make to it, to allow for more accurate voltage measurements when it is wired up to my main system. The voltage drop from the cable leading from the meter to the battery is just enough to throw the numbers far enough out to irritate me. I'm going to add an 'internal/external voltage sense' switch to it with an additional connector that will carry sense leads directly to the 8D. For portable use with the 36AH SLAs, the cable to the battery is short enough that this is not nearly as much of a problem.

In other news, I picked up the satellite tracker today, but haven't had much chance to play with it yet. But rest assured guys - for those that are waiting for me to conjure up the 'cheap' version of TrackerJack's controller, It's coming!

One thing for sure, by looking at the dish, I've got my work cut out for me!

More pics as I play...

Steve

[commanda]

It never ceases to amaze me how many different ways there are to skin a cat (apologies to any and all feline lovers present).

I solved this metering problem with a picaxe, an lcd text display, and a pair of current shunts with amplifiers.
I measure battery voltage, charge current, and discharge current.
I also accumulate AHrs in & out, and show the current remaining battery capacity as a %.

To illuminate those displays, you need to point leds in from the side.

Amanda

[DamonHD]

Neat.

My cheap meter is almost exactly the same as those (has a transistor doodah in the bottom LH corner too).

How long will those batteries last with the displays running (as alkalines or better, 'hybrid' low-self-discharge NiMH rechargeables)?

Have you considered some sort of isolated floating supply run from your main batteries?

Rgds

Damon

[Simen]

Have you considered some sort of isolated floating supply run from your main batteries?

Like those dc-dc converters found on old coax network cards? most of them have 5V to 9V converters, but some of the older fullsize cards have 12V to 9V converters. *I have been using one of the latter ones (a CFUS1209-C ), powering a voltmeter for a year now without problems. (It has even survived 16V+ charging during wintertime... )

[ghurd]

"These things are $15 a piece at my local auto parts place"
You have heard of the $3 HF meters, right?
If you get lucky, the $3 version of the week is lighted.
G-

[klsmurf]

Hi Steve,

 I enjoy reading your threads and look forward to your version of a solar tracker. I think I must have missed something, however. $30 plus 10 hours of labor? A turnigy meter is only $30 and cumulative readings are available with it.      Kevin

[Madscientist267]

Wow, a little more response than I thought I'd get heh...

...different ways there are to skin a cat... ...point leds in from the side...

Very true. It was the quickest and dirtiest way I could think of to pull it off in a hurry... Since it's still a 'learning' system, I didn't want to dump a whole lot into this part of it, but wanted my real meters back. When I go to do the real system, I'll probably even go as far as to do a logging design for both so that if I should desire, a computer can keep track of what is going on.

The LED problem is tricky, there's not a lot of room in there next to the displays, so I'm even considering a little bit of fiber optic 'feed' to bring the light where I want it. They don't even need to be bright, just enough to kill the darkness.

How long will those batteries last with the displays running (as alkalines or better, 'hybrid' low-self-discharge NiMH rechargeables)?

Have you considered some sort of isolated floating supply run from your main batteries?

Not sure about the battery life just yet, still running the cheapies that came with them for the moment. I am kinda disappointed that Energizer quit making the 9V lithiums, those were awesome for meters - lasted forever it seemed like!

I actually did consider the rechargeable approach, so that when the power switch was in the off position, I could bleed charge into them from the mains, but unfortunately, it would have required a quad pole switch to pull off (or some really intuitive FET usage), and so I bailed on it at the last moment and just decided to go with regular batteries.

Isolated supplies are a possibility, and that thought also crossed my mind, but anything more complex (or expensive) than a cheap buck converter (such as a modified cell phone charger) gets outside the range of what I was willing to put into it. Since bucks aren't isolated, I shied away from it.

I have been using one

Where do I get them? They don't seem to be readily available (from a quick search online). Are they super simple, efficient, and are they cheap enough to make it worthwhile for this? Gotta keep parasitics to a minimum for a system this size. 

Might have an old network card laying around somewhere... ?

If you get lucky, the $3 version of the week is lighted.

Thought I told you guys to psychically notify me when I'm about to do something like this so that I don't spend too much money! LOL

In actuality, the funny part is I had just picked them up with a slightly different use in mind, and then I realized I had the project box that was big enough to hold them both and poof! A combo meter was born.

A turnigy meter is only $30 and cumulative readings are available with it.

Figures. LMAO. See above.

My standard response to that is "yeah, but I didn't build them, so what's the fun in that?" 

Not that I built the cheapies in this one either, but eh... It's all good. It does a couple other things as well, so I guess it's a wash. 

