Wouldn't a USB connection presence sensing circuit be really easy to make and would be extremely low powered. It's really just sensing whether there's termination between + and GND, and if so turns on the higher amp circuit to provide, and as soon as connection is lost, powers down again. 5V at about 5-10mA should suffice.
As long as muted power consumption is held beneath the battery self discharge there's really no need to go further as up to that level is basically free anyways. I mean, the self discharge rate will discharge the batteries regardless of the power up to that rate is being used by the circuit or not.
As long as muted power consumption is held beneath the battery self discharge there's really no need to go further as up to that level is basically free anyways. I mean, the self discharge rate will discharge the batteries regardless of the power up to that rate is being used by the circuit or not.
Last edited:
Self-discharge is unwanted chemical reactions in the battery happening even when current is supposedly not flowing. From my experience with both SLAs and LiFePO4 batteries both will have a "free current" point up to about full (for SLAs) and 2/3rd (for LFPs). It's easily tested though. Just take 2 identical batteries, charge them to the same voltage, and put a low power circuit on one, being lower than the discharge rate, and the other will just reamin unconnected. After a month, measure both batteries. For a 12V 7Ah self discharge is about 15mA, so about the same as a low power single LED.
I dunno about 15mA... 7Ah/.015A = 466 hours = 19 days. The "general google concensus" is that 40%-50% self discharge happens in a year... going with the 50% number (3.5aH) that's 3.5Ah/(365*24h) = 400uA.
Granted, self discharge isn't constant. I imagine it varies with battery voltage, number of cycles on the battery, temperature, etc.
I've been killing myself trying to make my card do 100uA standby, but I'll probably move that up to 500uA or even 1mA.
Granted, self discharge isn't constant. I imagine it varies with battery voltage, number of cycles on the battery, temperature, etc.
I've been killing myself trying to make my card do 100uA standby, but I'll probably move that up to 500uA or even 1mA.
You might be right. I've only tested it with freshly recharged batteries and only for a few days. Maybe self-discharge is hyperbolic instead of linear. Or maybe it's because I've only tested with batteries in parallel which could also affect the result.
Either way, aiming for less than 1mA is pretty pointless as that would mean 291 days of standby. Any battery should be recharged at least every 3-4 months, or 3-4 times a year to avoid damage anyways.
Either way, aiming for less than 1mA is pretty pointless as that would mean 291 days of standby. Any battery should be recharged at least every 3-4 months, or 3-4 times a year to avoid damage anyways.
Using LT3680 for the +5V output, running all the time, estimated power consumption is 334uA. If the micro senses the battery voltage is getting low, it will shut that down and power consumption drops to 260uA. Most of that is the ADAU1592's (82uA each) and the LT3652HV solar charger (85uA). Think I'll go with this method... only things on the user interface will be a pair of USB jacks for charging, an audio input using jack switching for the enable, and perhaps a bicolor blinkenled which blinks with the battery state.
If I go with TAS5414A, shutdown consumption is around 100uA. I'm really debating this change since it means not having to solder QFNs - the next worst parts on the board are exposed pad MSOPs and they're not all that bad to do. Unfortunately the TAS pulls 240mA idle, 160mA with two channels active instead of four... while the two ADAUs pull 44mA/pop idle.
ADAUs are nice because they've got decent current draw, fast mute/unmute (planning on using this if the audio's connected but not playing, to save a bit of power), two of them won't beat together since they're self-oscillating, etc. Only real issue with them is the package.
And I just realized they're going obsolete. Dangit.
If I go with TAS5414A, shutdown consumption is around 100uA. I'm really debating this change since it means not having to solder QFNs - the next worst parts on the board are exposed pad MSOPs and they're not all that bad to do. Unfortunately the TAS pulls 240mA idle, 160mA with two channels active instead of four... while the two ADAUs pull 44mA/pop idle.
ADAUs are nice because they've got decent current draw, fast mute/unmute (planning on using this if the audio's connected but not playing, to save a bit of power), two of them won't beat together since they're self-oscillating, etc. Only real issue with them is the package.
And I just realized they're going obsolete. Dangit.
99% done on the PCB now - just have to hook up the microcontroller I/O and fix a couple of thermals. It's a ridiculously tight fit, but I managed to keep a decent layout.
PCB dimensions are 6x10cm (fits in a Hammond 1590N, barely) with 12-pin 3.5mm terminal blocks on each short side and four mounting holes in each corner. It's a 2 layer board, compatible with the free Eagle tools.
I've kept two ADAU1592s for the audio output, despite their obsolescence and annoying package. No other chip was jumping out at me with the same performance, operating/shutdown current, etc.
I'm also making a slight variation on the board with one terminal block replaced with "user interface" connectors (dual USB jack, audio connector, bicolor LED) which I'll be using myself. Hoping I can keep this within 6x10cm also to allow use with free Eagle, I haven't laid the board out yet with these changes.
