Author Topic: LOW POWER/COST LM2596 POWER POINT REGULATOR (Read 65767 times)

vtpeaknik · « **Reply #33 on:** June 05, 2012, 05:02:27 PM »

Good point. The datasheet for the MAX8212 says 2V to 16.5V supply-voltage range, 18V abs max. But, since it only needs a few microamps to function, can easily arrange for it to run on some fraction of the input (PV) voltage? I.e., if you're using PV with Voc between 16 and 32, use a 2:1 voltage divider (two equal resistors around 10 kohms, and perhaps a capacitor). For the voltage sensing (pin 3 of the MAX8212, 1.5V threshold) use a separate voltage divider.

OperaHouse · « **Reply #34 on:** June 05, 2012, 05:29:15 PM »

Or use a LM431 driving an opto isolator for a wide adjustable range. Or a simple zener and a pot. It really isn't that critical. Just wanted to get out there that this method is pretty easy to implement and these boards at under $2 can be useful. Any regulator feedback input can be forced high to turn off switching so you are not limited to use a control pin. This would be a really good method to control speed with a treadmill/pedel generator or cut in with a windmill. Good to see some interest in this.

dnix71 · « **Reply #35 on:** June 05, 2012, 10:47:05 PM »

Not much point in kit building when for $15 including shipping you get a 150 watt boost module straight from China.

http://www.ebay.com/itm/150W-DC-DC-Boost-Converter-10-32V-to-12-35V-Step-Up-Voltage-Charger-Regulator-/110773897686?pt=LH_DefaultDomain_0&hash=item19caa389d6

$9.40 gets a booster that runs on much lower voltage and uses the LM2577

http://www.ebay.com/itm/LM2577-Boost-DC-DC-Step-up-Converter-3-34V-to-4-35V-3A-max-regulator-module-/110761077095?pt=LH_DefaultDomain_0&hash=item19c9dfe967

OperaHouse · « **Reply #36 on:** June 06, 2012, 04:28:38 AM »

It is easy to get confused. This discussion is about modifying those Chinease boards so they can be used as power point regulators. Those boards you listed have nothing to do with the buck/step down converters we are talking about. I wish they had one for that price that could be converteded into a power point regulator. Units offered in that power range are generally 24 to 12V vehicle units that are potted and can't be modified. And when you read the fine print on China boards the power is usually half of what they say.

independent · « **Reply #37 on:** August 10, 2012, 02:06:25 AM »

I've reconfigured my LM2596 regulator. I've wired in the plain single pot LM2596-adj cheapo regulator with the LM723 MPP backend. It's more efficient than the 3 pot version of the LM2596. OperaHouse's advice about looking out for the regulators with the highest voltage caps is gold.

One thing I was able to do is de-solder actual LM2596-adj chip from the board it was attached to and attach a proper heatsink to the regulator standing up as it ought to be. Some might be welded on but I doubt it. It's a bit tricky and you might ruin your regulator but it worked for me.

These are very versatile regulators and one use for them is with low wattage 12v panels for providing a regulated 5v out to a USB charging device of which there are many (it's an international cellphone charging standard now). 5v regulated is a difficult trick to pull off without a battery to regulate it. I've seen another guy's version where he boosts his using massive solar cells and that looks good too.

A problem with the 5v regulated output is all smart-phones have different charging requirements but that's another story (mainly to do with data pins, shorted or not, some requiring voltage there..).
An advantage of using the LM723 regulator is that it needs 8v to function so in theory, when there isn't enough light it will just turn the regulator off.

With 4700uf cap on the input (salvaged) plus two smaller Low-ESR caps on the input and 1200uf of Low ESR capacitance on the output with very moderate evening (sun just about going down), 300mA was measured going into a battery at 5v. I know this particular device ramps down it's charging requirement when the battery capacity registers over 800mAh on the BME chip, so 300mA might be it's lower requirement (it was exactly 300mA at it's highest at that time of the day). I am curious to see how the USB charging device works when it needs the full 900mA as it will do when the battery is less than 800mAh and if the 5w panel + the LM2596 can supply it.

