Another La Crosse Meltdown

[willchueh]

Hence the reason why I'm now running my two BC-900s on a 2.8V power supply. See this thread.

jtr1962,

I believe you got the right idea. However, there is a slight problem. BC-900 uses a single output power supply. When you reduce the output voltage of the PSU, you are simultanesouly reducing the voltage to the MOSFET and the voltage to microcontroller.

The BC-900 uses a form of inexpensive mounting called Chip-On-Board (COB) to cut cost. If you've seen the guts of the charger, it is under the black epoxy. It is basically a silicon wafer without a case. These types of ICs are pretty sensitive to input voltage. Without knowing the exact specs, your supply voltage to the IC could be below the required voltage and may cause stability (latching) and reference voltage problems.

All the chargers we design almost always have a secondary voltage regulator (1% regulated output) right before the controller IC. However, I believe the BC-900 lacks this.

The best way would be to just reduce the voltage to the MOSFET but not to the main IC. I would imagine that, however, it is not an easy modification.

William

 

[jtr1962]

Without knowing the exact specs, your supply voltage to the IC could be below the required voltage and may cause stability (latching) and reference voltage problems.

I've had the chargers running continuously refreshing batteries for the last week to see if there are any problems. So far, so good. The voltage readings still seem correct even at the reduced power supply voltage so the lower voltage isn't causing problems with the reference voltage.

Yes, you could reduce the voltage to the MOSFET but not the main IC by putting a Schottkey diode in series with the MOSFET drain. Of course, that will still result in the same amount of heat being dissipated on the circuit board even if it is spread among two components instead of one. I consider the plastic case too small and poorly ventilated to dissipate that amount of heat.

BTW, some here have said that the voltages displayed by the charger are incorrect. To a first approximation this was true unless the charger was modified with added bypass capacitors. See this thread for details. However, once modified the voltages seem accurate to within 0.01 or 0.02 volts. The problem is that you can't simply measure the voltage at the battery terminals while it's in the charger and compare that to the value displayed by the BC-900. Since the BC-900 pulses the charging current, it measures battery voltage during the time when the MOSFET is off so as to get a reading independent of voltage drops in the contacts and PC board traces. However, if you try to measure the voltage at the battery terminals with a multimeter you'll end up with an average of the voltage while charging and the resting voltage. The readings will be higher than what the BC-900 displays if it's charging, and lower if it's discharging (yes, it pulse discharges as well). I just thought this info would be useful. Also note that it's important to measure the battery voltage right after you pull it out of the charger since it will immediately start to drop once it's no longer receiving charging current. I find that if I measure within 10 seconds of pulling the cell the measured voltage pretty much agrees with what the BC-900 displayed.

 

[willchueh]

The problem is that you can't simply measure the voltage at the battery terminals while it's in the charger and compare that to the value displayed by the BC-900.

Engineers usually refer this voltage as the "off-line voltage." A quick and dirty way to see this voltage is simply connect a scope to the battery terminal. You'll be able to observe the up and downs in the voltage.

The only tricky thing is a battery is typically modeled as a resistor and a capacitor in parallel (plus other stuff) so there is a capacitive discharge effect when the current pauses.

William

 

[wptski]

I had four "HR" stamped made in Japan 2500mAh Energizer AA cells so I discharged two at a time on a Triton and ICE charger down to .9V/cell, so each cell isn't really equal. I have two cells charging at 200ma on a v33 BC900 monitoring the temperature on both with K-type bead probes on a Fluke 54-II.

Since I don't want to leave it unattended, I'll only be able to let it charge for a max of 12 hrs. if it don't terminate or melt by then! I'll be able to post a graph tomorrow evening but right now after almost four hours, they are at 86F which is 8F higher from the start.

 

[evan9162]

I can't believe that LaCrosse is using an SOT-23 package to disspate upwards of 1W. I wouldn't dare touch that level of power dissipation in a package smaller than SOT-223, and would rather use a DPAK or D2PAK instead. Sounds like a poor design decision.

 

[wptski]

I haven't looked at the log so temperature maybe a bit higher! One cell finished charging this morning at 11:12, 2150mAh at 101F. The other cell was still charging at 12:01, 2310mAh at 97F. So the calculation above for 100F at 200ma was accurate.

 

[wptski]

I didn't post a graph from the 200ma temperature test because there wasn't anything really interesting to look at.

Below is a temperature graph of two 2.5Ah Energizer AA cells again in Slots 1/2 but at 1A charging rate. The green dash line is Slot #2 and the blue line is Slot #1. Notice how Slot #1 turned off for high temperature lower than Slot #2 was allowed to go and finish the charging cycle. Slot #1 cooled down and restarted. It finished charging but was allowed to go to higher temperature which had cut it off earlier!! No fan is used.

Looks like I cut off the time scale words on the pic, it's 1,2,3 hrs.

 

[koala]

Stack em up! Double/parallel the tiny mosfets . Maybe a tiny heatsink will cure it...

I don't mean to bash BC-900. The electronics in it is a joke. It's the cheapest looking circuit board I've ever seen. It seems like the whole charging system depends alot on the software.

I'll try to find out what's the max operating temperature of the mosfets in BC900 running at this week.

Question: I have an Infrared Thermometer I use it for everything. I find it pretty accurate and very effective. Is there anything I should be aware of as I will be pointing it directly on the tiny surface on the mosfets of the BC900.


