Big Bass Amp
Fri Jul 6 23:58:52 PDT 2007
Don't expect this to be organized. I'm just going to string stuff in here
in a linear fashion along with some notes. It's going to be a mess.
Here's the schematic so far:
I decided that using inverted amps for one half, and non-inverted for the other is kind of silly. Everyone seems to use a DRV134 to give them the inverted signal. I'm just going to use an opamp. It's there to buffer the input anyways, so I may as well use the other half to do the inversion.
I've decided not to bother with servos for now.
I'm slightly concerned that my choice of output resistors may be too small. Since I bought most of the parts, I've seen a few suggestions that the .1 ohm called for in the appnote is a little on the low side and that .2 ohm is probably better. I bought .07 ohm, so I guess I'll have to match things really well. Shouldn't be too difficult. I think I should have enough resistors to match them quite well. I'll probably figure out a setup to match the LM3886 chips as well. I'll have about 20 to choose from so I should be able to find two good sets of 3 with similar dc offsets.
The speakers I bought look like they'll be well suited to the purpose. I decided to go with 8" drivers since I wanted to use hifi speakers rather than guitar speakers. I expect that a larger woofer would probably not have enough midrange, since hifi speakers don't usually have the same kind of cone breakup as a guitar speaker. My theory is that 8 inch speakers should reach high enough in frequency that I won't need a seperate mid or tweeter.
The cones seem quite light as well, which should help.
On another subject I decided a couple days ago that I was going to use the 3886 instead of the 3875 which I was originally planinng (Already have some 3875). I found some 3886 for $2 each, and they'll give me a lot more flexability if I decide that I need a second cabinet (maybe a 12" or 15" woofer to fill out the bottom?) For the extra few bucks, I figure I'll keep the option open.
For the power supply I'm currently leaning towards a switching supply. It looks like an adequately sized transformer would be expensive. Looks like from plitron I'd be looking at about $135 for a 625VA up to about $172.01 for a 1kVA. 350W computer power supplies are available for about $25. One per rail should do it. It seems that the newer ones are all a little bit tougher to modify than the older ones were though. Ideally I could find a simple supply, rewind the transformer for a higher voltage and modify the regulation to suit. I've done the regulation changes on a few older supplies, but all of the newer ones I've pulled apart seem a lot more complicated on the regulation/protection side of things.
I may just try and build a very simple switcher using components from the computer supplies. I tore apart a Fortron 350W, and based on what I could find specs on it looks like the switching controller is internally clocked. Probably ideal for hacking something together with an unknown transformer, since I at least know that the switching frequency is correct. (Update: Actually it looks like that controller just switches the standby power, and it's a custom controller that does the rest)
Sat Jul 7 21:23:45 PDT 2007
I just added the extra parts for the servos to the schematics. I don't know if I'll use them right away, but this way if I decide that I do want them, I can just replace the DC blocking caps with wire, swap Rb for a 1k and then add all the extra parts. The schematic looks gross though, and it took me 20 minutes to screen cap and stich it together. Hopefully I don't need to make any more changes. I'll do the power supplies on a seperate schematic, and the preamp too.
I changed my mind. The amount of board space I want to fit the amps into doesn't have room for all the parts. Removing those huge caps should help. I'm thinking the servo is probably a better choice anyways, given that my output resistors are such a low resistance.
I'll put the (hopefully) final schematic up tomorrow.
Sun Jul 8 11:12:22 PDT 2007
The schematic is up. With the exception of some hf filtering it's pretty much in line with the appnote version, except it's all non-inverting, and inverted by an opamp at the input.
Sun Jul 22 21:10:55 PDT 2007
I've been slacking off on the updates lately. The PCB has been through several complete iterations. After several attempts at a single board design, I decided to go with a more modular design. The routing of the power was always ending up quite ugly, and would either need a large area of PCB or would require adding wire to increase thickness. The other advantage to this design is that more or less boards can be built into the amplifier depending on requirements.
I Also decided that the star grounding at this level shouldn't be necessary. I made two seperate ground planes, one for high current (PGnd), one for low current (SGnd) which will all be wired back to a star grounds at the main power supply connection and the two stars connected at that point.
I also went back to putting all of the power stuff on the left side of the chip (Same as in my previous amp). Most people don't do it that way, but it certainly seems easier, and probably leads to a much better overall layout, as all the high current stuff is more easily seperated from the low current stuff.
