2

I have been working on a project of a commercial LoRaWan node. I want it to be powered by 3.7 battery and I consider to add a solar charger. For now, I have implemented a circuit based on an LTC3652 IC and it works quite good.

I found such solution today. What do you think of this one?

enter image description here

I'm not hiding that this solution is way cheaper and easier. Have you ever dealt with this circuit? What's your experience on this one?

Is it safe to implement this circuit in commercial solution or you think I'd better stay with the circuit mentioned in the beginning?

  • 5
    You first mention a 3.7 V battery so that is likely a Lithium based cell. The circuit shows a 6 V battery which is likely a Lead Acid type. You cannot and should not just charge a Lithium based cell with a charging circuit designed for a Lead Acid battery. Lead Acid batteries are much less "picky" with regard to how you charge them so then you can use this simple circuit. For Lithium cells you should not, you must use a charging circuit which is designed for Lithium cells. Unless you do not mind your battery overheating/smoking/catching fire etc. – Bimpelrekkie Feb 14 '19 at 11:09
  • This is not a good design. Bypass capacitors are missing and T1 should have a base pulldown resistor of 100K or so. Voltage clamp is not effective until about 7.4 volts. This may work for 7.4 volt NiMH batteries or 6 volt lead-acid batteries. –  Feb 15 '19 at 02:20
  • LM317 was an amazing IC when it first came out. Nonetheless, and with no ill-will directed at the original designers (who were giants among men) at this point, I would recommend that it not be used for new designs, regardless of its official production status. – user57037 Feb 16 '19 at 02:35

2 Answers2

7

Since you say this is for a comnercial product, I'll be up front about this: Given the choice between buying your product with an LM317 regulator or a product from your competition using the LTC3652, I'd rather buy from your competition.

You are saving a few cents, and delivering an inferior product.

The LM317 is in all ways inferior to the LTC3652.

  • The LTC3652 uses Maximum power point tracking (MPPT) to make the best use of the power from the panel. The LM317 can't.

  • The LTC3652 is an efficient switching regulator so as not to waste power from the solar panel. The LM317 is an inefficient linear regulator, whose efficiency gets worse when the panel provides more power.

  • The LTC3652 is a charge controller. It has selectable charge termination, and maintains the battery at full charge by monitoring the voltage and switching back to charge mode when needed. Your LM317 circuit has charge termination, but it depends on the properties of a zener diode and a transistor - making it properly match your battery could be difficult.

  • The LTC3652 can easily set the charge current - the datasheet tells you how. The LM317 circuit you have can also set the charge current - does the source of your circuit tell you how?

That LM317 circuit is simpler and easier than the LTC3652 in the same sense that it is simpler and easier to light a campfire on the kitchen floor rather than installing a stove and oven. It's cheaper, but it sure ain't better.

As already mentioned in the comnents, the LM317 circuit is not a good choice for lithium batteries.

The LTC3652 isn't either, except for LiFePO4 which is explicitly mentioned on Analog's web page.

Both circuits seem to be better suited to lead acid batteries.

longneck
  • 237
  • 1
  • 6
JRE
  • 71,321
  • 10
  • 107
  • 188
  • you're talking about a 7$ chip where entire lorawan modules powered by an esp32 go for <$20. Adafruit sells a solar charging breakout board that uses MCP73871, which seems quite well suited to the task. (and only ~2$ on digikey vs ~7$) – Aaron Feb 14 '19 at 15:06
  • The LM317 was designed to be a linear regulator. It was never intended to be a solar-panel battery charger. You are criticizing an orange for not being an Apple. – Edgar Brown Feb 16 '19 at 05:30
0

Commercial use and outdated, inefficient linear regulators are not good partners. Modern switching supplies cost little more and when factoring in things like heatsinks, can and often do cost less overall.

For lithium batteries, a battery manager designed for such batteries is an essential. If you don't use one, expect lawsuits as you're essentially creating a firebomb. Lithium batteries are great but, improperly handled, they are dangerous. If you're asking such questions, you really need to hire an experienced design engineer to do the design for you.

SamGibson
  • 17,870
  • 5
  • 40
  • 59
Peter
  • 21
  • 1
  • 2
    I agree with you. HOWEVER, let's take a moment to appreciate how good the LM317 was for its time (1976). It has been continuously in production for over 40 years! It is older than most of the people who ask questions on this site (but not older than the people who answer them, I suspect). – user57037 Feb 16 '19 at 02:39
  • 1
    @mkeith The dynamic Widlar-Dobkin duo made the LM317 do double duty- a change to the metalization mask only resulted in the LM395 "ultra reliable power transistor", it was otherwise the same die. – Spehro Pefhany Feb 17 '19 at 00:07
  • Ha ha.... mkeith must be more than 40 years old. – soosai steven Feb 17 '19 at 02:45