Vince.

Tom,

Frankly speaking, I believe LaCrosse used the improper MOSFET packaging (SOT23) for this application. Most SOT23 are only rated for Pd = 1.0 ~ 1.25W. Given their linear design, the best case efficiency = (Vin - Vout)/Vin.

The best bet would be to replace the MOSFET with a bigger packaging ones, say SO-8 or TO-252. But there might be problem fitting these components in.

William

 

[wptski]

Stack em up! Double/parallel the tiny mosfets . Maybe a tiny heatsink will cure it...

I don't mean to bash BC-900. The electronics in it is a joke. It's the cheapest looking circuit board I've ever seen. It seems like the whole charging system depends alot on the software.

I'll try to find out what's the max operating temperature of the mosfets in BC900 running at this week.

Question: I have an Infrared Thermometer I use it for everything. I find it pretty accurate and very effective. Is there anything I should be aware of as I will be pointing it directly on the tiny surface on the mosfets of the BC900.


Vince.

You might want to experiment on something of similar size. The low end IR temperature probes have 1" spot at 8" most often and they suggest staying at least 8" away from the target.

 

[tacoal]

The MOSFET used in BC-900 is NTGS3443, 2A, 20V.

 

[jtr1962]

That MOSFET would be fine if they were using a switching type circuit. For the linear circuit in the BC-900 it's marginal at best.

BTW, both my BC-900s have been happily operating on ~2.8V with the modified supplies for 11 days continously now. I haven't noticed any problems at all. The only thing I've noted is that all the batteries I've refreshed are reading lower capacity than the last time. Some are as much as 25% less. I'm guessing the warmer room temperature in my work room is the culprit (80+ °F versus ~65°F last time). I noticed the same thing even with the stock supply so it's not a voltage reference problem.

 

[wptski]

That MOSFET would be fine if they were using a switching type circuit. For the linear circuit in the BC-900 it's marginal at best.

BTW, both my BC-900s have been happily operating on ~2.8V with the modified supplies for 11 days continously now. I haven't noticed any problems at all. The only thing I've noted is that all the batteries I've refreshed are reading lower capacity than the last time. Some are as much as 25% less. I'm guessing the warmer room temperature in my work room is the culprit (80+ °F versus ~65°F last time). I noticed the same thing even with the stock supply so it's not a voltage reference problem.

Your saying that higher temperature gives lower capacity would go against the grain of RC'rs wanting to run hot packs. Some chargers like the ICE has provisions to allow timing so packs are hot just when they need to be used.

So, what about using a fan likes some are doing? Does that cause less capacity too?

There's some odd stuff about high temperature cufoff too because form my graph posted above you can see that it cutoff at a lower temperature once but allowed a higher temperature when it neared termination.

 

[evan9162]

The MOSFET used in BC-900 is NTGS3443, 2A, 20V.

unless there are huge copper pads coming off the leads of this device, it's maximum power disssipation is just 0.5W.

 

[jtr1962]

Your saying that higher temperature gives lower capacity would go against the grain of RC'rs wanting to run hot packs. Some chargers like the ICE has provisions to allow timing so packs are hot just when they need to be used.

I don't know why my cells are coming out at lower capacity. I thought higher temperature might be a reason.

There's some odd stuff about high temperature cufoff too because form my graph posted above you can see that it cutoff at a lower temperature once but allowed a higher temperature when it neared termination.

That's when it's charging. When discharging even if the cells are warm at the end of charge they'll cool off fairly quickly with or without a fan. Yes, I use a fan which is on all the time in order to keep my cells from overheating. I do this with any charger capable of charging over 500 mA, not just the BC-900. I just don't trust the charger to shut down in the event the cells overheat. Also, I modded my BC-900 to have a lower cutoff temperature because the cells could get too hot with the stock setup. If I don't use a fan at the higher charge rates the charger would go into thermal shutdown long before the cells were charged, and the charging would take much longer to complete.

 

[willchueh]

It would be interesting to monitor the MOSFET package temperature. Then compensate for the thermal gradient across the junction and case. Compare that to spec to see if it's close to the "kill zone."

William

 

[willchueh]

The MOSFET used in BC-900 is NTGS3443, 2A, 20V.

Thanks tacoal for the exact packaging information. Based on the spec, the maximum disspation for minmal thermal padding is only 0.5W. Even if the thermal padding is improved, at best it can take 1W (and quoting the spec, this requires "package mounted onto a 2 in square FR-4 board with 0.06" thick cupper, which I doubt there is room in the BC-900).

William

 

[wptski]

I don't know why my cells are coming out at lower capacity. I thought higher temperature might be a reason.


That's when it's charging. When discharging even if the cells are warm at the end of charge they'll cool off fairly quickly with or without a fan. Yes, I use a fan which is on all the time in order to keep my cells from overheating. I do this with any charger capable of charging over 500 mA, not just the BC-900. I just don't trust the charger to shut down in the event the cells overheat. Also, I modded my BC-900 to have a lower cutoff temperature because the cells could get too hot with the stock setup. If I don't use a fan at the higher charge rates the charger would go into thermal shutdown long before the cells were charged, and the charging would take much longer to complete.

By my graphing, it takes about 15-20 minutes to resume charging after a temperature pause.

 

[Bones]

Bah, wrong thread, but interesting reading nonetheless...

 

[alphazeta]

......