2.1 amp with DSP -- preview/teaser

[Sydney]

... The only requirements for the tester would be a powered subwoofer, preferable one that is small, and some time to play with the different bass enhancement algorithms. I need to think about this some more, but it's got some potential for making the Act 4 phase an interesting "community" project.

This brought to mind the Sept 2016 Audioxpress article "The DSP Assisted Reflex System" and Elliptic filters.
Think on

[neildavis]

While I'm not in a position to do so; what would prospective tester(s) require?

I'm glad you asked that question. Not because I have a ready answer, but because it got me thinking. I've got 6 of the older DSP boards that I can give away. The older types used the 16-character LCD display, an 8-bit CPU and only have 4 channels. Functionally the same as miniDSP except designed as an Arduino shield and using different software. I think they would make great "bass enhancement algorithm testers". The only requirements for the tester would be a powered subwoofer, preferable one that is small, and some time to play with the different bass enhancement algorithms. I need to think about this some more, but it's got some potential for making the Act 4 phase an interesting "community" project.

[Sydney]

... and I think eventually I'll find a balance I'm happy with. Right now I don't have a plan, but I'm working on it.

IP ownership is a thorny issue these days.

especially anyone willing to help, by testing, contributing or evaluating.

While I'm not in a position to do so; what would prospective tester(s) require?

[neildavis]

No this:
Pardon the ambiguity

My mistake--thought the discussion had turned to the Sure board. Since some people had asked about that board, I provided some details on where that stood.

Regarding the code...I've benefitted from some of the open source code and I'm sympathetic to making some or even a lot of this DSP code freely available. But, it's been a lot of work, and at this point I'd prefer to keep this code "among friends", especially anyone willing to help, by testing, contributing or evaluating. I keep thinking about ways to give away some of it while protecting other parts, and I think eventually I'll find a balance I'm happy with. Right now I don't have a plan, but I'm working on it.

[Sydney]

I think you mean #32,....

No this:

but testing them and tweaking the parameters and evaluating the results could take a while. If someone wants to help, this could get done a lot quicker. I've got extra hardware that I could give away to someone willing to test and evaluate.

Honestly, I'm still agonizing over how much code I really want to put in public domain

Pardon the ambiguity

[neildavis]

I certain can appreciate the considerations stated in Post #30.

I think you mean #32, but, yes, some of the decisions Sure made are disappointing. Why use an oddball clock frequency? Why not use a more conventional 1V nominal input instead of 2V? The reconstruction filters may be a non-issue for class D amps that typically use a low-pass filter on the input (internal to the chip), and it doesn't matter for subwoofers. And there doesn't seem any artifacts on the headphone amp outputs according to the response shown in post #34. But all of that aliasing garbage on the rear outputs in post #34 will show up if you bypass the headphone amp and use amps without input filters.

Nonetheless, the Sure board is only $20. The chip itself sells for $11 in small quantities, and with the connectors, PCB and supporting parts this board would cost me about $35 in parts alone. Assembly with enough beer to finish the job would double that cost. So it makes sense to see if this board can be used, even with those odd design decisions. As I said, I'll probably try it, since I already have the parts I need, but it might not prove to be a good approach.

[Sydney]

I certain can appreciate the considerations stated in Post #30. -

but testing them and tweaking the parameters and evaluating the results could take a while. If someone wants to help, this could get done a lot quicker. I've got extra hardware that I could give away to someone willing to test and evaluate.

Honestly, I'm still agonizing over how much code I really want to put in public domain

[neildavis]

Neil,
How is your ADAU1701 DSP board differs from that produced by Sure Electronics? Is your s/w to to drive the DSP where does it come from, and is it the same as that used by Sure electronics?

The DSP board I'm using has an extra stereo DAC to provide a total of 6 outputs, plus it has a DC-DC converter circuit to allow direct connection to the 24V amplifier supply. It's also got a "socket" for the teensy Arduino boards -- either the LC or the 3.2 version. You can use a Sure Electronics DSP board for this plate amp, and I will probably try that, but I'll need to add this additional circuitry. Unfortunately, the Sure board has some "quirks" that require some modification to make it more suitable for this plate amp design. See this thread for more info: http://www.diyaudio.com/forums/digit...01-module.html

The software I'm using is an outgrowth of software I started writing about 15 years ago for the TAS3004. There is also a version for the STA326 series and the larger STA308a series, all written in 6801 assembly code. The code is intended to work with a wide range of DSP chips, including the ADAU1701. There is a companion PC application for designing multi-way loudspeakers and that provides real-time control of the DSP and allows loading any ADAU1701 program developed with SigmStudio.

