freeDSP V2.0 (ADAU1452) developement thread

greetings,

did you already try out sigma_tcp? during waiting for my cypress fx2 programmer i tried it out with raspi but didnt continue working on it because in the meantime the programmer arrived. somebody already forked it for adau1452- with this one i had compiling issues but maybe you can work it out.

sigma_tcp:
SigmaStudio Network Utility for Linux [Analog Devices Wiki]

fork for adau1452:
GitHub - aventuri/sigma_tcp: a C gateway IP/I2C between Analog Device SigmaStudio and ADAU DSP

the original sigma_tcp startet to work and i could build the bridge- there only seemed some i2c adressing issues left, but i cant tell.

(sorry if this is already postet somewhere here in the thread, i didnt have time to read the whole thread)
 
drive ADAU1452 with Sigma Studio thorugh a Linux SBC

hi,
currently there's a tool, called sigmatcp that should let you configure/control an ADAU1452 from sigma Studio passing on TCP/IP and a Linux SBC. this original code used not to work for me so..

..here it's my git repo with a fixed version that i use on an A20 Allwinner SBC Cubieboard.

it's plain i2c writes, so i suppose it should compile/work on RasPi with proper toolchain.

bests


Your idea fits really nice! The original idea was for Pi to be just playback host and programmer - in a sense that it writes to self boot eeprom trough the DSP.
 
Hi,

I just wanted to point out something quickly that caught my attention. Windows 10 IoT is available for Raspberry Pi models for $0. This might be another way to control the PiDSP, and we may be able to harness the capabilities built into SigmaStudio because it can run Windows Applications.

The Win 10 IoT doesn't run on the PiZero, so we won't get USB OTG to make the Pi a device. None of the other Pi Models have USB OTG modes available because they have a USB HUB integrated into the USB buss. But we could still use TCP/IP from the host to control the Pi to control the PiDSP over ethernet.

I just wanted to share my recent thoughts.

Jay
 
DSP Preamp

What To do now?
When you've got a (free)DSP?

Well you really need to give a lot off thoughts. A DSP give us so much possibilities it is amazing.
I put this here because I would like to inspire Pitrsek a little to let him know how much we appreciated everyone working hard on such a great and time consuming project.
Also it might inspire others to get involved as well
My goals where:
1 build active speaker
2 use DSP to equalize the used single loudspeaker.
3 add 3D enhancement.

I game up thinking it would be nice to make a preamp with the Freedsp board.
In this way I would have multiple inputs and the ability to reach the goals.
Personally I am a fan of using real switches and potmeters.
Also with a wink to Crown Preamp I wanted to have a SPI led (Signal Presence Indicator) and a CLIP led (IOC).
All the used schematics are based upon the documentation from Freedsp and Analog devices.

Have much fun and with building your own project.

For my next project I would love to use the adau 1452 board so to have digital inputs and also love to have a balanced output board ect ec
 

Attachments

  • dsp-pre-sch.jpg
    dsp-pre-sch.jpg
    126.8 KB · Views: 1,084
Hello FreeDSP followers,

I'm designing some audio modules. Specifically some ADC and DAC parts. I want to build these modules compatible with the FreeDSP expansion header so that maybe it would be useful for a few different MCU/DSP boards. I also happen to be designing an Audio Bridge for the Teensy, because it's working for me now. But that's not what my question is about.

The FreeDSP expansion connector... Is this design confirmed? Does it work well? What are the Pros and Cons to it? Is there anything you would change? I find this header connector to make a ton of sense for many different projects. I want to design some audio modules, bridges, and development parts so that I can contribute some things to the community around digital audio mixing and I want to make sure this design is working. I designed a similar "Audio Bridge Port 1.0" specification that I haven't finalized. But the idea is to just make it easier to connect audio modules to DSPs.

I have a discussion over here about these concepts I'm working on. But I reach out to you here today to see if you had any specific recommendations on my Audio Bridge Port. My current question is about the entire ROW of GND connectors. Is that necessary? See my post for my inquiries.

http://www.diyaudio.com/forums/digi...audio-bridge-i2c-i2s-lines-2.html#post5505660

Thank you for your insight!
 

