DIY - NI-Mh battery charger, no delta-v and no delta-t.

[Avelino Sampaio]

Hello everyone.

I'm starting a project for a charger, with characteristics very different from traditional chargers. I want the charge to be defined by the battery itself. Observe the circuit in image 1 - see through the calculations presented, that the charge decreases as the battery voltage rises. See image 2, the behavior of the voltage and current I want to obtain. As the final charge decreases, there won't be the possibility to measure the delta V or the delta T. My real intention is that the battery doesn't heat up as it approaches the final charge. I intend to build a circuit in which the cut will be determined, when the battery voltage reaches 1.495v. I am aware that the final charge may not reach 100% but I want to reach at least 80%. My question: what kind of problem might occur with "non-linear" loading?


Image 3: Reply #22 - Final project with excellent result!

[TheMG]

In practice this is really not much different than charging the battery through a simple resistor from a regulated 1.5v supply. No need for the opamp transistors etc.

[ogden]

Never ever charge batteries w/o current limiting/regulation

[Avelino Sampaio]

In practice this is really not much different than charging the battery through a simple resistor from a regulated 1.5v supply. No need for the opamp transistors etc.

Exactly! The idea applied here is very simple. I will use either 5v or 3.3v from an ATX power supply. I will keep the transistors and the opamp, as I intend to carry out other functions with them (cut and protection).

Never ever charge batteries w/o current limiting/regulation

The circuit is still pretty incomplete. Current limiting will either be built in with a transistor on the output, cutting off Q1 and I will use a simple fuse on the input.

[floobydust]

I found it did not work. With Eneloops or Ikea Ladda LSD NiMH, they need the "top up" charge or else their capacity drops. This is the peak delta-V phase. Then you must lower charging voltage, then stop charging completely.

Steady voltage (after bulk charge) will kill NiMH either from overheating or lack of top-up. I have a cordless phone with constant-current 1.50V (it was NiCd) and NiMH don't do well, they don't last. I just pulled Ladda batteries that had dried out from staying in the charger too much, because they have low capacity without the top-up phase. They do behave differently from Eneloops I have used the same way.

You cannot "trickle charge" NiMH, this is not recommended. Max. is 0.033-0.05C which is way less than NiCd, so they dry out if charged too long or in a CMOS memory-backup use.

Graph taken from https://lygte-info.dk/info/batteryChargingNiMH UK.html

[Benta]

The best strategy for charging NiCd or NiMH cells is d²V/dt² detection. This gives the cells a long life, I'm speaking from experience, my cells have until now survived 1000s of charge cycles without significant degradation. All other strategies that I've seen result in overcharging and damage.

To understand the idea, p.4 of this data sheet helps:

https://pdf1.alldatasheetde.com/datasheet-pdf/view/26117/TEMIC/U2402B.html

Today, you'd let an MCU with associated ADCs run the algorithm.

[ogden]

Never ever charge batteries w/o current limiting/regulation

The circuit is still pretty incomplete. Current limiting will either be built in with a transistor on the output, cutting off Q1 or I will use a simple fuse on the input.

For battery charder it does not make sense to address voltage regulation first. Fuse?? You clearely do not know what you are doing.

[Avelino Sampaio]

Floobydust,

thank you very much for the page indicated, it was of great value. Thanks also for sharing your experiences. I will still persist with the tests, using my idea of applying a variable charge to the battery. As you can see in the image, I have some batteries, ranging from new (eneloop) to fully zeroed. My fun will start....

[Avelino Sampaio]

ogden,

have a little patience with me. I certainly couldn't interpret your message correctly. What's the problem with starting with a quick charge (0.5C) and ending with a standard charge (0.1C)? Where is the lack of control in this?

[NiHaoMike]

This is the best open hardware NiMH charger I'm aware of:
http://www.ultrasmartcharger.com/

[ogden]

have a little patience with me.

Idea of fuse as current regulator made me impatient, sorry

What's the problem with starting with a quick charge (0.5C) and ending with a standard charge (0.1C)? Where is the lack of control in this?

Problem is that your circuit is voltage, not current source. If supply is not limiting factor and transistor of your charger does not melt, then it can easily put more than 2C current into poor battery and quickly overheat it. Don't reinvent the wheel. Look for existing circuits, pick one that you like, copy it & test. After that modify and improve - if needed. You can test/simulate your circuits without blowing-up transistors or batteries: https://www.analog.com/en/design-center/design-tools-and-calculators/ltspice-simulator.html

[Avelino Sampaio]

Ogden

Sorry, it's not yet time for me to give up until I'm fully convinced of my idiocy.

Benta and NiHaoMike

as I mentioned, my intention with this project is not to achieve maximum efficiency (100%) in loading. Any value above 80% I will be satisfied. I don't care if this will shorten battery life, as long as it's not terribly significant.

[Avelino Sampaio]

The Eneloop

On the eneeloop packaging it says that the batteries were ready for use. I unpacked and measured the tensions. All batteries (AA and AAA) measured 1.31v. To find out how much charge they had, I used my electronic charge, set to 500mA, and tested the discharge until I reached the 1v limit. The AA battery (2550mAh) had a load of 1000mAh and the AAA battery (950mAh) was only 300mAh.

Project (image 1)

I managed to add to the project the part that interrupts the battery charging, when it reaches the limit voltage, which I stipulated at 1.495v.

First battery charge (image 2).

