Low Power circuits (10 years on CR2026) design feedback. Is it worth using a ...

[V3r50n]

Is it worth using a 300F-400F supercapacitor with a solar cell for a project that would be outdoors all the time vs just using an AAA or coin cell with no solar cell? We are talking about what would make the product last longer in the field without replacing anything.

Here are my personal pros and cons for using the cap and cell.

Pros:

• Possibly last longer in the field (Solar cell 25 year+ life span, super cap if treated nicely over 10 years?)
• More efficient charging from solar (for cloudy days)

Cons:

• Cost more money, would customers be willing to pay for it?
• More complicated design, leading to higher BOM
• Device will need to be larger to be able to hold the cap.

So if you wanted a product to last longer than 10 years (since lithium bats have 10-year shelf life) or is it not worth it to worry about a product after 10 years? what approach would you take?  Anyone have insight on the actual life span of these caps?

[thm_w]

I see some 300F super caps on digikey for ~$12. Thats a lot better than I thought, but still a huge cost difference compared to a lithium.I don't know what a CR2026 is, but you can get CR2032 for <15c.
From what I compared its not close to being worth it. Size is much larger, capacity is much lower, you will have to have some supporting circuitry (although not much). Of course some special applications it may be of use (high discharge/charge).

What is the power consumption of the product?
10 years is too long to worry about IMO.

[mariush]

You could also look into primary batteries (non-rechargeable) like the Lithium Thionyl Chloride  ones... low self-discharge (usually under 0.1% per year, about 10 years life or more but low currents .. let's say less than 10mA if you want them to last a long time.
See http://www.eetimes.com/author.asp?doc_id=1322276 and http://www.tadiranbat.com/ and farnell : http://uk.farnell.com/webapp/wcs/stores/servlet/Search?catalogId=15001&langId=44&storeId=10151&categoryId=700000004383&showResults=true&pf=110151629

An AA style battery gives you about 3000h at 1mA, while a D-style one can give you about 25k hours at 1mA : http://uk.farnell.com/eve/er34615/battery-lithium-3-6v-d/dp/1973584

Tadiran advertises on their main page a version that's supposed to last for 40 years if you do short pulses only.

OH ... and if your product uses very small amounts of current, or the small amount of current can be stored somewhere for example in ceramic capacitors and then do a burst a discharge the caps ... you could look into nuclear batteries. 

There are chips that use betavoltaic effect to produce energy and you can buy them (but they're expensive at around 3000$ ) : http://www.citylabs.net/index.php?option=com_wrapper&view=wrapper&Itemid=25
As you can see, they're at best 2.4v at 300-540 nanoAmps but like I said, you could charge a bank of ceramic capacitors and then do something with the energy.
You can make your own cheaper version of these batteries by using regular photo cells and some tritium

[bitseeker]

If it already can run for 10+ years with a single battery, I'd just use the battery. No need to worry about availability of sunlight, cleaning the panel, replacing if the panel breaks/dies, etc. Simple solution FTW.

[Neganur]

Isn't the leakage of those super caps too high for that kind of application?

[SeanB]

How about a solar cell and a lithium pouch cell. simply charge to a lower voltage ( around 3V6) and you get a massive life boost, at a cost of reduced capacity.  Just use a cheap management unit to handle the discharge cutoff, and use the solar cell to drive a LDO to set the float voltage and the charge current, and use a FET switch to connect the charge circuit when light is available.

[jbb]

If you're talking about a commercial product, I can tell you that people get a mental blank about shelf life.

Example conversation:
Sales: "Oh no! This customer has had units failing with flat batteries after only 6 months!"
...panic panic stress stress...
Engineering: "Hang on, those units had already been in storage for a few years. They're on spec."

If you want customers to give the units an occasional recharge while in storage I can tell you now that it won't be done reliably and they will blame you when units go flat.  So if you can't accept any period of power loss (e.g. you'd loose your Real Time Clock) you need to just stump up with a battery.

[Dubbie]

Be aware that most solutions involving storing charge in caps for a long period will be defeated by cap leakage.

[Teuobk]

It turns out that you can get rechargeable lithium coin cells (e.g., LR2032), but the capacity tends to be lower than their primary-cell cousins. Also, if you're subjecting them to daily charge/discharge cycles, they'll wear out within a few years.

I'd slap in the biggest primary-cell lithium coin cell you can fit and concentrate on lowering the power consumption of the circuit instead of screwing around with rechargeables or supercaps.