BC-900 power supply discovery

I was recently refreshing some solar garden light cells with my pair of BC-900s when a turn of events led to an interesting discovery. I was running one charger on the stock power supply and the other on this supply. I had already done all of the mods described in this thread, as well as the temperature sensor mode described in this thread (see posts 38, 40, and 41). Anyway, when charging some cells at 1000 mA (with a fan blowing on them, of course), the current on one of the charging stations dropped very low, and the bottom of the charger right under it was extremely hot. I removed the cell right away. Apparently the MOSFET had gone into thermal runaway to the point that with its increased internal resistance it just couldn't deliver the set current. Surprisingly, the charger worked fine upon cooling so the damage wasn't permanent but some have had their chargers go into complete meltdown. I did have to fix a melted button shaft however.

Anyway, the problem occurred on the charger with the alternate supply. As mentioned, the circuit board does get notably hotter when using this supply as opposed to the stock supply. However, the charger gets too hot on 1000 mA even with the stock supply. This got me thinking that maybe the BC-900 can happily operate at less than the 3 volts of the stock power supply so I started experimenting. To make a long story short the BC-900 will happily operate at 2.5 or so volts while still being capable of delivering 1000 mA. Any less than that and it works, but can't provide 1000 mA to the cells (although lower current settings work fine).

Armed with this information, I decided to see if I could mod the alternate power supplies. Yes, they were sealed and a pain to open, but open them I did. :devil: I found the area of the circuit board which regulates voltage. The mods were actually quite simple. First, I placed an 82K SMD resistor on top of a 12K one. Since this is a switching power supply with no minimum current it already has several load resistors so that it can regulate with no attached load. Unfortunately, these weren't enough to allow the power supply to regulate at the lower voltage so I had to add 6 68.0 ohm resistors placed on top of the existing load resistors. Of couse, no project is complete without changing an LED so I exchanged the stock dim yellow-green power on indicator LED for a much brighter ChiWing true green one. With these mods completed, one supply read 2.78V and the other 2.84V. When the BC900 was charging 4 cells at 1000 mA the power supplies delivered about 2.6V at the charger due to the voltage drop in the rather long wire. Even at 1000 mA the bottom of the charger barely felt warm. I'm not sure but I also think that the charge termination is working better with less heat on the board. After all, higher temperatures make more noise in analog circuits. I'll have to try my chargers without the fan at 1000 mA to see if the charge terminates before thermal shutdown occurs.

While unfortunately the alternate power supply is sold out (at least by the vendor I used), the idea if anyone else attempts this is to aim for a voltage of about 2.6V at the charger while charging 4 cells at 1000 mA, and to mod whatever power supply you're using appropriately to deliver this.

Edit: I found the alternate supply available for sale here.

I have one of the alternate supplies but never used it. Away form home now and can't compare. Why does the alternate cause more heat? Was a higher voltage that it can't contend with??

wptski said:

Why does the alternate cause more heat? Was a higher voltage that it can't contend with??

Click to expand...

My guess is that the BC-900 uses a linear constant current regulator which in turn is pulse width modulated to average out to the set current. I have no idea exactly what value this current is, except that by definition it must be higher than 1000 mA. Let's say for argument's sake that the current is set for 2000 mA (just an educated guess). Furthermore, let's assume that the voltage on a NiMH cell charging at 1000 mA is 1.5V. With the stock power supply the regular must absorb the difference between 3 volts and 1.5 volts while supplying 2000 mA. This produces 1.5x2 or 3 watts of heat. However, since the 2000 mA current is only on 1000/2000 or 50% of the time the average heat is only 1.5 watts. Some of this is dissipated in the current sense resistor and the rest in the MOSFET. Let's say that the current sense resistor is 0.25 ohm (I remember seeing 4 1 ohm resistors in parallel). Therefore, at the 1000 mA the current sense resistor dissipates (0.25 x 2² x 0.5) or 0.5 watts while the MOSFET dissipates the remainder of the heat, or 1.0 watts. Note that the power in the MOSFET plus the sense resistor totals 1.5 watts as calculated earlier.

Now let's say that you increase the supply voltage to 3.3 volts. Since the sense resistor by definition can't change in value it still dissipates the same 0.5 watts as before. However, the total power that needs to be dissipated by the MOSFET and sense resistor has increased to (3.3-1.5)*2*0.5 or 1.8 watts. This means that the MOSFET must now dissipate 1.3 watts instead of 1.0 watts. Remember that given the small size of this part it was marginal even with the stock supply. Now it can't cope with the heat at all.

However, consider what happens when the supply voltage is reduced to 2.6 volts. The total power to be dissipated drops to (2.6-1.5)*2*0.5 = 1.1 watts. The resistor still dissipates 0.5 watts as before which means the MOSFET must dissipate only 0.6 watts. In practice the contact resistance and PCB trace resistance absorb some of this heat so the MOSFET actually dissipates even less. Also note that if we continue dropping the supply voltage at some point we won't be able to supply 1000 mA even with the MOSFET on all the time. I figured based on my tests that maybe the sense resistor, MOSFET, PCB traces, and contacts total roughly 1 ohm. 1000 mA flowing in one ohm is a voltage drop of 1 volt. Added to the battery voltage of 1.5 volts means that at a supply voltage of 2.5V or less the charger just can't supply 1000 mA.

