24v Amplifier

The selection of viable 24V amp is severely limited due to there not being many chips/chipsets to choose from.

On paper there's tons of chips that work up to 24V. However, on batteries what we're really after is amps that can tolerate at least 27V and run 4 ohms speakers which means they must be capable of delivering at least 4A per channel as well.

That actually only leaves TK2050 chipsets, and the TAA4100A chip. Both from Tripath. The first is well know although it must be stressed that other designers of TK2050 amps than Sure use the TC2000 controller with far more capable output drivers than Tripath's own TP2050. Ones that tolerate far higher voltages (up to 40V) and higher currents (up to 6A).

In any TK2050 based amp it must be stressed that the current output of 6A can only be gained if the output driver chips are used in mono mode. As the chip package simply cannot handle more current than 6A.

The TAA4100A is the chip used in the amp9. I regard it as the most suitable chip of all for this purpose. Tolerates 27V under load, and up to 10A. The 10A rating makes this 4 channel amp a shoe-in companion for use with a miniDSP. I my opinion it also surpasses TK2050 amps in sound quality.

The TDA7498 chip which is also used in some 24V Sure amps, are not suited for 4 ohm. Remember the I said the chip package limitation of 6A above for TK2050 output driver chip package. This is the same package so it's completely out of the question to use it with 4 ohm spakers.

There are a couple brand new chips that look interesting that have just come out from NXP (formerly Philips) but they have not been used in any design yet.

Don't buy this. There's an updated version out that makes up for the shortcoming this amp has. It's discussed over in the 48V boost converter thread.

amlu wrote: spring is coming may give it a go.... need to borrow a sound meter from someone and build some test rig

It's the TPA3116D2 that is the most promising. However, it's has a 27V +-1V overvoltage protection which basically means 26V maximum operating voltage. 2 12V batteries in series will after reverse polarity protection diode forward bias of -0.7V have 26.9V, if the batteries are fully charged to 13.8V. LiFePO4 batteries which are 14.6V fully charged (for a 4S standard package) are naturally not possible either. 

Like several other similar chips they almost meet the requirement but only almost.

EDIT: Actually I might have found a way it can be used after all. Changing the reverse polarity diode to a linear voltage regulator (lowers voltage over a certain point by dissipating the difference as heat) then with a trip point set to 26.6V (prior to diode drop), the regulator will in worst case scenario only "steal" 0.2% of the total battery capacity even with LiFePO4 batteries.