The best set up for 2 Vcc levels...

Projects General Guidance
1

So, I am running one of these 433Mhz Tx modules. Apparently they like up to 12V supply.
The Attiny85s like 5.5V max.

So I have come up with 2 solutions for solar powered units...which is the one that would be best suited?
Or is there another cheaper way (i.e. trying to avoid any boost converters that cost like £3 a module!)

Do I want the 2 6V panel option...where each module (tiny85 and 433Mhztx) is powered "separately".
OR do I want the 12V panel that uses a voltage divder that will top up the super cap.

Full size link:

Thanks!

2 Panel Set Up.png
2 Panel Set Up.png3326×1072 82.6 KB

2

Look closely at the diagram on the left - your ground connections completely short across the RHS solar panel. I have a feeling that you won't get much voltage out of that panel.

In the diagram on the right, I see 40mA on the solar panel. Where did you get 40mA? At 12V, there'll actually be 60mA flowing through the resistors. That's wasted power. How much current is the panel capable of providing?
Why don't you consider a regulator?
Also, there's nothing to stop the supercaps from charging to a higher voltage than their ratings, if the solar panel can provide enough current.
For instance, if the 12V panel hits 14V, there'll be 6.3V applied to the ATtiny and it's supercap, and 12.6V applied to the RF module and it's supercaps. Again, assuming the panel can supply enough current.

3

Ah I see!

Didnt notice I shorted that panel :|.

So a simple 12V panel with some regulators for each voltage I need is best?

How about this in a TO-92 (easy soldering for me!)?
Low Power Low Quintescent Current 5V regulator

And then for the 12V supply, just a simple fixed 12V LDO like this:

With 2 general silicon diodes to drop that to around 10.6V?

I was more worried that components like regulators may just sap power when not in use...

PS> The 40mA was what the panel "could" produce.

4

Oh god I dont know now.

These seem nice and cheap and plentiful. Says 5.5mA Quintescent current...but thats when operable? I assume when no input is provided, these are dioded so they do not draw back from the super caps?

Cheaper 5V regulators.

5

You just posted again. Regarding those supercaps, see below.

Johnny010:
Ah I see!
Didnt notice I shorted that panel :|.

So a simple 12V panel with some regulators for each voltage I need is best?

How about this in a TO-92 (easy soldering for me!)?
Low Power Low Quintescent Current 5V regulator

And then for the 12V supply, just a simple fixed 12V LDO like this:
http://www.ti.com/lit/ds/symlink/lm2940c.pdf

With 2 general silicon diodes to drop that to around 10.6V?

I was more worried that components like regulators may just sap power when not in use...

PS> The 40mA was what the panel "could" produce.

There are other problems as well. 4F caps won't hold anywhere near enough charge to power your circuit overnight and through heavy overcast periods.
These are the basics:-
1 Farad = 1 Coulomb per Volt
1 Coulomb = 1 Amp-Second
1 Farad = 1 Amp-Second per Volt

So if you charge a 4F cap to 5V, and don't want it to discharge below 4V, you can only draw 4A for 1 second, 66.7mA for 1 minute, 1.1mA for 1 hour, 110uA for 10 hours etc

Then there's the panel. It needs to be able to supply enough current to run the circuit and charge a battery or supercaps to allow for night and overcast periods. Unless you're in a particularly sunny place, it needs to do that without full sun.

Finally, for a 12V LDO regulator to work, you don't want the panel output to drop below 12.5V.
Still beats drawing 60mA through those resistors from a panel that can only provide 40mA total.

I was more worried that components like regulators may just sap power when not in use...

Not as much as those resistors. :slight_smile:

I don't get why you want to drop the 12V supply to 10.6V for the RF TX module, either. If it's a typical 3V to 12V module, 12V is better.

Anyway, overall you need to put a fair bit more thought into your design, I'm afraid. And while you originally didn't like the idea, switching regulators aren't so evil or expensive. Unfortunately, you can't design something like this without spending a little money.

Sorry to be such a wet blanket. :frowning:

6

Thank you for the reply.

Well I have watchdog on the Attiny85, it loops in watchdog for ~15 minutes, wakes up, takes a reading from a DS1820B temperature sensor, and then sends the value (can do it say 3 times as apparently these 433Mhz Tx modules are "flakey"). So overnight, say 12 hours, the module wakes up for about 1 second ~50 times overnight.

So as far as the caps I am pretty happy they should last! The Attiny85 is happy to around 2.7V regardless...

I think I see what you mean about powering the Tx module, I cant seem to find high capacity caps at 6V or 12V that are "cheap". The 12V ones seem to be massive and built for car amps.

I thought about placing 5 2.7V (100F) ones in series, giving 1/(5/100) = 20F equivalent capacitance. Much larger of course...but I assume the Tx is the power beast. It would be 13.5V, but a diode after the panel will pull the voltage down to 12V max ish anyway. This adds a lot of cost and size mind.
A pack of 6 of these 2.7V 100F caps seems to be around £15...

I can get the 4F 5.5V ones for about £1 each! So I guessed 2 5.5V 4F caps gives 2F capacitance equivalent...and still should be enough for the use of the Tx 50 times in a night?

Dont the regulators still pass voltage although they are under volt on the input? Is it just the drop out voltage loss over a range? So even if the panel was giving say 8V for the whole day, at least ~7V will get to the caps?

7

Johnny010:
So, I am running one of these 433Mhz Tx modules. Apparently they like up to 12V supply.

KISS.
433Mhz transmitters work perfectly on 5volt.
Only less RF output power on a lower voltage (less power consumtion too).
Don't confuse module output power with range.
A module running on 12volt with a simple coil aerial will have less range than a module on 5volt with e.g. a DIY "sleeve" aerial made from a piece of thin coax.
For range, a proper aerial on the receiver (not a small coil) is a must.
Coil aerials and inbuild aerials e.g. in keyfobs are for convenience, not for range.
Leo..

8

As Wawa says, those RF transmitters work on 5V. You only really need 12V if you're going for absolute maximum range. (For a given antenna, you definitely get much better range with 12V than with 5V.)

I was picturing the micro being awake the whole time, and also more things connected than just one DS18B20 and the RF module.

If you do use a single 5V supply, you need to ensure that it doesn't drop below 3V, or the DS18B20 and/or RF module might fail to operate. Also, down at 3V, the RF module's range will be more limited than at 5V or 12V.

You still need to ensure that your supercap(s) can't charge over their rated voltage.
And to see if you have enough capacity, you need to add up the total current consumption of the DS18B20, RF module and tiny85 for the time they're active, then use the methods I showed to calculate time vs capacity.

The RF TX module is the hungriest. The simple 3V to 12V 10mW ones that I have here use 45mA at 12V. The DS18B20 only uses a maximum 1.5mA when active and the tiny85 uses up to 8mA when active, (5V, 8MHz external osc). I couldn't immediately see the consumption for an internal 8MHz oscillator. It'll be higher.

Edit:

A pack of 6 of these 2.7V 100F caps seems to be around £15

If you find you need to go this route, you might be better off looking at batteries.

9

Thanks again OldSteve.

May have found a solution.
The panels seem to only produce 60mA @ 13V in strong direct sun.

From this, with a bit of internet searching and remembering seeing something like it elsewhere...
I have come up with this:


FullSize: Zener Diode Regulator

About to breadboard it up and test it with some usual LEDs before trying to attach any caps to it...incase I burn the house down.

Update:

Built this up and left it in strong sun for 10 minutes. The caps are charged (10.5V on one and then 5.05V on the other) :). So it looks like it will work!

Thanks for the input though!