Mystery of an old H-bridge

I have dissected an old device and found a lot of “ancient” chips. One of them is BA6418N - a H-bridge. I want to use it for something because I love the SIP package (for no particular reason - it simply looks cool). But there is a thing in the datasheet I don’t understand. They claim maximum output current it 0.7A:

But they also say maximum power dissipation is 800mW only (at 25°C).

But from the electrical characteristics the saturation voltage is 1.2V at 200mA. 700mA*1.2V > 800mW, and I did not consider the saturation voltage will be surely higher at 700mA.

What is worse is the supply current. Typical value for Icc_1 (the logic section) is 34mA and the Icc_2 is 54mA at no load conditions. This mean the total current consumption at no load conditions is more than 80mA - the quiescent current of the device alone exceeds the abs. max power dissipation when supply voltage is more than 10V! It can provide no current to the motor despite recommended voltage range is up to 15V. Do you have an idea how is this device supposed to be used? Did I miss something in the datasheet?

BA6418N Motor driver.pdf (61.6 KB)

My interpretation of the data sheet, which has an unfortunate typo, is different.

Supply current 1 is when the motor, if attached, would be running either forward or backward.

Supply current 2 is with both input pins high, brake mode. The motor would provide current only very briefly during braking.

The quiescent current is with both inputs low and both outputs off, and is less than 1.5 mA.

So the package power dissipation would be maximum when the motor is running, and would be typically Vcc*34 mA + 1.2V*(motor mA) in mW.

The power dissipation limit is an average over the PWM cycle, so for low duty cycle the power dissipated is low, and high if high.

The fact the current limit is 0.7A doesn't mean you'd use it at that level, or anything like, in normal operation, just like you don't expect a car to drive at top speed continuously...

For a motor with 0.25A normal load, 0.7A stall, it would be a sensible choice I think.

Sorry for late reply.

@jremington:
Where did you get your crystal ball? I want it too!

In other part of the datasheet they “clearly” define ICC1 as a logic section current and ICC2 as the power section current:

(8) Current consumption
The ratio of current consumption (ICC1 versus ICC2) is 1 : 2 for the logic section GND (pins 2 and 5) and the power section GND (pin 8).

They also define VCC1 as logic section supply voltage and VCC2 as the power section supply voltage int the Electrical characteristics chart.

But despite this you are right - I have measured the total supply current and it is < 1 mA while the outputs are high impedance, 30 mA when the motor is powered to either direction and over 50 mA when braking. The current does not change much with load.

They also define VCC1 as logic section supply voltage and VCC2 as the power section supply voltage int the Electrical characteristics chart.

No crystal ball! There are several obvious problems with the documentation, so I understood it to be unreliable and open to interpretation. The one I posted is the only one that made sense.

Since there is only one Vcc pin, the distinction between Vcc1 and Vcc2 is simply wrong. My guess was that someone sloppily edited the documentation for a different chip to produce this one.

jremington: Since there is only one Vcc pin, the distinction between Vcc1 and Vcc2 is simply wrong. My guess was that someone sloppily edited the documentation for a different chip to produce this one.

There is single Vcc pin but 3 ground pins. As I understand it 2 of them are for the logic (pins 2 and 5) and should have the same level. The last one (pin 8) is for the power part and intended to reduce the speed of the motor by increasing this voltage. The Vcc1 and Vcc2 definitions make sense (as difference between the common Vcc pin and different "grounds" as defined in the table).

I do not agree with that interpretation of the data sheet. The application example shows all three "ground" pins connected to the same ground.

Who knows what the chip designers actually intended?

The logic ground is not connected to the others internally according to the equivalent circuit, which makes sense as you don't want to inject ground noise into the logic circuit - Its not clear if the logic ground can be a substantially different voltage though, or is just intended to be separate for noise-spike purposes only.