I'm not sure of the exact values given here but this is roughly the way the charger works, and why it makes more heat with the alternate supply (and less with the modication).

Thanks! I posted some graphs of the PWM current output on my v32 BC900 before it melted!

You know the Lacrosse engineers(if they have any) could be laying back on their chairs collecting information off cpf rather than working on it. Good find there jtr.

jtr1962: I'm very interested by your findings on using a lower power supply voltage into the BC-900.

This may be a naive question - I've only a passing knowledge in electronics - but I was wondering about just putting a medium power (say 6A rating) silicon rectifier diode in series with the 3.3v output of the alternate supply? Would this be a viable alternative to dropping the voltage to about 2.7 volts?

Thanks in advance for any advice on this,

Regards,

David.

Hello Jtr,

Do you still have the 1500 and 1800 mA charging rates with two cells?

This is a very interesting development.

Tom

May 25, 2018

Details on power supply lower voltage output mod

Hello,

Last week I dug-out my old La Crosse BC-900 charger which was giving me constant reset after years of use. I bought in way back in 2007 while still working in Texas & around 2015 back now in the Philippines the unit started to reset everytime I charge my any cell.

Last year (2017) I read somewhere in the web that adding a tantalum caps at the input jack on the charger motherboard would fix the problem. I soldered a 22uf/35v tant (the one I have lying around) & indeed it seems to fix the problem : no more reset,the charger runs cooler even at 1000mAh charging current but I was getting an erroneous reading after charging : like my AA batteries where reading more than 3,500mAh capacity after charging!

So back it went in my storage...

A few days ago I saw this video in YouTube ===> https://www.youtube.com/watch?v=WoAISvO-tek&t=93s

Out went the charger & back on the operating table...

I cut-open the power supply/adaptor unit & indeed the 3 electrolytic capacitors where bloated.

I replace those 3 caps with a G-Luxon low ESR caps from an old PC motherboard & removed the tantalum cap

Problem seems solved...charging warm to hot again & capacity reading back to normal. However I still get a reset now & then. Perhaps due to the 2nd-hand/used elec.caps?

The other day while test charging my Eneloop cells the charger reset again so I decide to solder a 0.1uf/75v ceramic caps at the input jack of the charger motherboard (where I soldered the tantalum caps in 2017) - the result was worse : now the charger keeps resetting like it did before in 2015. I remove the ceramic capacitor & the charger behave well again but as usual it runs hot if charging at 1000mAh or higher.

Today I decided to resolder back the 22uf/35v tantalum caps on the same location; perhaps this will correct the intermetent reset. When I tested just now the reset problem came back! Comfounded it!

Now I want to try modding the power supply unit to a much lower voltage output

My La Crosse BC-900 is v33 (purchaed in 2007 or 2008) & the adapter has NO (IV) on it. Unloaded output of the adapter is 3.05v & while charging it fluctuates from 2.89v to 2.94v (measured at the input jack of charger motherboard)

Anyone here still got any patience or ideas on how to solve this pesky problem? Replacing this unit is not a solution as I'm in the Philippines now...

regards

Dennis/dmcbohol

I think you are doing a good job getting to the bottom of this problem. If you have a bench power supply, you might try to power the charger with it and see what results you get. That way you could adjust the power to a slightly lower level if you want to experiment.

Back in the days of the VGA monitors, I use to repair them as a side business and bad electrolytic capacitors was the first thing I would look for. The electrolytic solution would dry out and then you would end up with a series resistance with the capacitance. It would not happen as often if the manufacture would use good quality caps. Once there is a series resistance in there, that would cause them to heat up and they would eventually bulge. I have an ESR capacitance meter and so that helps to identify which ones are bad. I never bothered checking the capacitance......... just the ESR.

Placing the 22uf cap on the power line at the input of the BC-900 is a good idea. The larger the better. You have the power supply and then you have several feet of cable. So I would think its necessary to have a cap at the end of the cable. Also if you know where the power enters the microcontroller (probably a black blob) you could place a 0.1uf ceramic cap to bypass the power to it. The reset seems like the voltage drops out too low at the microcontroller. Now from what I remember, the means by which the LaCrosse charges is via pulses to the battery. So you're going to have some surge current. A separate cap at the entry to the switching mosfet for each bay would also help alleviate voltage drop at the microcontroller.

Just some thoughts but I know that's easier said than done .

I had a BC-900 once, but later I bought a BC-700 and liked it better and so the BC-900 didn't get used anymore and I eventually sold it. I didn't like the delay and how hard the switches were to press. The BC-700 works better. I digress. Now I was always suspicious of charging 1000mA on that unit........especially with all 4 bays. And I am also suspicious with charging all 4 bays at 700mA on the BC-700. When I saw the electronics in there (and if there is any heat issue), I don't think the margin of safety with component derating is adequate with the LaCrosse chargers. That is just my opinion though. I limit max charging with all 4 bays at the same time to 500mA charge rate.