grid charger/balancer

[Mr. Mik]

Originally Posted by Mike Dabrowski 2000

MR Mik,
I can see by your questions that you need some lessons in basic electronics, and this is not the purpose of this thread. Get a good basic electronics book, and experiment with what you learn, and eventually it will all become clear.
It all gets down to ohms law.
Fully understand that formula and how op amps work, and you will understand how the value of the 4.7 M resistor will effect things.
No shortcut to understanding.

I did buy a book today! And a bunch of capacitors and other assorted bits and pieces.

Have not read the book yet, but played around with massive hysteresis some more and found it very useful indeed!

Adding a 1000uF electrolytic capacitor between the pos and neg outputs of the 6V DC helped against the on-off oscillations to a degree. But not enough to prevent continuous on-off oscillations with only a 4.7MOhm resistor for hysteresis.

Adding a 100uF electrolytic cap across the PTC made it all a lot worse! Changing the resistance of the simulated PTC strips (a 25 turn 500 Ohm trimpot and another resistor) was just like a smooth dimmer for the LED! I don't understand it, will need to read that book! Maybe I misinterpreted what "across the PTC" means?

Similarly, a 100uF electrolytic cap on the setpoint potentiometer centre tap made things worse, but the "dimmer" function was nowhere near as smooth as with the cap across the PTC strips.

I did not bother trying a cap on the centre tap for the "Warning" setpoint. I like it that it blinks for a while before going solid.

So it was back to trying out lower value resistors for more hysteresis.

I ended up putting a 560KOhm resistor in parallel with the 4M7 resistor, and it works BEAUTIFULLY!

The effect of this massive hysteresis is such that I can still get the shut-down to happen within 10 Ohms PTC resistance (or less if I wanted) of the warning light coming on solidly. About 15 Ohm from the warning light beginning to blink.

And, once the LM324 switches with this amount of hysteresis, it stays switched and that is very good for this particular charger. It should always stay turned off for a minute or so after it has been turned off, to avoid the risk of excessive voltage across the motor-run capacitor. If the sine wave was at one extreme during shut-down, and is at the other extreme during turn-on, then there could be almost 600V potential difference upon turning back on. The resistor across the motor-run cap serves to discharge the cap to avoid this. But it takes time!

And, the best part, the massive hysteresis causes the shut-down to persist past the warning level. Once the charge current has been turned off, the continuing cooling impeller action cools the battery down. The massive hysteresis causes the shut-down part of the circuit to hang on to the shut down state for about -10 Ohm past the warning point.

So the shut-down lamp stays on, and the charge current off, until the temperature is again lower than the warning temperature level set.

Only then will it begin charging again. I think this is very handy and good for the batteries.

With little hysteresis the charger would always turn on again as soon as the temperature has fallen just a little bit, effectively charging the battery as closely as possible to it's set "too hot" point. Large hysteresis gives the battery a breather when it has become too hot and only commences charging again if the temperature falls low enough so that the conditions are favourable for EQ charging.

Here is the schematic:

The problem stopping the charger altogether (yesterday) turned out to be a snapped-off PC-stake, caused by all the wiggling about during installation of the comparator circuit into the somewhat crowded box. It's all working now!

The next step will be to try it out on a real NHW10 battery again, there might be more surprises and challenges ahead...

[Mike Dabrowski 2000]

Rush
The LED is not really required, and after blowing out 2 on my prototype charger, I am going to use a slightly different system. A standard led any color, and a parallel resistor. The resistor will pass most of the 350ma to the batteries, and the led will light up showing that current is flowing. I just need to play with the resistor values to determine the best combo.

MrMik
The LM324 like all op amps can operate as a comparator, but you should also look into true comparator chips like this one.
LM311 - Voltage Comparator

they have more flexible input and output setups, and many have special internal hysteresis circuits, and usually have an open collector transistor output so they can safely drive a relay or higher voltage device than the power supply for the comparator inputs.

You must have noticed that as you increase the hysteresis the difference between the set-point and the actual turn off point got bigger. this makes it difficult to simply set a value and have the turnoff happen at that value, but if you adjust the setpoint dynamically what you have will work fine. 500K is still a reasonable number, but the real issue is that you must have a lot of HF noise in your setup.
A look at your schematic shows why you are still having issues.

1. The capacitor across the setpoint pots should be cap - to negative 6V, cap + to pot wiper
2. The cap across the PTC strip wants to be cap- to negative 6v, cap + to junction of r5 and r4.
3. The lm324 is rated for 25ma output current. pin14 is driving an led and a relay, and may be exceeding its max current which will make the circuit unstable. I much better way to drive the relay is to drive it with a transistor that the op amp turns on, so the relay current is handled by the transistor.

adding more hysteresis of course will let it switch cleanly even with the noise, so if it is doing the job leave it as is.

The accuracy of switching will be effected by the actual 6v power supply voltage, so for a more accurate system, one must regulate the voltage feeding both the PTC strip voltage divider and the setpoint pots.
keep playing you will understand it all eventually.

[Mr. Mik]

Originally Posted by Mike Dabrowski 2000

...
...
The LM324 like all op amps can operate as a comparator, but you should also look into true comparator chips like this one.
LM311 - Voltage Comparator

That's something for much later on, I think! The LM324 will keep me busy for quite a while.....I like the solar tracker you built with it.

As you know, I'm a bloody beginner with this and the more a single device does, the more it will confuse me!
...