I'll post the design files on here when it's done, most likely later tonight. This will be under a permissive license, probably Creative Commons, letting anyone change the design for their own purposes.
PCB dimensions are 6x10cm (fits in a Hammond 1590N, barely) with 12-pin 3.5mm terminal blocks on each short side and four mounting holes in each corner. It's a 2 layer board, compatible with the free Eagle tools.
I've kept two ADAU1592s for the audio output, despite their obsolescence and annoying package. No other chip was jumping out at me with the same performance, operating/shutdown current, etc.
I'm also making a slight variation on the board with one terminal block replaced with "user interface" connectors (dual USB jack, audio connector, bicolor LED) which I'll be using myself. Hoping I can keep this within 6x10cm also to allow use with free Eagle, I haven't laid the board out yet with these changes.
I'll post the design files on here when it's done, most likely later tonight. This will be under a permissive license, probably Creative Commons, letting anyone change the design for their own purposes.
Sounds really great. And I really can't thank you enough for doing this. My skills at electronics design are unfortunately woefully inadequate for such a task. Loudspeaker design is another matter entirely 
I will naturally try and see if I can't make one if my soldering skills are up to the task and use it for the signature Boominator version I'm building for this years festival season.
I also very much appreciate you putting all this effort into fitting it into the Hammond case although I know it must have been a herculean task. But it would be a shame if people who have already built it and future builders who are building it after the drawings available would be excluded from using it.
I will naturally try and see if I can't make one if my soldering skills are up to the task and use it for the signature Boominator version I'm building for this years festival season.
I also very much appreciate you putting all this effort into fitting it into the Hammond case although I know it must have been a herculean task. But it would be a shame if people who have already built it and future builders who are building it after the drawings available would be excluded from using it.
WRT soldering skills - there's a few "bitch components" on there:
- ADAU1592 (48 pin QFN) - contacts are 0.5mm apart, and there's a thermal pad underneath that has to be soldered.
- The three switchers have thermal pads underneath that have to be soldered too.
- Lots of fine pitch SMT (0.5mm, 0.65mm, 0.8mm)
- 0603 ceramics, which aren't that bad if you've got experience soldering them.
The card could be done 100% by hand, but it would seriously be a bitch to do. At the very least, I'd try to get access to a BGA rework system - we have one at my workplace that I'm planning to use for the thermal-pad parts.
Using solder paste, a stencil and a toaster oven is another practical option: Openhardware - Toaster Oven SMD
After prototyping is done, given enough interest a group buy on assembled PCBs is probably the best route.
- ADAU1592 (48 pin QFN) - contacts are 0.5mm apart, and there's a thermal pad underneath that has to be soldered.
- The three switchers have thermal pads underneath that have to be soldered too.
- Lots of fine pitch SMT (0.5mm, 0.65mm, 0.8mm)
- 0603 ceramics, which aren't that bad if you've got experience soldering them.
The card could be done 100% by hand, but it would seriously be a bitch to do. At the very least, I'd try to get access to a BGA rework system - we have one at my workplace that I'm planning to use for the thermal-pad parts.
Using solder paste, a stencil and a toaster oven is another practical option: Openhardware - Toaster Oven SMD
After prototyping is done, given enough interest a group buy on assembled PCBs is probably the best route.
Oh: one thing I'll add, I'm not expecting this to be a 100% turnkey board for the end user.
There's an onboard SigmaDSP part which does audio processing - same part used on a MiniDSP actually, but unlike the MiniDSP there's no nice, corresponding GUI to set the thing up. If you want to make any changes to the DSP processing, speaker arrangement, etc. you'll have to reprogram the board. This requires:
- An Atmel AVR programming pod - official AVR ISP MKII recommended.
- Analog Devices SigmaStudio (free off their website, but you have to register to get it)
- Atmel AVR Studio (free from Atmel) with WinAVR.
I actually encourage people to play around with this, because it opens up all sorts of interesting possibilities for experimentation/customization/etc.
I thought about putting a couple of DIP switches on there which select between a few basic modes of operation (2ch plain, 4ch plain, couple of biamp modes, etc) but I don't think I have room for it.
There's an onboard SigmaDSP part which does audio processing - same part used on a MiniDSP actually, but unlike the MiniDSP there's no nice, corresponding GUI to set the thing up. If you want to make any changes to the DSP processing, speaker arrangement, etc. you'll have to reprogram the board. This requires:
- An Atmel AVR programming pod - official AVR ISP MKII recommended.
- Analog Devices SigmaStudio (free off their website, but you have to register to get it)
- Atmel AVR Studio (free from Atmel) with WinAVR.
I actually encourage people to play around with this, because it opens up all sorts of interesting possibilities for experimentation/customization/etc.
I thought about putting a couple of DIP switches on there which select between a few basic modes of operation (2ch plain, 4ch plain, couple of biamp modes, etc) but I don't think I have room for it.
