Hopefully not too much to ask

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
What I want to do is build a subwoofer as a present for my dad for christmas. He is planning to get a surround setup, and will need a sub. I already have the speaker and box, but now need the amp.

I am going to be using this amp circuit:
http://www.mif.pg.gda.pl/homepages/tom/files/simp200W.gif

With lm3875 chips and MJ15003/4 transistors.

Like I said, hopefully this isn't too much to be asking for for free, but would anybody be able to help me design a PCB for this project?

I am not in a big hurry since christmas is still over 3 mo away, so if somebody is willing to spend some free time doing this I would really appreciate it.

Thank you so much!
 
soundNERD said:
No, 120W is no where near enough. I have tried bridge 3886, not at all impressive.

Must be the driver/enclosure combo you're using then.

I have 200W driving a pair of 15's in an IB configuration (in a not too huge room) and I can watch the wall oppposite the driver manifold flex sometimes. Playing back T-rex footsteps will cause the "ripples in the water glass" thing that tips the people in the movie off to there being something scary coming, even when the stuff is too low to hear (though, I do get a tingle on my spine...). and the glass is sitting on a table that's on carpeted concrete basement floor. And that 200W is more than I need (the drivers run out of excursion before the amp clips, and my ears are already bleeding at this point)

But, given that I spent $100 on the 200W amp (which also includes phase control, crossover, line level or speaker level inputs, speaker passthrough, etc.) I still question the value in trying to use one of these chip-amps in this situation.

I should add, I'm still working on PCB layouts of boards (buffers, NI and I 3886 for regulated power, regulated power boards) that will lend themselves to parallel configuration well. Doing a BPA wouldn't take much effort either. Though, the LM338 reg shouldn't go beyond supplying two chips... Progress is ongoing, but it's at a very leisurely pace. :) I'm at the point where I think I need to switch programs so I can actually send the design out for proto boards. It's not a custom layout though.

There are layouts out there for various paralleled, briged, etc. setups as well already.

C
 
200watts?

i think your being optomistic with 200watts. i could be wrong, ive built a large array with 6 x tda2030 driver and bd907 bd908 using roughly the circuit in the data sheet and tda2030 chip runs best at about 36v. what about adapting the low pass circuit in figure 20 and running 4 amps with 4x6 inch drivers. or use two of your circuit with paired 8ohm drivers. c ya soon, steve.. ..
 
what layout are you talking about?

Would I even notice much of a difference with 400W using this circuit than a bridged chip? I would use the LM3875 so I don't lose sound quality with the mute.

The speaker is actually a car speaker I couldn't resist on clearance for $20. It was a $100 12" Rockford Fosgate Punch Z that somebody had bought then returned because it didn't fit. It was still brand new and works perfectly. I know it is no where near the best, but well, its all I can afford, and still sounds decent.

Anyway, I believe the SPL is somewhere around 86-88. Would a bridged 3875 be enough? Wish it was a DVC so I could bridge 2 amps for each VC. Then I'd have at least around 300W.Now that I think when I tried the LM3886 bridged, I didn't do too well (just soldered some wires on 2 seperate boards to make one invert). Then I was trying it with a cheapy speaker (Some no-name 10" speaker). Then I realized the trace that kept the mute off had broken, so only one amp was running. I would think 2 3875 bridged would sound pretty good would it?

Or would the other circuit put out much more volume before distortion? Is it really worth the extra time trying to design a PCB for it then figuring out how to mount 4 TO-3 chips plus 2 to220 chips to a heatsink and having to mess with insulators for the to-3 chips?

When I design this one, I want to have phase control built into it. Anybody know of a good circuit for this?

Finally, would I be better off running both chips in either iverted or non-inverted (which one is better for this?) and using a DRV134 to split the input, invert one side then balance them?

Thank you all so much.
 
Um...

you could only afford a $20 driver, but you want to spend how much putting together a GC to power it?! :rolleyes:

Take a look at the BrianGT 4870 board group buy (not quite available yet, but very soon it seems). It sounds to me like this is the perfect solution for you here.

Also, the subtle SQ things you might find with a GC are largely going to be moot driving a subwoofer. You mostly need power. You may want to increase capacitance significantly as well.
 
soundNERD said:
what layout are you talking about?

Would I even notice much of a difference with 400W using this circuit than a bridged chip? I would use the LM3875 so I don't lose sound quality with the mute.

The speaker is actually a car speaker I couldn't resist on clearance for $20. It was a $100 12" Rockford Fosgate Punch Z that somebody had bought then returned because it didn't fit. It was still brand new and works perfectly. I know it is no where near the best, but well, its all I can afford, and still sounds decent.

Anyway, I believe the SPL is somewhere around 86-88. Would a bridged 3875 be enough? Wish it was a DVC so I could bridge 2 amps for each VC. Then I'd have at least around 300W.Now that I think when I tried the LM3886 bridged, I didn't do too well (just soldered some wires on 2 seperate boards to make one invert). Then I was trying it with a cheapy speaker (Some no-name 10" speaker). Then I realized the trace that kept the mute off had broken, so only one amp was running. I would think 2 3875 bridged would sound pretty good would it?