Steve

[Simen]

Where do I get them? They don't seem to be readily available (from a quick search online). Are they super simple, efficient, and are they cheap enough to make it worthwhile for this? Gotta keep parasitics to a minimum for a system this size. 

Might have an old network card laying around somewhere... ?

The best source for these are old nic's; try your local junk yard or wherever electronics are recycled. (it's the old cards with bnc connectors, popular in the 90's that have them.)

They're not extremely efficient; the 12 to 9V i'm using are at 68%, and uses around 25mA unloaded, but they ARE super simple; a capacitor on the input and one on the output are all that's needed. And sourced from an old (free) nic, makes the price right.
They come usually in an 24pin dil package, often with only 8 pins.
http://www.datasheetarchive.com/pdf-datasheets/Datasheets-314/285595.html

[DamonHD]

Steve,

By the NiMH rechargeables I meant, um, charge them somewhere convenient (eg from mains or RE) for the few hours required and physically put them back into the meters once every few months!  Not massively elegant, but in keeping with the excellent KISS already at work in your stuff!

(I checked, my near-twin of your meter that was £5 on offer is now £9 at Maplin.  Mine is the M-830B and badged 'Sinometer' so can presumable tell me the local density of Sinos per km^2 on the right dial position; is it facial recognition SkyNet tech or audio do you reckon, down-weighting loud Cantonese speakers of course?)

Rgds

Damon

[Madscientist267]

charge them somewhere convenient (eg from mains or RE) for the few hours required and physically put them back into the meters

No, I know. But my major deterrent for regular batteries isn't the cost of replacing them, it's opening and closing the case. Things got a little cramped around the edges and it's a fight (pronounced 'b1+<h') to get the front cover back on when I have to go inside... I'd rather have them that last for eons and not have to mess with them at all... so who knows... I might come up with a better scheme, but until then, Duracell it is...  

Okie, well I had the chance to dink with the satellite tracker today... I got it working, although nobody warned me going in that some assembly was required.  

I spent the first hour or so wishing I had the captain crunch decoder ring that gave the location of the repair manual for this thing. LOL

The two steppers (each as bad as the other) can be mounted in one of something like 45 different configurations, and to the untrained eye, would appear to work just fine. Which is what happened... I'd try it one way, and a snag would reveal itself, so I'd try it another way, and a different snag... so on and so forth until I finally cracked it. Doesn't seem like it would be all that difficult, but the motors are 'universal', in that the units are designed to use the same motor for both azimuth and elevation, only the little adapter plate is different. Add this to the fact that at first glance, almost any configuration appears that it would work just fine, and you have a recipe for disaster. I digress... here are some pics...


This is the dish, the two motors, the cables, and the lower half of the composite weather shell, that was conveniently (or not, depending on how you want to look at it) secretly loaded with nuts, bolts, washers, and so on. Once I picked all the little bits out of the bottom of the unit, things began to look up. They were causing the azimuth mechanism to jam for seemingly infinite reasons. Don't get me wrong, I'm not really complaining, as the price was definitely right, but I personally can guarantee that the afternoon was not without frustration...


Here is the back side. The silver box at the back of the dish is the gyro... Turns out that even if I keep using the steppers with the original circuitry, this box will do absolutely nothing, so off it will come to make for other interesting toys.


Here it is, after assembly and testing, with the weather shell 'mod' to make the parts more accessible for working on it. It will be an interesting ride to see if this thing can be modified easily to take the seemingly overshadowing load of the panels and move them about. Geared motors (rather than steppers) and counterweights come to mind...  

And, this afternoon, after I mocked up the data cable, I had some fun with it and wrote a quick program in QuickBasic (yes, a M$ product, I know... LOL) to make it dance around a little bit. I did basic controls for azimuth and elevation, some direct 'pinpoint' commands, and a few macros to work up a few laughs for those that can appreciate my twisted sense of humor. Here's the video:

http://www.youtube.com/watch?v=G9zhmavQgKU

All in all it works great now, there was only one hitch other than the whole motor mounting thing. The azimuth limit switch wasn't activating properly at first. After removing the weather shell, I was able to determine that it was simply a matter of tweaking the lever on a leaf switch so that the cam on the limit mechanism would actuate it. A quick twist with the needlenose, and voila, all was working as planned!

We'll see how this goes. This part of the tracking system will probably sit just a little. I need to think long and hard about the best way to go about doing this so I don't dump a bunch of money into it for nothing. Then again, I'd have to actually have a bunch of money to throw at it first...  

I'm going to continue on the quest however for the low cost 'eye'. In the end, it will be needed anyway, so I'd be just as well off to do it sooner than later.

Until next time,

Steve