Attachments

  • audio_bridge_port_1_freedsp.png
    audio_bridge_port_1_freedsp.png
    29.4 KB · Views: 755
Last edited:
Hello FreeDSP followers,
The FreeDSP expansion connector... Is this design confirmed? Does it work well? What are the Pros and Cons to it? Is there anything you would change? I find this header connector to make a ton of sense for many different projects. I want to design some audio modules, bridges, and development parts so that I can contribute some things to the community around digital audio mixing and I want to make sure this design is working. I designed a similar "Audio Bridge Port 1.0" specification that I haven't finalized. But the idea is to just make it easier to connect audio modules to DSPs.
Thank you for your insight!

Hey, not sure if the topic is still open, my two cents:
The design is confirmed, and is being actively used and works well. The signal ground signal configuration is good for signal integrity as well as for EMI.
I think it's a good compromise with regards to size and use and emi. the I2c connection offers a possibility for a port expander, if you need GPIO/HW configuration. I do not know if you consider audio bridge port for ribbon cable connection, but if you do, I would not send low voltages over it - contact resistance and crimp quality might byte you. I If your power supplies are good(low flat impedance), your vcc pin is as good as gnd - with regards to SI and EMC. For ribbon cable connectors it's of not much use (1 pin saved), but for stacking bards this can be used to a greater advantage.

Personally I would keep the FreeDSP expansion connector, and create low voltages locally. It's way easier in the end.
 
I have decided to use the ADAU1452 and the codec and the same parts on the evaluation board. The CAD program that I have is very limited. I am trying to find a board that is about the size of my shirt pocket. I am at the point of needing hearing aids. I only need one stereo channel. Commercial hearing aids are for people who have lots of money. $4000 Does anyone know if there is a small board that uses the ADAU1452? I will take the output and feed the stereo signal into a Bluetooth transmitter. I am using a phantom Sony mike for the input.


Any help would be helpful. If someone is selling a DSP:confused: and codec board? Please let me know. Or send me a link where I can find a solution. I am on Social Security income. So $4000 is a joke.
 
Last edited:
Hey, not sure if the topic is still open, my two cents:
The design is confirmed, and is being actively used and works well. The signal ground signal configuration is good for signal integrity as well as for EMI.
I think it's a good compromise with regards to size and use and emi. the I2c connection offers a possibility for a port expander, if you need GPIO/HW configuration. I do not know if you consider audio bridge port for ribbon cable connection, but if you do, I would not send low voltages over it - contact resistance and crimp quality might byte you. I If your power supplies are good(low flat impedance), your vcc pin is as good as gnd - with regards to SI and EMC. For ribbon cable connectors it's of not much use (1 pin saved), but for stacking bards this can be used to a greater advantage.

Personally I would keep the FreeDSP expansion connector, and create low voltages locally. It's way easier in the end.

Hello Pitrsek, thank you for your response. As you might know I'm designing an audio mixer. My goal is to create a modular system that would allow me to choose and pick different modules based on the inputs and outputs required. Rather than designing an 8x8, 2x4, 6x6, etc board I am looking to modularize the Codec, adc, dac parts so that I can install as little or as many pieces of each part I need for the project.

The idea is that by doing this way I might get some support from the community that would help me move my project along. Also, I envision (and have not realized) some cost savings manufacturing wise. I figured if I were to create 100 pieces of a single modular PCM5242 for example, it would be cheaper than building a 4x board that may or may not be the right specs for any particular user. So ultimately, if I had a stock of 100 PCM5242 boards and a user needed four outputs, then it's easy to sell 2 boards with 2 outputs on each. For users needing 8 outputs they can buy 4 modules. I'm not looking to build a business off this but I need to mass produce this to get the costs down and then sell some to offset that cost. Depending on how much support I get from the community the hardware and or software would be open. That being said, I'm working on a concept for an "Audio Bridge" and then "Modules".

