Need an easy low power class D amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm in need of a small low power (5w to 20w range) class D amplifier circuit. I'm thinking class-D is purely to maximize battery life, but I'm not committed and open to suggestions. The application is to improve what has become a very useful, portable and small (4" x 4" x 2.5") guitar amplifier I made some time ago, to improve battery life. I'm already using rechargeable lithium cells, so further improvements must come from circuit efficiency and speaker choices. Anyway. here's my wish list. Any recommendations, pointers, or even offers for your own designs that are up for sale are much appreciated.


  1. Capable of operating from a low voltage single ended battery supply, as low as 5 or 6 volts.To maximize output at such low voltages, bridge circuits seem appealing, though I don't know if that's problematic for class D.
  2. Some adaptability in input gain. Typically the amplifier will be fed by a 50-100mV signal from a pickup.
  3. Adaptable to available speaker impedance, at least down to 4 ohms, without loss of stability.
  4. Simplicity. I'm not looking to cut my own boards. If possible, I'm looking for a ready made building block, or something close to a kit.
  5. Again, simplicity... I'm sure single ICs are available to do the lions share of the work, which at the very least will minimize my parts count.
  6. Availability... With the "just in time engineering" so common today, I can't work with a circuit that requires minimum purchase requirements on key components.
Thanks in advance!
 
I'm in need of a small low power (5w to 20w range) class D amplifier circuit [...] Thanks in advance!
Hi PeterPan,
The low operating voltage of 5-6V limits the choices. National makes a chip with a built-in supply booster from 5V, but it gives 5.4W into 4Ohm (3W into 8Ohm) only.
Other chips I've seen need at least 8V (TFA9815T by NXP or Yamaha YDA148) and in BTL connection give someting like 8-10W into 4Ohm. Incidentally, I found this finished board with volume pot using the YDA148 (for $9.50!!). You might be able to wire it for BTL, the DS has the schematic.

Another potential problem with class D for guitar use is the clipping, which can sound ugly unless precautions are taken at the design stage.

Lastly, I am currently working on a small mono block for an unrelated use with volume and tone control pots, using the TI TPA3123D2 chip for 20W/8Ohm, but it needs at least a 10V supply (=less power). PCBs are still weeks away though.

How about using Li-Ions in series, eg the RC LiPo packs? With 4 cells (~14V) you'd be in business with just about any solution.
 
Thanks Elevator! We should stay in touch since I see we tend to make amplifiers for similar usage. I've poked around on the MAXIM, Analog Devices, and I'll look at TI too. As I look into the innards of a lot of these chips (courtesy of Mfgrs internal block diagrams), I see that they usually already include H-bridge's internally, so I can't bridge two whole amps to get more than the single circuit rating as I used to do with traditional AB chips. That's a bonus as far as I can see. It seems though that its tough to find designs that will accommodate a wide supply range as I'd hoped for. I suppose its just more difficult to do so with class D circuits. Its a problem only because of the small size I'm trying to achieve. My last AB design used 6 AA size NIMH batteries (all i could fit in the box) so fully charged I had 7.5 volts, and the amp would work reasonably well till the battery voltage started dipping below 5V. So in a bridge circuit, you can see where clean peaks up to about 20w was possible. Of course playing that loud would drain the batteries pretty quick, but that would be the case with any circuit.I guess I'll have to settle for a little less power to start, but as a practical matter for this size category and battery capacity, anything over 5W will probably be BS (that's NOT an acronym for some kind of bridge connection). :)

Thanks for the tip about clipping. I'm not sure why its more a problem with class D, but this probably explains why a lot of small class-D based guitar amps have such good limiter circuits... they NEED it! :)
 
You're welcome.
What's your speaker impedance? At low supply voltages, a low impedance helps. Also, many class-D chips are two-channel with each channel an H-bridge, and some allow paralleling the two channels. Case in point the Yamaha YDA148. I had a closer look at it and I like what I read. It can also avoid clipping by compressing when needed, and the behavior is configurable. I'm actually thinking about designing that one in, in place of the TI. Only caveat, I've had availability problems with Yamaha chips in low-to-medium quantities in the past.

Other thing, have you looked at LiPo rechargeable packs? They're half the size of NiMH for the same power, for example, this at $15 gives 12V (not 11.1V at stated) for a long time. Add the class D's ~90% efficiency (compared to 50-60% for AB) and you could be jamming until 7AM (ok maybe 3AM:D)
Cheers

PS On edit: I agree that you may not actually need a full 20W all the time. But it's nice to know that it's there, for the fraction of a second when you do want it, instead of sagging, limiting or badly clipping. And, with class-D, it comes almost free (power consumption wise).
 
Last edited:
Paralleling outputs? Very interesting! Since most of these chips already have H-bridge outputs, I'd never have guessed there would be any way to combine a pair to increase output, beyond the obvious use of 2 speakers ;-) Obviously to do this, there must be some way to force multiple amps within the chip to use the same internal oscillator, but then maybe they already do (DUH! ). Still, unless the outputs were combined AFTER the usually recommended LP filter, wouldn't the switching characteristics of the two amps need to be 100.0% identical to pull this off? Well maybe that's a "non problem" too these days. Seems that with Class -D, I did my usual ignoring of the technology until the 20 pages of inherent problems were solved. In the mean time, a couple of decades have passed and its likely every problem I can think of has already been dealt with!

