Accurately measuring voltage on a power supply

[dnix71]

I am trying to modify some Black and Decker chargers originally made for an 18v nicad pack. One of the supplies is transformer based and the others are switching supplies. Now I know this is B&D we are talking about, but these supplies seem to have no regulation at all.

The no load voltage on the transformer based supply measures 21.7 using a cheap Harbor Freight multi-meter, but jumps to 33.8 volts if I put a 470uF cap of the output to smooth the ripple. A Tenma 72-7720 meter says 22.7 no load and 34.3 with the same cap.

Oddly enough the no load and capacitor smoothed voltages on the switching supplies is the same for both meters and with and without the capacitor, however the rated output for three is 24v, where the measured is 28.0, 28.8 and 29.6. The 4th switching supply is rated 26v but is measured at 24.7. Stranger still is that if I plug those chargers into a lithium pack they will severely overcharge the packs to about 2.8 volts above the respective measured outputs and do it quite quickly..

I want to charge a 6-cell 18650 3.7v each lithium pack, so the "real" voltage matters. It took 4 diodes facing forward to cut the voltage on one of the switching supplies down to a level that was safe for the lithium pack. I left it plugged in to 2 lithium packs for 8 hours and neither went overvoltage on any cell.

It shouldn't take an oscilloscope to figure the voltage a supply is putting out. As it is, I have to improvise with diodes to drop the finish voltage to something safe for my packs. Nicads won't explode and burn like lithium, but the still burn up and out prematurely.

Any suggestions?

[Flux]

NiCd are current charged so you would need to measure current to make any sense.

The transformer based units will be virtually useless for Lithium without a separate regulator. The switching ones may be modifiable to regulate at lower voltage but without circuits you are guessing.

Diodes do not have a sharp knee for precise dropping of volts, zeners would be better but a variable regulator chip added to the output would give much better control.

Rough methods may be ok for quick charging if you don't leave them on too long but if you go away and leave them you need precise voltage or have a big safety margin and work well below full charge

A NiCd charger is not a good starting point for Lithium ion as the charging process is very different.

Flux

[OperaHouse]

I get all my Li from computer packs so haven't cared that much about charging them right.  One thing I have noticed is the charging circuitry has more computer power than the lunar lander did.  Without balancing you will eventually be in trouble.  That said, I might do a TL431 voltage monitor driving a relay.  When the final voltage was reached, the relay would latch on and disconnect the battery.  At that time, it would be on float charge by a simple dropping resistor.    There was a good post on EEVblog  where a guy charged up batteries with several smart chargers and performed discharge tests.   Even with the same charger results varied greatly due to reaching trip point too soon.  NiMH batteries but a good read     http://www.eevblog.com/forum/projects/battery-capacity-measurement-gives-wildly-varied-results/msg528923/#msg528923

[dnix71]

For the intended purpose, a high current weed whacker, nicads are the best choice. But they just fail without warning because the current is so high it flips/shorts the weakest cells during a hard discharge. With a protect board on each lithium cell, those are less likely to die suddenly.

A better solution might be a dc-dc voltage booster, but that would be costly and bulky. A parallel stack of lithiums, boost the voltage by 5 or 6 and no cell would fail because of current pushed through it by it's stronger pack mates.

[joestue]

If you really want to potentially create a disaster, wrap buck converters around each battery, like this
http://johansense.com/bulk/bms_1.1.png

You then drive each fet with a 50% duty cycle fed from a gate drive transformer.. yeah, i know you need a metric crap ton but if you can find a half bridge driver that will work with just 7 volts (two lithium cells) that is another option.

Unfortunately it takes quite a bit of brains to measure the current and send it back up or down the food chain to monitor which cell is weak.
I recall a 6 cell bms that uses coupled flyback converters to dump power from the strongest cell and push it out the battery pack, but its not able to push power into the weaker cell.

[dnix71]

I got rid of gasoline powered lawn equipment to try and simplify my life. I use a manual reel push mower and edge the sidewalk with a flat half-round spade. There are areas where I can't use a swing blade to edge and manual shears/scissors on my knees is hard at my age.

For me to need a computer controlled battery charger just to whack the weeds just seems wrong.

[joestue]

For me to need a computer controlled battery charger just to whack the weeds just seems wrong.

I think my rototiller tines are air hardening tool steel.
i'm pretty sure in a previous age such a thing would be regarded as unobtainable, much like the crucible steel that the vikings bought from india during the dark ages.

-back on topic, it would be great if someone made a 4.1 volt zener with a knee as sharp as a razor blade.