The circuit is based on Figure 14 in the MC3334x datasheet, with some improvements and modifications. Although you could probably figure out the circuit's operation by staring at the datasheets long enough, I'll walk you through the circuit here so you don't have to.
You should have a copy of the schematic handy to follow the discussion.
The regulator (U1) is set up as a constant-current source. (As opposed to a constant-voltage source, which is how linear regulators are used most often.) An LM317 style regulator strives to always keep the voltage drop from OUT to ADJ equal to 1.25V. If you put a resistor between the two (R2), the regulator puts out whatever current is required to keep the voltage 1.25V while satisfying Ohm's Law:
I = 1.25 / R2
R3 is technically involved here, because it interacts with R2 to form a voltage divider. As long as R3 is much higher than the value of R2, this effect is very small. Calculate it if you like, you'll see. The real purpose of R3 is to limit the current from the ADJ pin to just enough to keep the regulator operating. See your regulator's datasheet for details, but a typical range is 5 to 10mA.
During fast-charging, the charge controller chip (U3) monitors the battery voltage through its Vsen pin. The battery voltage runs through a voltage divider (R4/R5) to get to this pin. The divider drops the voltage down to within the sensitivity range of the Vsen pin.
The charge controller's F/T and Vsen Gate pins are connected internally to the collectors of NPN transistors. Normally these transistors are open circuits, but the chip can push a bit of current into the base of the transistors to open up a path to ground. The ADJ pin of the regulator is indirectly connected to these pins, so that when these pins open up, it shunts away the current the regulator needs to operate, so the regulator shuts down. Voilá, the circuit is in trickle charge mode.
The F/T pin opens up any time the charge controller wants to put the chip into trickle mode long-term. Some reasons it might do this are because the Vsen pin senses that the batteries are charged, or because the voltage at the Vsen pin is out of range, or because the fast charger has been running too long, or because the battery pack's temperature is too high. See the MC3334x's datasheet for full details on the conditions under which the chip will change modes.
The Vsen Gate opens up every 1.38 seconds during fast-charging to turn the fast-charger off long enough for the chip to test the battery pack's voltage. It does this because it's less accurate to test the battery's voltage during fast-charging. The chip watches the voltage of the battery pack through this pin, and turns off the fast charger when the voltage at this pin drops during two successive tests. This is called -ΔV charge termination. If you look into the datasheets for NiMH cells, you see that at the end of fast charging, the voltage levels off and then starts falling, which is exactly what the MC3334x looks for. This charge termination method ensures that the cells are as fully charged as they can get. Time, temperature and voltage threshold termination methods look for indirect indications of fully charged cells, so it's hard to make them work optimally in all situations. They're okay methods for backup charge termination, but I don't think they're very good for primary charge termination.
Since the charge indicator LED is along this F/T and Vsen Gate path, it is off when the fast charger is running, and it turns on when the fast charger is shut off. Since the fast charger is shut off for a short time periodically while the chip is in fast-charge mode, the LED blinks.
The TLE2426CLPs (U2 and U4) are there to cut high supply voltages down below the 20V limit of the MC3334x chip. Because you need higher voltages to charge more cells, you end up crossing the 20V line at around 10 cells, the exact number depending on the charger's circuit configuration. The MC3334x is just a controller, not the charger per se, so it doesn't need to see the full supply voltage; scaling the voltages down to safe levels works just fine. I did try resistor dividers where these TLE2426es are now, but that didn't work for various reasons. You can read the gory details in the PPA Battery Board thread on Head-Fi. Fair warning: My understanding of this circuit has improved since my first post in that thread, so several of my posts in that thread contain incorrect information.
The trickle charging path is through R1. As you can see, it's always in the circuit. While the fast-charger is running, the trickle charge current is still there, but it's swamped by the much higher fast charging current. The trickle charger serves several purposes: to top off the batteries after a fast charge, to maintain the charge level while the amp is plugged in, and to bring severely depressed batteries up to a level where they can be safely fast-charged.
The second schematic page shows a way you can modify the fast-charging circuit to run in a constant trickle mode. It still uses U1 as a constant current source, but it sets it up for a much lower current level. Many components can be removed or jumpered in this simpler arrangement. Even if you use an LM317 with a fairly high drop across it, you probably won't need to add an extra heat sink in this circuit.
I could have used the R1 path as the trickle charging path, but while the batteries are charging their voltage drop increases, so the current would change over time. This isn't as big an issue in the fast-charging circuit because the voltage is more stable in the situations where the fast-charging circuit goes into trickle mode.
Notice the switch across the F1 position on the schematic. Modern NiMH cells will tolerate trickle charging for quite a while, but it's better if you stop the charger once the cells are charged. This switch lets you turn the charger off without unplugging the amp. It's not required, but I think it's a good idea to add it.
Because there is an output diode in this circuit, you can put multiple boards in parallel. This helps the boards share the load without interfering with each other, even when the boards are in various states of charging.
You might think that you could use one charging circuit for both boards, but this doesn't work well. For one thing, NiMH cells just plain don't like being charged in parallel. Even if you were able to work around that, you'd have to configure the regulator to put out 2× the current to charge two battery boards in parallel to get the same charge time. The heat from fast-charging just one set of batteries is bad enough, managing multiple times that dissipation might well be untenable. If each board has its own charge circuit, the heat generated is the same, but it's spread out.
You could also lash multiple boards up in series. This is probably only a good idea in AA configurations, to get higher voltages (20 cells) than is possible in AAA configurations. You have to figure out how to make one charge circuit charge both sets of batteries, or you have to rig some kind of switching scheme that separates the boards and lets them charge separately. You'd also have to disconnect the batteries from the amp while they charge if you took this path. No matter how you lash it up, I suspect it would be more work than it's worth.
<< Step-by-Step Assembly Guide
|Helpful Links >>
|Updated Fri Dec 16 2022 12:23 MST
|Go back to The PPA Project
|Go to my home page