I did the first charging test with 2550mAh eneloop AA, which I had discharged up to 1v. Charging time was 9 hours, until reaching the voltage limit I stipulated of 1.495v

Time to unload and measure

The battery voltage, after 8 hours of rest, is at 1.40v (image 3). I'm going to start the unload test and see how much loading I got.

[ogden]

Time to unload and measure

Not even trying to introduce current regulation. Hopefully others do not learn from you. I am out.

[Benta]

Hopefully others do not learn from you. I am out.

Me too.

[IanB]

Some time ago I did some charging experiments to learn about the characteristics of NiMH cells. Attached is a charging graph for a 1900 mAh Eneloop. You can see that when the terminal voltage reaches 1.48 V in this test, the cell has only received 50% of a full charge. You can also see why charging at constant current and monitoring rate of change of voltage, and/or rate of change of temperature, makes for a sensible charging strategy.

[Avelino Sampaio]

IanB

Thank you so much for bringing this question to me. See in image 1 the graphics that I extracted from the link below (Ansmann).

https://datasheet.octopart.com/5030852-Ansmann-datasheet-5400527.pdf

Note that for Standard Charger mode the ceiling voltage has decreased while for Fast Charge mode the ceiling voltage has increased significantly.   You're absolutely right, if I use the cut-off voltage at 1.495v in Fast Charger mode the charging will be much lower than expected.  See in image 2 the results I got with this project of mine - the AA battery performed well (87.6%) as it spent most of the time in standard charger mode. The AAA battery remained part of the time in quick Charger mode and here I assume that the cutoff voltage would then have to be a little higher than 1.495v, in order to be able to surpass 80% charging.

Image 3 - Some updates to the project.

[Avelino Sampaio]

Hi!

Image 1 - The project is advancing and the results so far are very promising. The project is capable of charging batteries ranging from 300mAh to 2700mAh, with an average charge exceeding 85%.

Image 2 - I added a timing when the voltage reaches 1.495v and with that I managed to overcome the issue of the final cutoff voltage. There was never any change in the temperature of the batteries, as charging always ends in Standard Charger Mode.

It's been a lot of fun putting this together with discrete components. At the end of the tests, I will make a version using a PIC12F.. processor.

[Avelino Sampaio]

Hi

image 1: A few more eneloop's arrived.

image 2: I will start a new version of the project, using the PIC12F.

[Avelino Sampaio]

Hi!

Errors, hits and assembley programming, resulting in another project update.

[Avelino Sampaio]

Hello

I'm managing to carry out a load between 95% and 100%

The programming I am adopting has as its main objective the monitoring of the slope of the voltage curve, and thus getting as close as possible to the top voltage. Monitoring starts from a voltage of 1.48v and every 15 minutes a measurement is taken and then compared with the previous measurement. If the result is less than 3.4mv in the 15 minute interval, it means that we are approaching the top voltage. At this point I decrease the voltage over the battery (from 1.61v to 1.53v) and leave it with a slow charge for another 60 minutes.

[Avelino Sampaio]

Hi!

Battery inversion in socket

I had already added resistors R2 and R12, in order to protect the PIC and Opamp. Note in the image that the current flowing through the internal diodes is below 1mA. They are well within the safety margin. The problem that I didn't count on is that transistor Q2 (BD139) is fully conducting to the point where Q3 is acting and limiting the current to 0.6A. I need to interrupt the Opamp or transistors so that it blocks this strong conduction. Could someone suggest me how to perform this block?

[Avelino Sampaio]

Hi

I managed to solve it, adding a mosfet (Q6).

I finished the PIC programming and the end result was way above my expectations. I will now plan the assembly of the PCB.

[Avelino Sampaio]

Hi

For this project, the 5v power supply will have to show stability under any load conditions, which will vary between 75mA to 2.2A. At first I would use the 5v from an ATX source but I decided to change my mind and use an independent source. I ordered a power supply from banggood (Image1) but the question remains whether I will have the 5v regulated, under any load conditions. While I'm waiting for my order, I found here in my electronics scrap, a 28v+12v source (image2) that has a transformer that seems to guarantee at least 30w of output. So I decided to modify it, taking advantage of the 12v output. See in image 3 the layout of the original output and in image 4 the modifications I made to get the 5v. I've had success with this adaptation, managing to keep the 5v at any required load (images 5 and 6). This source will remain as plan B, if the source I ordered doesn't fit.

[Benta]

It seems you're speaking with yourself?

[SilverSolder]

The best strategy for charging NiCd or NiMH cells is d²V/dt² detection. This gives the cells a long life, I'm speaking from experience, my cells have until now survived 1000s of charge cycles without significant degradation. All other strategies that I've seen result in overcharging and damage.

To understand the idea, p.4 of this data sheet helps:

https://pdf1.alldatasheetde.com/datasheet-pdf/view/26117/TEMIC/U2402B.html

Today, you'd let an MCU with associated ADCs run the algorithm.

That seems an interesting principle, haven't seen that before.  Not sure how the low current phase (after d2V/dt2 detection occurs) is handled, is it just a timer or does it still need -dv/dt to stop charging?

[vk6zgo]

Never ever charge batteries w/o current limiting/regulation

There are "batteries & batteries".
Lead/acid "wet" cells have been charged with simple transformer plus rectifier combinations for decades.

[Avelino Sampaio]

Hi

I finally managed to finish the tests and I'm very satisfied with the results. The font I ordered from banggood met the requirements.

I thank the participants for sharing important information and thus contributing to my learning process.

[Avelino Sampaio]

Hi!

I took advantage of the box of an old stabilizer and thus concluded my project.  My eneloop, at the end of the charge, reaching around 1.48v.