So, 300mA in low light, 250mA as the sun went down measured with an ammeter is very good in my estimation of how useful this circuit is.

independent · « **Reply #38 on:** August 11, 2012, 01:21:13 AM »

This morning peaked at 800mA 5v from a 5w solar panel

independent · « **Reply #39 on:** August 12, 2012, 04:02:27 PM »

So, peaked at 800mA, no problems at that current, heatsink slightly warmer than ambient. No problems, a bit of inductor squeal to be expected I guess.

independent · « **Reply #40 on:** August 18, 2012, 01:08:18 AM »

As an aside but totally related to this thread.

These are not genuine LM2596-adj in these boards. I checked digikey for 1000 units and the cost is ~$US2.50 per unit. The whole board is being sold for that price (including shipping anywhere in the world) so they must be counterfeits. I know there are many counterfeit chips on the market now but it didn't occur to me to check the bulk unit prices for these components until recently

independent · « **Reply #41 on:** September 01, 2012, 12:31:45 AM »

Hello again

I've been using this circuit to charge a laptop 3s2p battery in the laptop. The AC charger uses 16.5v on this particular laptop. It's an older IBM Thinkpad. I was curious to see the built in charging circuit would behave when presented with less current than usual (and a variable one at that) but at the correct voltage.
I used a 15w solar panel and it was peaking at around 0.7A this late in the day. The charging circuit was accessed through the acpi monitoring installed in my GNU/Linux install (gentoo). I can monitor how the computer is seeing the charging process as I have compiled a full acpi suite of tools.
It works best when it is charging or discharging. That might sound strange but if the computer is discharging overall but there is some current going in it is fine. For example a 10w load and 7w of power in, the battery monitoring will say it's discharging 3000mW. If the computer is at the cusp of charging and discharging it is not so happy. That is the way the circuit monitoring the charging process deals with this is it dwells at "charged" even though it is either being charged or discharged. And if it's got enough charge to get past the base load it's fine as well as if it's in it's discharging state.

It was charging when it was sleeping (in standby). Then it stopped. I am not sure what happened there. The battery is well worn and might have met it's prerequisite charging voltage. I don't know.

It will not charge when the computer is off with only 0.7A available. I have no idea how it can sense there is only 0.7A as there is not the slightest bit of current going through, but there you go.. It might have sensed it had reached a charged state. These do have ICs and possibly battery tank chips inside the batteries themselves.

So, overall, my little experiment left me a bit nonplussed. It will work but really needs a decent amount of solar 30-60w and perhaps two of these little circuits in parallel to do it properly. That the laptop will not charge in an off-state has left me a little bit baffled but again, I aren't too fussed as the experiment told me what I needed to know. That charging directly off solar will work but there are many possible pitfalls with the process. The main one being how happy the laptop (or the battery) will be in accepting the amount of charge available or how happy the laptop will be in accepting a variable charging rate (even though it as the correct voltage).

I had in my possession an OLPC 1.75 for a week or two a month back and that laptop has built Maximum Power Point technology built into the charging circuit. LiFePo4 batteries and can handle voltages of between (11-40v from memory). It will also limit the amount of power the laptop requires depending on the charging profile. It really is the way to go for a fully solar charging laptop.

OperaHouse · « **Reply #42 on:** September 02, 2012, 01:07:35 PM »

I get a lot of computer Li battery packs for free and take them apart. It is amazing how much electronics is in them. For sure you are driving them crazy First thing I would try is a really massive capacitor on the input. These controllers expect to be able to draw a massive amount of current when they start up and they may be quickly be going into shutdown. It just might cycle between charge and standby in its own. In that same thought, I wonder with a really large capacitor bank or maybe some rechargables, it would be possible to use a 555 to trigger a FET to switch power and charge for a second or two (longer with batteries) and turn off and recharge for several seconds.

On the subject of " must be counterfeits." By some sense of the dfinition everything from China is counterfeit. $2.50 each is no where near a manufacturers price.

independent · « **Reply #43 on:** September 03, 2012, 03:27:33 AM »

I should have qualified that counterfeit claim on the LM2596s.