Well if paralleling of outputs is possible, this opens up a lot of possibilities for me! With automotive chips often boasting 4 channels, combining them would be a convenient way of adding power and driving lower Z loads. maybe even 2 ohm speakers!

Thanks for the tip on LiPO batteries. I never felt comfortable with the LI_Ion counterparts, fearing my little amplifiers (which I often make small enough to wear on a belt or bag) would catch fire when I tried to show off its power (LOL). But then I also though the LiPO batteries had a problem with sudden peak demands, something even a small guitar amp would have to deal with. Maybe this too is a "solved" problem?
 
Paralleling of H-bridges must be supported by the architecture, of course. For one, self-clocked designs (aka ucd) are unsuited. In clocked designs, usually the chip outputs are paralleled before the LPF. Makes sense, as basically your only worry then is on-chip propagation delay skew, as opposed to LPF component tolerances potentially making a mess.

Another thing, if the speaker is close to the amp as in your case, you can do without the LPF (using the speaker as the LPF; some might get slightly warm though, depending...).

LiPo packs for RC use support huge discharge currents: reading the specs, "20C" means that it can be drained with 20 times its capacity, eg a 1000mAh pack can be discharged at 20A, continuously (for ~3 minutes, ie until it is empty...). In short, I don't see a problem here.
 
@Elevator: That Y148 chip looks pretty good, and I see from the data sheet that a parallel output connection is pretty easy. I'm tempted to buy the board to try it. I wonder if its just a 2 layer board. Pretty tough to modify I'd think if not.Did you say there was a schematic somewhere? I do see what you mean about availability though. I can't find the chip itself anywhere. Buying that board might be the only way. Which I wouldn't mind if I knew that have a decent supply in stock. I don't have any illusions about making and selling a lot of these, but invariably a few friends will ask about things like this that i make, and I like to know that a future availability of maybe 20 isn't impossible. Guess I'll give them a call (their chat line seems un-manned).
 
Peter Pan,
The changes required to the board depend on whether it uses a differential or single-ended input. I'm guessing SE (I don't have the board myself): use INLP as input, tie INRM and INRP (at the chip, ie after the DC blocking cap) to AVDD, then check if both channel outputs have the same signal. If yes, connect the speaker to both outputs, minding the polarity. The first time I'd use series resistors:cool:

Back in 2004 or so I was trying to design-in one of their chips, talked to their US rep at the time, and he had a guy from Yamaha fly in from Japan, with NDAs, PPTs, meetings with coffee and all. By the time they were ready to quote, I had already implemented their chip's function in firmware so I called the circus off :clown:. if you want I can give you the contacts (by PM).
 
Hey everyone, especially 'Elevator' that mentioned the LiPO packs. I did order some of these, as the price was right and the energy storage seems amazing! But I've never dealt with LiIon or LiPo before, and didn't realize that charging them required such a complex circuit, that is able to balance the charging of each cell individually. Now I can understand such a scheme if you wanted to really charge in a hurry, like 1C rate or higher. But suppose all you want to be able to accomplish is overnight charging, so you were only going to charge at maybe 1/2C or even less. Is all that extra circuitry to balance the individual cell charge really necessary? Will it ruin the batteries (or worse, make them explode!) If I just give them a constant current through all cells of maybe 500mA for a 1200mAh 3 cell pack?
 
Back to my original request for a simple modular low power Class D amp. Can anyone recommend any other Class D amplifier modules in the 10-20 watt category? Now that I've found LIPO packs, I'm going to be using 12 volts and using a 4 ohm speaker, that should make it easier than my original 6 V requirement.

I did order some of those Y148 boards ELEVATOR recommended and they do work well, but they are pretty near impossible to modify. I need to increase the gain some and I'm sure those options are available, but I can't find ANY data on the chip itself or a board schematic. So I guess I'm looking for something small that ALSO provides some documented configuration options.

I know I'm being lazy not building my own board and such.
 
Can anyone confirm that YDA148 is switching at 1MHz from the datasheet?

The board from dealextreme comes with ceramic input capacitor. You can replace it with film caps of 1uF but make sure that all 4 caps of the differential inputs are replaced, even if you're using single ended input.
 
By the way, a little off topic now, but I was indeed able to modify the gain on these boards to the higher level, by removing one 0-ohm resistor and replacing it at a different position. In my originally intended application, to build an amplifier w/speaker and battery pack into a small travel guitar, this amplifier has exceeded all my wildest expectations, and sounds awesome. I think because the lowest note on a guitar is about 82Hz, there is no need for me to consider changing the capacitors. Thank God really! Its good to know it can be done to improve bass response, but my eyes and nerves are just not cut out for modifying components this small. ;-)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.