BR-100 LM3886 pcb
Hi all,
This is the first time I've attempted to design a pcb; any helpful critiques would be much appreciated.
https://easyeda.com/Rabenvald/Gainclone_AMP-dd3acd02dd2c4f52b5c5c5702145f766
https://easyeda.com/Rabenvald/Gainclone_PSU-180fbe86786c4e61b3ddca0cd6aa66ab
Hi all,
This is the first time I've attempted to design a pcb; any helpful critiques would be much appreciated.
https://easyeda.com/Rabenvald/Gainclone_AMP-dd3acd02dd2c4f52b5c5c5702145f766
https://easyeda.com/Rabenvald/Gainclone_PSU-180fbe86786c4e61b3ddca0cd6aa66ab
I presume those BIG capacitors C11-C13 are polypropylenes used for speaker crossovers. DONT do this. Any advantage of them being polyprop's is cancelled out by the huge antenna they form.
Wima MKS2's are film caps if you must use film, polyester but they have 5mm pin pitch.
Power supply board - dont waste one side on being a largely useless ground plane. Either make it a single sided board, or put the same tracks on both sides to double current capacity. I wouldnt bother with the bleed resistors R1 and R3 either. I also wouldnt bother with so much capacitance, 20,000uF per rail is crazy for a chip that will only give you 50W at most.
Wima MKS2's are film caps if you must use film, polyester but they have 5mm pin pitch.
Power supply board - dont waste one side on being a largely useless ground plane. Either make it a single sided board, or put the same tracks on both sides to double current capacity. I wouldnt bother with the bleed resistors R1 and R3 either. I also wouldnt bother with so much capacitance, 20,000uF per rail is crazy for a chip that will only give you 50W at most.
Also put some smaller power supply decoupling caps eg 470uF on the amplifier PCB, near the chip. Dont rely on the capacitance on the PSU board alone, as you have wires connecting one to the other. The PSU board's capacitance should be for energy delivery, and the amp board's capacitance should be for supply decoupling.
Don't forget that the output impedance of the preceding stage will change the gain of the inverting LM3886 and attenuate slightly the input to the non-inverting LM3886. Also R5 and C11 are now parts of the input impedance seen by the inverting LM3886. You probably need an op-amp buffer or better two (one for each LM3886) on board to make sure that no matter what is connected to the input it does not alter the behavior of your amp.
The amplifier board, especially before I reverted it to a previous state, is a bit of a mess; I know. I've only touched it in my free moments after work.
Anyway, I really should have made a separate copy for revisions earlier, but here we go:
https://easyeda.com/Rabenvald/Gainclone_AMP_v1_1_copy-9a95558d9b6149e39af9cecec184a42f
@jaycee: Regarding those large caps acting as antennae: I suppose that's probably not something that can be overcome with shielding? Bleh, that's a shame...
@OlegSh: I didn't need the mute functionality, so I omitted that part of the datasheet. Does that pin need to be connected somewhere?
Anyway, I really should have made a separate copy for revisions earlier, but here we go:
https://easyeda.com/Rabenvald/Gainclone_AMP_v1_1_copy-9a95558d9b6149e39af9cecec184a42f
@jaycee: Regarding those large caps acting as antennae: I suppose that's probably not something that can be overcome with shielding? Bleh, that's a shame...
@OlegSh: I didn't need the mute functionality, so I omitted that part of the datasheet. Does that pin need to be connected somewhere?
@OlegSh: Not only did I read it, I misread it too! Oh joy!
@Mark Whitney: I was building this entirely off of the supplementary document from TI: http://www.ti.com/lit/an/snaa021b/snaa021b.pdf which may or may not have been a good idea at this point... I chose the bridged design over the parallel one, purely for the decreased distortion claimed in that document. Honestly though, I am not an EE, nor do I desire to be one. This is exactly the sort of thing I would have loved to pick up in kit form.
@Mark Whitney: I was building this entirely off of the supplementary document from TI: http://www.ti.com/lit/an/snaa021b/snaa021b.pdf which may or may not have been a good idea at this point... I chose the bridged design over the parallel one, purely for the decreased distortion claimed in that document. Honestly though, I am not an EE, nor do I desire to be one. This is exactly the sort of thing I would have loved to pick up in kit form.
if you.. take the best of adices i can think of... that is.
stay with something simpler, a monoblock lm3886 sounds like a right choice.
do ask a lot on pcb design, it has far more to do with distortsion in case of these chipamps then bridge/paralell stuff.
learn good grounding principles, and when you are confident i might think an opamp buffered lm3886 will suit your needs.
starting out with a project that is most proably more complex than what you can overcome, is a pretty close call for a project ending without joy.
actually the tipical starter amplifier i recommend..
http://www.st.com/content/st_com/en...s/class-ab-audio-power-amplifiers/sta540.html
thisone, in the bridged configuration. as it requires far less components, has a fixed internal gain, and its a pretty pretty hefty chip.
give it a go.
stay with something simpler, a monoblock lm3886 sounds like a right choice.
do ask a lot on pcb design, it has far more to do with distortsion in case of these chipamps then bridge/paralell stuff.
learn good grounding principles, and when you are confident i might think an opamp buffered lm3886 will suit your needs.
starting out with a project that is most proably more complex than what you can overcome, is a pretty close call for a project ending without joy.
actually the tipical starter amplifier i recommend..
http://www.st.com/content/st_com/en...s/class-ab-audio-power-amplifiers/sta540.html
thisone, in the bridged configuration. as it requires far less components, has a fixed internal gain, and its a pretty pretty hefty chip.
give it a go.
It's a mobile amplifier designed to run off 12 volts. It's a car radio amplifier. It's about as good as it gets without using a switching power supply or maybe a discrete circuit, which would be silly for this application.
I think it is a good chip to get your feet wet. You can hardly go wrong. A 3886 seems simple to us because if it has a problem we know right away what's wrong. A newbie might be clueless if it doesn't work. Add that to the fact that many internet circuits are incomplete and will only work under specific conditions, and you might be in store for disappointment.
I think you could build a 3886 diff amp by getting some decent mono or stereo boards, assembling them with high stability 1% (or even 0.1%) metal film resistors (like Vishay Dale CMF series), and a dedicated differential line driver like DRV134 https://www.digikey.com/product-detail/en/texas-instruments/DRV134PA/DRV134PA-ND/275875 . I know I'm going to try it one day, and I'm going to shoot for around 100 watts RMS into 8 ohms.
Last edited:
It's exactly like the chips you find inside modern car radios "50 watts RMS x4".
I've taken a few apart. There's various chips in them, but it's always a 4 channel bridge amp in a "multiwatt" package. It seems to me that it was designed for this application.A bonus for the DIYer. It could also be used in a "boombox" or "mini hi-fi." But the modern trend in those is to use TDAxxx chips. I've taken a few of those apart too.
But I haven't seen this chip in any commercial products. There must be a reason for that.
surely, yess it has 4 amps inside, just like car radio chips.
but car radio chips are not ment for 22 volt supply. nor does the datasheet says it was ment for any automotive application.
12V systems would be too low, 24V systems would be too high.
but works perfectly from 3x6V battery config, if you want a boombox.
but car radio chips are not ment for 22 volt supply. nor does the datasheet says it was ment for any automotive application.
12V systems would be too low, 24V systems would be too high.
but works perfectly from 3x6V battery config, if you want a boombox.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.