Home         What's New         Products         Manuals         Technical Documents         Contact Us         Support         Links         Buy         Dealers    

View as PDF

Ground Bounce



Ground bounce is the name of unwanted noise in a signal that is caused by ground currents. Any noise caused by unwanted current in sense leads can be called ground bounce even if the current is in positive wires.

The figure at the top left is a simplified wiring diagram for an alternator, battery and alternator regulator.

Two resistors are shown, R1, and R2, which are in series with the leads to the battery from the distribution points. The resistors aren't real in the sense that they are distinct components wired into the system. The resistors represent the wires, all of which have some resistance. Large diameter wires have less resistance than do small diameter wires.

When current flows through a resistor, there is a voltage drop from one side to the other. The voltage drop is proportional the the amount of current flowing, according to Ohm's law, E = I * R. E represents the voltage drop, I represents the current flowing, and R represents the resistance. (The asterick symbol implies multiplication.)

As Ohm's law indicates, the higher the current, the greater the voltage drop. Likewise, the higher the resistance, R, the greater the voltage drop.

Current Flowing

When no current is flowing there is no voltage drop, even with a large resistance, R.

Now, for the sake of illustration, let R1 and R2 be 2 Ohms and let the alternator be producing 3 Amps. According to Ohm's law, each resistor has a voltage drop of 6 Volts. At this point the reader should be clear about the application of Ohm's law, even though the values used for illustration are not practical.

In the case of zero current, the regulator sees the battery voltage. In the case of 3 Amps, the regulator sees the battery voltage plus 12 Volts! In the case of a 12 Volt system, the regulator with 3 Amps current will see 24 Volts or more.

A More Realistic Scenario

In a more realistic situation, the resistance of the wires R1 and R2 might be on the order of 0.0016 Ohms, and maximum current might be 150 Amps. Now the voltage drop across each wire is 0.24 Volts.

Now the difference that the regulator sees is 0.48 Volts between no current and 150 Amps. What problems does that present?

Truncated Absorption Cycle

With 150 Amps the regulator is reading about 0.5 Volts higher than actual battery voltage. If the Absorption trip point is 14.4 Volts, the regulator will trip at 13.9 Volts. That doesn't make for a fast full charge.

At least one regulator sold today does a kludge around that by temporarily raising the trip point. Sounds good on paper. Of course one size never fits all, so how effective that kludge works is very system wiring dependent. The regulator doesn't have the information to deal intelligently with the problem on a case-by-case basis.

Set Point Instability

Since the voltage fed back to the regulator varies with changes in alternator current, the regulator ends up chasing itself. If it reduces field current to lower battery voltage, alternator current also declines. This means that voltage drops also decrease, so the regulator may now decide it's below the desired set point, and raise field current again.

If the cycling up and down happens rapidly, the user may never notice. However, alternators are slow to respond and response delays lower the cycling frequency so it becomes very apparent to the user.

Some regulators may overcome this instability by lowering the resolution used to measure battery voltage. Resolution is the amount of change in input voltage that the regulator can discern. For instance, a regulator using an 8-bit analog-to-digital converter, (ADC), can resolve about 0.06 Volts with 15 Volts full scale. Changes less than this would not be noticed by the regulator ...it would ignore such changes.

If the regulator uses a 10-bit ADC, then it can resolve about 0.015 Volts, or one-fourth that of the 8-bit ADC. While more susceptible to chasing itself, the 10-bit regulator will provide more accurate set points, including the point where it trips out of the Absorption cycle. That translates to a faster, full charge.

A regulator designer is faced with many choices, but the one germaine to this article is the choice of resolution. Higher resolution will yield better charging performance and longer battery life, but runs the risk of chasing itself in a cycle of up and down excursions.


The Ample Power Way

The engineers at Ample Power have always strived to provide fast, full charges while protecting the battery from under and over charge due to temperature changes.

What to do about the ground bounce issue?

A simplified drawing on this page shows the solution. The B+ and B1V are now two inputs rather than one. Likewise there is an ALT_GND wire added.

Note also that the wires now sense at the battery. Voltage drops across R1 and R2 no longer change what the regulator reads. If the installer is really paying attention, placement of the lugs on the battery will be done such that no current flows through the sense lugs. The back side of the battery post is a good place to attach the sense leads.

So, at the cost of two additional wires, a user can have the benefits of higher regulator resolution, while avoiding cyclic voltage excursions. But it only works if wired correctly.

Many people, who are semi-literate about electricity, see two wires on the wiring diagram that connect to the same conductive place and decide that one wire will suffice ...just jumper the terminals at the regulator.

Sometimes they're lucky ...R1 and R2 are small. Convincing them that they haven't wired the regulator properly isn't always easy.

"Ground is not ground" is a saying from the ancient days of analog circuit design. Now you know that all wires have resistance and connections along that wire are not at the same voltage when current is flowing.

And now you're also an insider who can speak ancient geek, "ground is not ground". Fashion guidelines? A pocket protector for you pens.