I've done a bit of research on this and in particular looking at National Semiconductor (Texas Instruments) LM2596s. LM2596s from reputable sources cost around US$5 a chip. There is evidence that people have been buying LM2596s and buying fakes as the units have been faulty by the batch. There is a guy who's written quite an in-depth page on LM2576 regulators he's built. Have a look. in particular one chap pays $AUS6.50 per unit, he is an electronics repairer and very weary of ebay. From my local electronics store they are $NZ21.50. Check your Mouser/Digikey catalogues. These aren't cheap chips.

Counterfeiting chips is widespread, high-end opamps especially.

The best evidence I can find is from the datasheet. In the past I have ordered LM2595t-5.0 (1A version) as a sample and found this to be true.
TO220 LM2596 pins are staggered. All pins on the cheapie units are inline.

vtpeaknik · « **Reply #44 on:** September 08, 2012, 03:08:23 PM »

"Counterfeit" or not, the question is: do the cheap ones work as they should? When you wrote "faulty by the batch" do you mean some work and some don't? What would be really annoying is if they work for a while and then quit, potentially in a damaging way (e.g., overcharging your battery).

DamonHD · « **Reply #45 on:** September 08, 2012, 03:26:22 PM »

If they've started telling you lies by selling you a fake, just how much more do you want to believe about that fake with your expensive equipment, and even that different fakes would behave consistently?

Rgds

Damon

dnix71 · « **Reply #46 on:** September 08, 2012, 05:17:32 PM »

I buy cheap Chinese components on eBay, but send them to someone I trust for destructive testing prior to use. I got 0.01uF 2Kv ceramic caps for 4 cents each including shipping so I wasn't convinced they were as stated. I can test for uF value but not high voltage failure. The caps passed the 0.01uF value test, but there was little margin on the breakdown value.

The circuit won't put more than 1200v on the cap and even spiking the caps with a cheap power supply only caused 1 to fail at 1200v. All the rest seemed to breakdown above 3600v, one even held to 6kv, so they will work.

It's easy to test a cap or resistor, but not so much a logic chip.

OperaHouse · « **Reply #47 on:** September 11, 2012, 03:52:53 PM »

I didn't find anything on the internet really convincing as to fakes. A lot of people just pretending to know something or re hashing someone elses story. What I do know is that I would be taken out and shot if I proposed a $2 regulator. We have sold microcontrollers that didn't have $2 in parts total in them. Pricing is likely based on location and if you can't sell a boatload to China you might as well stop production. You are probably more likely to have actual "factory chips" being sent out the back door as no name products. Let's face it, these are just sand. The metal tab is the most expensive part. They have to price it so they don't get coppied.

independent · « **Reply #48 on:** September 12, 2012, 04:28:10 PM »

OperaHouse. I believe you didn't read my post correctly. The pins on the genuine LM2596T-adj -12.0 and -5.0 are staggered. The pins on the fakes are inline. These are for the TO-220 ("T") version. That is as good a proof as you are going to get in my books.

OperaHouse · « **Reply #49 on:** September 13, 2012, 02:49:01 PM »

Watched the evening network news last night and they did a segment on the economics of the new iPhone. They said the cost to manufacture it was $8. It must have a lot of those "counterfeit" chips in it. I think that is a big ZERO on the staggered pins. It would be unfortunante if all this talk convinced some people not to use them.

independent · « **Reply #50 on:** September 13, 2012, 03:28:29 PM »

EDN states $167.5 for iphone 5, isupply $188 for the iphone 4s. I don't see how a news statement proves anything, it's a correlation not a causation. Those were the first two google results for "BoM iphone".

P.S. I have a deep respect for your electronics knowledge OH and recognise I have a long way to go before I even come close to your level of understanding in this area. I just don't happen to agree with you on this matter. This discussion of counterfeit chips doesn't take away the usefulness of these small LM2596 boards but it definately broadens the way we look at these little devices..

nickskethisnikske · « **Reply #51 on:** September 17, 2012, 01:24:49 PM »

I was just wondering about to use this "concept" as dump controller, would it work?
I think it would regulate his output, you won't get on/of/on/of cycles all the time like a normal dump controller.
The advantage would be that you can keep your batteries at float voltage.

News:

Author Topic: LOW POWER/COST LM2596 POWER POINT REGULATOR (Read 65767 times)