'fraid not. The USB connection is just a +5V DC output. Making it programmable via thumbdrive would require a micro with a USB OTG controller on it, and a lot of code. Putting a USB slave device on the board (say, ATMega32U4) so you could plug the card into a PC to reprogram it is a bit more practical, but I don't really consider it practical enough to do.
Relaxing the "AVR ISP MKII recommended" restriction, it looks like you can buy a USBTinyISP programmer for 10 bucks, including shipping, on eBay. There's cheaper ones on there but I wouldn't trust them.
I'm cramming a small, 2-slider DIP switch into the design to give four 'default' modes, as mentioned before. The four modes will be something like:
- One ADAU1592 disabled, one running but with no equalization (classic Boominator)
- Both ADAU1592s running, no equalization (double your power with 4 ohm speakers)
- Both ADAU1592s running, biamp mode. I'll set this up with my GW-1058/GT-1001 setup.
- Probably a second biamp mode with a different tweeter gain, cutoff frequency, etc - perhaps to suit a HP-10W/different-tweeter setup? Still have to think about this.
The default DSP program will have the following features:
- a sharp highpass to keep sub-bass out of the speakers
- soft mute/unmute
- automatic mute (no audio for 15 seconds, mute the ADAUs to save power)
- AGC for avoiding clipping (automatically turns down volume if overdriving amp)
The program running on the Atmel (ATMega168PA) is fairly straightforward - the chip handles low battery shutdown, blinking the status LED, reading switches/contacts ... and feeding the SigmaDSP its code and configuration via I2C. I figure it's a lot easier to program an AVR than an I2C EEPROM.
I'll provide the AVR code and example SigmaDSP code.
If you want to customize the DSP code to make the board do something different - drive subwoofers, do bass/stereo/etc enhancement, compression/processing for more volume, etc.. then you'll have to edit the SigmaDSP code. To do this you'll fire up SigmaStudio and make your changes, export a header file from it, recompile the AVR code to include the new SigmaDSP code and reprogram the AVR. You'll need the AVR programming hardware for this.
But again, it's $10 for the programming hardware, and the process isn't that hard. Hell, there's probably a ton of fun to be had with it.
- One ADAU1592 disabled, one running but with no equalization (classic Boominator)
- Both ADAU1592s running, no equalization (double your power with 4 ohm speakers)
- Both ADAU1592s running, biamp mode. I'll set this up with my GW-1058/GT-1001 setup.
- Probably a second biamp mode with a different tweeter gain, cutoff frequency, etc - perhaps to suit a HP-10W/different-tweeter setup? Still have to think about this.
The default DSP program will have the following features:
- a sharp highpass to keep sub-bass out of the speakers
- soft mute/unmute
- automatic mute (no audio for 15 seconds, mute the ADAUs to save power)
- AGC for avoiding clipping (automatically turns down volume if overdriving amp)
The program running on the Atmel (ATMega168PA) is fairly straightforward - the chip handles low battery shutdown, blinking the status LED, reading switches/contacts ... and feeding the SigmaDSP its code and configuration via I2C. I figure it's a lot easier to program an AVR than an I2C EEPROM.
I'll provide the AVR code and example SigmaDSP code.
If you want to customize the DSP code to make the board do something different - drive subwoofers, do bass/stereo/etc enhancement, compression/processing for more volume, etc.. then you'll have to edit the SigmaDSP code. To do this you'll fire up SigmaStudio and make your changes, export a header file from it, recompile the AVR code to include the new SigmaDSP code and reprogram the AVR. You'll need the AVR programming hardware for this.
But again, it's $10 for the programming hardware, and the process isn't that hard. Hell, there's probably a ton of fun to be had with it.
In a comment to standard set-ups. There's basically 4 standard set-ups
- Classic - HP10W (original version) + HP100 (or similar) piezo (both of these are no longer available anywhere in the world)
- Revised (North American version) - GW1058 + KSN1001A (or replica)
- Revised (Rest of World version) - HP10W (current version) + KSN1001A (or replica)
- Signature - Basslite S2010 + PHT407
Only the signature version really requires bi-amping but I'm not certain about if the limited dissipation will limit output when it's basically only one channel on each that is heavily loaded given that both the HP10W and GW1058 will dip below 3 ohms when paralleled, and the channel with piezos will only be loaded with about 100 ohms when paralleled).
Most that have already built the Boominator will have the piezos directly coupled to the HP10Ws and only one cable going to the amp so that should be accommodated.
So with the above in mind:
- Classic - both ADAU1592s paralleled, no eq
- Revised (NA) - both ADAU1592s, bi-amped, eq to match drivers
- Revised (EU) - both ADAU1592s, bi-amped, eq to match drivers
- Signature - both ADAU1592s, bi-amped, eq to match drivers
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.