Your response is greatly appreciated. It looks like sending the I2S and I2C lines to the modules over ribbons is OK. But it sounds like you wouldn't recommend the design as drawn in this schematic (see attached)? It sounds like you don't like the idea of sending the low voltages over the ribbon in this way? My only concern about this is now each module will have to have it's own LDO. It's just more parts. Of course I want this to work right but I want each module to be as simple as possible - like a cheap breakout board. If I put an LDO on each module it just adds to the parts on each module. What do you think? Is this schematic acceptable?

PS - I know the "audio bridge port" isn't drawn exactly like the FreeDSP Expansion Connector. Just look at what it's carrying, not it's organization. I think you might have sold me on organizing it with a row of grounds, but I'd like to include the low voltages... Even if I were to have a standard FreeDSP expansion connector and then a second row of 1.8, 3.3, 5.0v, and some GPIO. No? Yes?

PPS - (GPIO is for some extra features in ADC/DAC parts such as a "RESET" in the TLV320AIC3104 or a "XSMT" aka "soft mute" in the PCM5242.)
 

Attachments

  • TEENSY_BRIDGE_C1.pdf
    85 KB · Views: 150
Ribbons, especially in S-G-S-G... are a very fine connection - provided you keep the length reasonable. With I2C bus - keep in mind that there are limits how much capacitance the bus can tolerate - that translates into limit on cable length and/or number of modules.

Local LDO is not a bad thing. Impedance is low, every module has it's own short circuit protection, no crosstalk/dustrubances over the power supply lines. And It's not expansive. If you share LDO for different boards, the total power is limited by your LDO. Your current limit is given by your ability to cool the LDO. Do the calculations, it's not that much... If you let module create it's own voltage, power is limited by the adapter.

What if I need 1.2V, or 6V for some reason? LDO are cheap. IMHO The biggest PITA is the negative voltage for opamps. I would not be too hung up on the parts price/count, if there is few $ in bom, nobody cares(as long as it's not a 1000s series bussines). Place where I used to work had a bad experience with low voltages over ribbon cable(contact resistance).

With regard to GPIO - I was considering providing them on FreeDSP connector, decieded not to based on connector size - would not fit 3x to eurocad board edge, and there is a limited amount of GPIO in ADAU dsp. For most part, I2C is all you need. And if you for some reason really need indpenedend GPIO from the DAC/ADC, i2c port expander is again 1$ in single quantity...
 
I have been working on using a few different concepts which could result in a 'longer bridge' than ribbon cables and thus serve as an extendable interconnect, while also bringing galvanic isolation.

Optical SPDIF ... a working board that could output i2s to the FreeDSP board
CS8416 Source Selector Module 2 SPDIF Inputs to 1 I2S Output - Audiophonics

A few other concepts utilizing AVB (Ethernet) and ADAT optical

XMOS based ADAT encoder/decoder
https://www.xmos.com/download/private/AN01004%3A-ADAT---Replacing-Wavefront-AL1401-402-with-xCORE%281.0.1rc2%29.pdf

XMOS ... AVB to I2S
https://docs.wixstatic.com/ugd/fe41be_8ab4eee8e9ff4e108cf9b72281310ad6.pdf
 
Is it possible to update the design files on Github with the B3 version?

Old thread, which is slightly disconcerting, as it's supposedly the official discussion thread for the PiDSP, but I still don't see a B3 there.

Anyway, I was wondering if someone could also convert the PiDSP from Altium to KiCAD, which the FreeDSP project is supposed to standardize on anyway, and then verify that the KiCAD version is indeed equivalent?
I can run a free converter as well as anyone else, but I don't trust myself to verify the result.

Thanks!


Edit: Perhaps this MIT-Licensed driver can also be adapted to work with the PiDSP, so that it no longer needs a separate programmer? The Pi then becomes a bridge from LAN to the DSP's programming interface, much like the freeUSBi is a bridge from USB to the same place. Just point SigmaStudio to the Pi's IP address, and go:
https://www.hifiberry.com/wp-content/uploads/2017/10/Screenshot-2017-10-20-11.28.37.png
 
Last edited: