IBM PC AT battery circuit question

[alank2]

When the MB is on and the switch is on, it says it is feeding +5V to the RTC components.  Here is my question - NONE of that turns OFF the battery.  What if the battery is a fresh new battery running at a higher voltage through the two diodes than the +5V ?  They say it is a lithium 6V battery, would that be around 6.4V fresh?  Subtract the 0.8V and you have 5.6V connected to 5V right?  What happened in that scenario?  Did the lithium battery get worn down before its time even when on until it matched the +5V?

[JS]

Doesn't look like so, first there are two silicon diodes, so you are closer to 1.4V drop, so you are about 5V there, then the forward bias of the colector base junction of the 2n3906 and the 10k resistor to ground, but that resistor will see about 4V, so 400μA wich will be coming from the emitter base polarization.

if you started with a full battery at 6.4V, 0.7V at 1n4148s, 5V from bat, 5V from supply so let's say no current flows through Q2 collector. The RTC would be powered from the bat even while the PS is on, wasting battery energy powering the thing while not needed. In any case it should last for years powering the RTC, close to it's shelf life, and not much energy can be drained out of the battery over 6.1V or 6.2V where this won't happen any longer. So you wasted one month ot two of battery life from a few years, not a biggy. This if the thing happend to be powered most of the time during the first few month of opperation, otherwise it would just be the same.

At the end, nothing to worry IMO.

JS

[bitseeker]

Yeah, a bit suboptimal for battery life when the battery is new. Of course, that also assumes that the mains-powered 5V rail is at or below 5V. It might be just high enough (5.1?) not to drain the battery too much initially.

[German_EE]

A tale regarding the battery in the original IBM AT (yes, I am that old)

Years ago I managed to get hold of one of those batteries and there was a label which reminded the service engineer not to pierce the case or dispose of the unit in a fire. Too tempting. I lived in a shared house with a bunch of medical students and one night we had a barbecue in the garden using a real fire rather than an enclosed affair. The food was cooked and consumed, the booze had been drunk, and I suddenly remembered that AT battery down in the cellar. Two minutes later the battery was in the middle of the embers.

The initial results were very disappointing, a brief bit of flame as the plastic case went up, then nothing. We sat there for about a minute wondering what to do next and then it happened.

Remember the scene at the end of 'Close Encounters' when the spaceship opened up and there was that brilliant white light? It was like that, only brighter. There was also an amazingly large bang which left our ears ringing. When our eyes had recovered from the glare we looked around and the fire had gone, there were however dozens of small fires throughout the garden due to scattered embers and LOTS of neighbors looking out their windows wondering what had happened.

It took a chemistry student to explain what had happened (maybe). Either the lithium in the cell had been superheated in the fire then exposed to air when the case melted. or. there was lots of hydrogen produced as the contents heated up, this then exploded once the case was breached.

Either way, it's not an experiment I would choose to repeat 

[alank2]

It looks like getting a replacement 6V lithium battery isn't so easy.  What do you guys think about the 3 x AA lithium route?  The battery curves I saw showed 1.72V fresh for lithium AA.  That would be 1.72*3=5.16-0.8=4.36V.  I'm be concerned that 4 x AA might be too much 1.72*4=6.88-0.7=6.08V.

[Nusa]

A 6 volt lithium photo battery is still easy to find, although anything else you come up with to create the right voltage should work as well. The lithium aspect is for longevity, not necessity. Other chemistry's won't last as long, but still good for many years.

You just have to solder on the connector wires, making sure to get the polarity correct. If the voltage is a tad too high for your taste, include an inline resistor or diode to drop it a bit.

[bitseeker]

It looks like getting a replacement 6V lithium battery isn't so easy.  What do you guys think about the 3 x AA lithium route?  The battery curves I saw showed 1.72V fresh for lithium AA.  That would be 1.72*3=5.16-0.8=4.36V.  I'm be concerned that 4 x AA might be too much 1.72*4=6.88-0.7=6.08V.

Alan, note that in this example and your previous one you're subtracting one diode drop. However, there are two diodes in the schematic. You have to subtract the diode drop twice.

[alank2]

Thanks - I am going by what they say on the bitmap above "the combined voltage drop of these two diodes is 0.8V".