You must have noticed that as you increase the hysteresis the difference between the set-point and the actual turn off point got bigger. this makes it difficult to simply set a value and have the turnoff happen at that value, but if you adjust the setpoint dynamically what you have will work fine. 500K is still a reasonable number, but the real issue is that you must have a lot of HF noise in your setup.
A look at your schematic shows why you are still having issues.

1. The capacitor across the setpoint pots should be cap - to negative 6V, cap + to pot wiper

Aha! I might try that.

2. The cap across the PTC strip wants to be cap- to negative 6v, cap + to junction of r5 and r4.

I had it connected that way, but made a mistake in the diagram. That's the one that caused the smooth dimmer function.

3. The lm324 is rated for 25ma output current. pin14 is driving an led and a relay, and may be exceeding its max current which will make the circuit unstable. I much better way to drive the relay is to drive it with a transistor that the op amp turns on, so the relay current is handled by the transistor.

Are the specs for current per output or the sum of all 4 outputs?

The relay powered by pin 14 is a solid state relay with very little current draw. Additionally, it has a 20KOhm resistor in series and cannot draw more than a few uA.

But indeed, I did not pay enough attention to the current ratings. I figured that the 900mW rating means I'm staying within specs, but of course there might be a maximum current rating when it is not used at 30V. Another beginners error! Need to read that book!

The red LED has a forward voltage drop of 2.2V.

6V - 2.2V = 3.8V

3.8V/120 Ohm = 32mA

The yellow LED has the same specs.

That means the current draw is 64mA when both the yellow and the red LED are on. The maximum current rating from the spec sheet seems to be -60mA (at 15V) and short circuit current is +/-60mA at +5V / -5V . I might read these specs wrongly, of course!

But anyway, it's too close to the maximum rating and I better put more resistance in the path of these two LED's!

Increasing the series resistor on each LED to 240 Ohm results in:
(3.8V/240 Ohm) x 2 = 32mA combined; right on the "typical" specification.

I hope the LED's will still be bright enough, because I use them for illumination of the panel meters and would like to be able to read them through tinted car windows in daylight. A locked car is often the safest place to charge these batteries. But the green LED will do this without causing a load on the LM324, anyway.

adding more hysteresis of course will let it switch cleanly even with the noise, so if it is doing the job leave it as is.

The accuracy of switching will be effected by the actual 6v power supply voltage, so for a more accurate system, one must regulate the voltage feeding both the PTC strip voltage divider and the setpoint pots.
keep playing you will understand it all eventually.

I really like the effect that triggering the "Too Hot" state will not allow charging to re-commence until the battery has cooled below the warning level. So I'll keep the hysteresis high for now.

Thanks again for all your help and patience with this!

[retepsnikrep]

Some very interesting info on cheap switch mode psus and modding them for chargers.

I'll be looking at modding some to use with my own batteries.

Endless-sphere.com • View topic - Making a CC/CV charger

Endless-sphere.com • View topic - Switch mode supply mod

[buzzkill]

Originally Posted by Mike Dabrowski 2000

The grid chargers are actively being worked on, and the code development is started, but new projects keep coming up.

I got a call from the WWU x-prize team last night, and they want to drive the complete x-prize car, trailer and crew out here to get the intermittent IMA problems sorted out.
They will spend several days if required.
Western Washington University | Progressive Automotive XPRIZE

How can I refuse a real consulting job on a state of the art scratch built car.
The car has a big lithium pack, and an additional electric motor after the clutch for pure EV mode operation, along with a full Insight drive system.
Until then, I will see how far I can get with the grid charger design and fabrication.
Thanks for being patient.

never saw you post anything else about the car set up they brought you. how did that all turn out?

how are the chargers coming? im debating on whether to do the mima grid charger or both. all depends on income

[Mike Dabrowski 2000]

Chargers are moving right along. I have the LCD displays, AC power cords, and several other parts on order, and am working on learning a new code development system, using C instead of assembly, which is where another of the holdups are. The chassis design is nearly finished but I have to have the LDC in hand to figure out the best way to mount it with the rest of the controls.

I put up a short blog with photos of the xprize car.
MIMA Honda Insight Modified Integrated Motor Assist - Western Washington University X-Prize car
We got the IMA to work normally, and they are still in the finals.

[Cobb]

I know this is an Insight First Gen thread, but with all my funny behavior I have experienced, I think my i2 can benefit from it. Id love to have one when its ready.

Id also be willing to test a booster too.

[thedhgroup]

I think the voltage is different between the I1 and I2 isn't it? What kind of odd behavior are you experiencing?

[uhtrinity]

Originally Posted by Cobb

I know this is an Insight First Gen thread, but with all my funny behavior I have experienced, I think my i2 can benefit from it. Id love to have one when its ready.

Id also be willing to test a booster too.

No reason a charger wouldn't work with an Insight 2, but the warranty would be voided. As far as voltages the Insight 2 nominal pack voltage is 100 volts vs 144 volts for the mk1. The charger should actually be cheaper. Same with any kind of booster technology.

[MN Driver]

Cobb, any funny battery activity would be something that I'd let slide, if possible, until the battery warranty is up. If the dealer seems to think you've been monkeying with the internals anywhere in the IMA system and tells Honda, you could easily set yourself up for trouble if the IMA pack ever has problems. Considering you have the first model year of a new generation where the BCM/MCM might later have updates to it to improve its behavior, I'd leave things alone.

...after the warranty is up, I figure that two appropriately set 48 volt supplies plus the rest of the equipment would do well, so you could skip out on two of those units, although those don't cost much so you might not save much money without them.