Recently I translated all of the microprocessor code to the Arduino environment and added a local user interface for the ADAU1701. That allows using an LCD display to select crossovers and other DSP features without an external PC. If I end up making a version of the plate amp with the Sure board, I won't use any of their "software". They provide a program for the ADAU1701 that implements a variable subwoofer filter, but it is isn't much code and it's not very useful for this application.

[ttan98]

Neil,
How is your ADAU1701 DSP board differs from that produced by Sure Electronics? Is your s/w to to drive the DSP where does it come from, and is it the same as that used by Sure electronics?

[neildavis]

Any plans to code for IOS?

No plans right now for me, but the interface is so simple that anyone who wants to implement an IOS client can easily do so. I'm just using the MIT App Inventor, which is fairly limited, but all it needs to do is send simple strings and display the results on buttons or textboxes. The micro on the DSP keeps track of the state and tells the smartphone what to display, so the app doesn't need to do much.

I may try an IOS client using Visual Studio sometime in the future, but right now it's not on my short list. With a more robust app you could have a lot more control over the DSP. For example, instead of using one of the 9 "pre-set" EQ frequencies and 11 gain settings, you could specify any frequency and any gain.

[jwjarch]

Any plans to code for IOS?

[neildavis]

ACT 4.... Waiting for it... !

I already laid a foundation for Act 4 with that series of articles on the Audiodevelopers web site. But there are still a lot of details and examples, and that will be coming soon, along with the PE SBDT write-ups. What may take more time is extensive experiments with the various bass enhancement algorithms. There are a number of algorithms that ADI has available, but testing them and tweaking the parameters and evaluating the results could take a while. If someone wants to help, this could get done a lot quicker. I've got extra hardware that I could give away to someone willing to test and evaluate.

Honestly, I'm still agonizing over how much code I really want to put in public domain, because some of this might be marketable, and my son is looking for opportunities like that. But I will give the code to any person who asks for it.

[neildavis]

Lots of digital noise in the area with the switch mode PS and the amp itself. Digikey has a number of chip based isolated DC-DC converters. Most of the reasonable priced units don't put out a lot of power (~100ma + IIRC).

Nifty way to mount the amp to use the AL plate as the heat sink. You shouldn't have a problem there.

Not to worry, the physical wiring is good. The only time I've seen better is by a guy who builds old-school tube amps for guitarists. But it's his business and he does it every day on repeated designs.

The grounding isn't ideal because there is no easy was to electrically isolate the amplifier from the chassis/plate. The mounting holes are connected to the amplifier ground plane, and in order to get a thermal transfer from the heatsink to the mounting plate I needed to use 20mm threaded standoffs with a .5mm thermal pad. The analog input connectors are isolated from the plate, and there is no noise at all when the analog inputs are used.

But the antenna ground isn't isolated from the plate, and that is probably an issue. I haven't tried isolating the antenna yet--that might also work. The noise appears to be related to the networking RF signals--there is a lot of random activity in the noise. When I powered the module from an AC-DC converter (USB charger), the noise went away completely. I bought some DC-DC converters and will try them next. I need up to 300ma for the WiFi module, but there are a lot of small 2W converters to choose from.

I have three switching power supplies in there, plus the 6 channels of class-D amp plus the DSP, so there's a lot of noise getting generated. But it is that WiFi RF amplifier that is being unsociable.

[ani_101]

ACT 4.... Waiting for it... !

[Sydney]

... Act 2 will be the plate amp itself, and it's many challenges. I'll probably start that write-up next week.

Act 3 will be the "turn-key" software, including the smartphone controller. That's coming along OK, also. I can control the volume levels of each channel and the upper and lower crossovers from my cell phone, and I'm debugging the EQ. I should have most of the basic DSP functions and Android control done this week.

Act 4 will be the bass enhancement algorithms and other extensions, and that will be deferred until a later time. Act 4 is where this project gets really interesting to a hard-core DIY'er, as it will address how to modify the DSP and get the most out of the speakers. It will show how to use different ADAU1701 algorithms, how to get that code into the micro, and how to add new controls to the cell phone software.

Did you set a time frame for Act 4?