But...now that you bring this up?  My experience has always been less than the 0.8V drop, and is usually 0.3V or 0.4V when measuring various diodes or drops in BJT's.  Why is that?  Is it dependent on the amount of current?

[bitseeker]

It depends on the makeup of the diode junction and, yes, the amount of current has some effect, too. So, let's have a look at the datasheet for the 1N4148.

NXP
https://assets.nexperia.com/documents/data-sheet/1N4148_1N4448.pdf
Page 3, V_F @ 5mA is 0.62–0.72V. Maximum @ 100mA is 1V.

On Semiconductor
http://www.onsemi.com/pub/Collateral/1N914A-D.pdf
This datasheet covers several models including the 4148. Similar V_F to NXP's version.

Maybe the source of the bitmap was looking at a 5170 that used different diodes.

Schematics are like construction blueprints. They don't necessarily match the final product/device/building.

[Nusa]

The clock circuit being powered here should draw microamps, not milliamps. Which is not a range the datasheet speaks clearly about.

Since you have the actual hardware in question, you can easily confirm by simply measuring the voltage drop across those diodes after you've supplied voltage on the battery connector.

[alank2]

The clock circuit being powered here should draw microamps, not milliamps. Which is not a range the datasheet speaks clearly about.

Since you have the actual hardware in question, you can easily confirm by simply measuring the voltage drop across those diodes after you've supplied voltage on the battery connector.

I drove and met a guy to pick it up today and was surprised with the hard drive spun up.  I hope it keeps spinning.  It took a bit longer to spin up the first time I turned it on so hopefully I can bring it back.  Did a low level format on it with 2:1 interleave and fdisk/formatted it and installed PC-DOS 3.30.  I was also surprised that the battery is still working - it says replacement on it, but I don't know how much power it still has left.  I'm going to measure it tomorrow and measure that voltage drop across the resistors as well!

[bitseeker]

The clock circuit being powered here should draw microamps, not milliamps. Which is not a range the datasheet speaks clearly about.

True that. We shall await Alan's measurements.

[viperidae]

The clock circuit being powered here should draw microamps, not milliamps. Which is not a range the datasheet speaks clearly about.

True that. We shall await Alan's measurements.

Look at the forward voltage drop graph in that Fairchild datasheet. It shows Vf goes from 250mV at 1uA to 500mV at 100uA

[bitseeker]

Oh, in the On Semi datasheet. I missed that one (Figure 3). The NXP one didn't have a graph with that much resolution.

So, the circuit in question could draw up to around 12µA and still have a total drop similar to that mentioned in the bitmap (2 x 400mV).

[alank2]

One of them measures 0.222V and the other 0.224V.

[james_s]

A tale regarding the battery in the original IBM AT (yes, I am that old)

Years ago I managed to get hold of one of those batteries and there was a label which reminded the service engineer not to pierce the case or dispose of the unit in a fire. Too tempting. I lived in a shared house with a bunch of medical students and one night we had a barbecue in the garden using a real fire rather than an enclosed affair. The food was cooked and consumed, the booze had been drunk, and I suddenly remembered that AT battery down in the cellar. Two minutes later the battery was in the middle of the embers.

The initial results were very disappointing, a brief bit of flame as the plastic case went up, then nothing. We sat there for about a minute wondering what to do next and then it happened.

Remember the scene at the end of 'Close Encounters' when the spaceship opened up and there was that brilliant white light? It was like that, only brighter. There was also an amazingly large bang which left our ears ringing. When our eyes had recovered from the glare we looked around and the fire had gone, there were however dozens of small fires throughout the garden due to scattered embers and LOTS of neighbors looking out their windows wondering what had happened.

It took a chemistry student to explain what had happened (maybe). Either the lithium in the cell had been superheated in the fire then exposed to air when the case melted. or. there was lots of hydrogen produced as the contents heated up, this then exploded once the case was breached.

Either way, it's not an experiment I would choose to repeat 

I suspect it was probably the latter, the majority of the energy came from a steam explosion which also happened to consist of flammable materials.

I'm reminded of a similar situation years ago when I had some friends over and we were sitting around a fire in the back yard. I decided it would be fun to toss a cup of gunpowder into the fire without telling everyone what I was doing. WHOOMP! There was a bright flash and a small mushroom cloud of smoke although nothing as dramatic as you describe. One guy tipped over backward in his chair and there were a couple of spilled beers.