Or would the other circuit put out much more volume before distortion? Is it really worth the extra time trying to design a PCB for it then figuring out how to mount 4 TO-3 chips plus 2 to220 chips to a heatsink and having to mess with insulators for the to-3 chips?

When I design this one, I want to have phase control built into it. Anybody know of a good circuit for this?

Finally, would I be better off running both chips in either iverted or non-inverted (which one is better for this?) and using a DRV134 to split the input, invert one side then balance them?


I seriously doubt, to the point of saying is non-existent, that the mute on the LM3886 and LM4780 will affect the sound quality. That is starting to become an urban legend that should be vanished from our DIY scene. If the project is well implemented there won't be any difference between "mutable" and "non-mutable" chips, or if there is they might favour the 3886 and 4780.

If you want to make a high power project, forget about the 3875 and go the 3886 or 4780 way. End of the story.

To how many SPL a bridged amp using these chips will take your speaker will depend on its impedance and other specs.

I would think of a 2 + 2 chip amp, parallel/bridged and using a summing circuit for both channels. Or use separate subs that may be better. That is another urban legend: that you don't need separate subwoofers. Two subs may sound better than one according to what you listen to.

But building a project with these chips will be simpler than a regular one using TO3 and TO220 parts.



Carlos
 
cjd said:
Um...

you could only afford a $20 driver, but you want to spend how much putting together a GC to power it?! :rolleyes:

Take a look at the BrianGT 4870 board group buy (not quite available yet, but very soon it seems). It sounds to me like this is the perfect solution for you here.

Also, the subtle SQ things you might find with a GC are largely going to be moot driving a subwoofer. You mostly need power. You may want to increase capacitance significantly as well.

yes, I do because once I get some money down the road, I will invest in a better driver. Good present for next christmas, by then I'll be old enough to get a job to afford it.

I never said I want to build a great amp, just want the most bang for the buck.

Another thing im thinking about is using mosfets. Would that be an easy and affordable way of building a high power amp?

And, another reason I do not want to mess with mute is the increased complexity. Less parts = less traces = larger traces in the same spot for what really needs them = better sound. Am I not correct here?

About the room, this is a huge room, about 40-45 ft by 25ft, with no walls between rooms, so bascially it would have to make enough sound to cover the entire room while losing some to the other rooms. Would a bridged 3875 be good for this?

Finally, cjd, are you saying a phase control would be expensive? Do you think it would make a large difference? All I know is that most expensive units have them, so if they are as simple as using a potentiometer to control how much of the sound goes to the inverted and non inverted inputs of an opamp, I might as well throw it in.

Thank you all again for your input on this project!
 
On the one hand, I am saying that a chip-amp is not going to be competitive price-wise with what you can buy, all said and done. I spent ~$120 on my basic BrianGT 3875 kit. I spent $100 ($120 normally) on a sub amp, which gets 250W into 4ohm, adjustable phase, built-in filter, and all that.

FWIW,
A phase reversal is ultra simple. I'm not sure on adjustable phase.

If you want to do a chip-amp, when Brian gets the 4870 kits finished that'll be the way to go for a sub-amp. Really. Ultra simple. I just don't think you'll be able to beat that deal with a stick, since you're requesting a PCB. Heck, just the parts when you do the Group Buy end up being a deal.

I wouldn't go with the 3875 in this application. 3886 if you want single-chip fun, but at that point I have to again suggest the 4870, . . .

And, that's a BIG room! hyooooj. I bet you're going to discover the driver you have is, well... disappointing.

If you want to DIY, there may be better options for a sub amp. But I'm not the one to ask about that. I do think there are some other threads here on the subject though.

C
 
"Thats a BIG room" Yep, thats why I kept insisting on high power. Now, I haven't physically measured the dimmensions, but it looks to be about that size.

Yes, i'm sure the driver isn't going to be the best, but like I said, I'm gonna work with what I have now, then upgrade later.

I have the PCB software, I have some boards, I have some etchant and I have a couple of hours, so I might as well throw a couple of subamps together using a bridged 3886 and another with a bridged 3875. They are going to both have a DRV134 for input inverting, a crossover, and a phase control.

Would anybody know if this would work as a phase control?
 

Attachments

  • phase.jpg
    phase.jpg
    4 KB · Views: 180
Try paralleling first? Bridge after that if you need more power. Among other things, it'll handle lower impedance drivers.

And, with subs, you generally run out of excursion before you run out of power. It's all about how much air you can move at those frequencies. More power can sorta make up for it, but only sorta.

You're always welcome to do what you want, just trying to point you down a road that's more likely to end in success and no pulled out hair. If you enjoy the process, well... start designing PCBs and post 'em for review if you're unsure!

C
 
the LM3875 aren't REALLY suited for less than 8 ohms...

also, if you bridge them you get half the impedence on each amp, so you'd be looking at 2 ohms per chip, and the protection circuit being perminently on....

if you want to drive a 4 ohm speaker from a bridged LM3886, you can, but you need to parallel/bridge them, so have 2 pairs paralleled, and then bridged..... thats the only way...
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.