Am i missing something? is this just another case of eastern asian OEM gatekeeping? how is one supposed to code for this dsp? a DSP from AKM has been used in mass production that i know of in the onyx studio 3. so there must be SOME way of coding it. i don't believe Harman Kardon used Audioweaver or did they? did AKM go through the trouble of manufacturing just to hinge their only chance of widescale implementation on a THIRD PARTY?
Please i'm at a loss and need group wisdom. does anyone have ANY experience or idea with these chips - and how to program them? thank you.
I have never heard of Asahi Kasei Micro before, so thank you for pointing out their web site. A site search turns up nothing on that chip for eval board or tools. Send the company an email, or even call them on the phone. They might be surprised to find out there are 1000's of engineers who can find uses for their products, if they help them use the product. If they want to grow their company, they will actually help you.
Hmm... AK7755EN datasheet: https://www.akm.com/content/dam/documents/products... mentions an evaluation board https://www.datasheets.com/zh-cn/reference-design/... which also includes control software.
Yes, and web searching shows Arrow has it as a reference design, but not as a purchasable part. So do you have to build the board and order the parts then assemble the design to get an eval board? And how do you get the control software? I see the youtube advertisement, but Arrow does not mention software.
The point of the OP is still valid I think. It should be easy to get eval boards and test software. The information should not be buried at the back of a datasheet.
It really should be easy to access right? Honestly, I don’t know what they are hiding. Texas Instruments makes exponentially more complex DSP chips and still has documentation on basically every aspect. It clearly can’t be an intellectual property issue, so it’s probably just retained for high >10ku OEM orders and not available in small quantities. I personally don’t think that’s good either because if the IC fails in the final product, repair shops/diy repairers will have a hard time finding replacement parts. But that’s just me.
The block diagram shows coefficient, data, and delay memories, which imply a focus on digital filters. The "PRAM", which I assume is program memory, is small, only 4k words. I might guess that compared to ADI and TI DSPs, the AK7755EN is not really -- or not nearly as -- programmable like you are thinking a typical DSP should be. Probably you can program it with different sets of filter coefficients and different sequences of operations (filters, peak detection, maybe some relatively simple algorithms like echo cancellation or multiple microphone handling, maybe some user-defined difference equations) and that's about it.
My suggestion would be to write to Asahi and ask for some example source code, something simple, like their version of "hello world" and see what it looks like. Maybe it's only asm with no C compiler ...
Yes. The chip turned out to be underpowered and we just went for a TI purepath part instead. sigmaDSP would have been a good choice but the avialable chips just didn't have the I/O needed. TLV320AIC managed just fine! thanks.
Sounds like good progress ! I hope the TI chip works out. One good thing about TI is the chip usually meets and exceeds data sheet specs, and on occasion when they miss they come out with errata fairly fast.-Jeff
For anything other than dirt cheap AKM are pretty much the only source I consider for high performance audio ADC/DACs. To me this DSP seems to be aimed at large volume OEMs. IMHO for smaller volumes SigmaDSP or a decent ARM chip is probably the way to go. Or if it's complex DSP but you want software vs. FPGA then ADI's SHARC chips.
Actually I also like their super straightforward converters and the velvetsound series of DACs is awe-inspiring but it’s a shame their DSPs don’t receive the same documentation treatment from AKM themselves. In a custom design, implementing AKM ADCs or DACs is easy. Can’t say the same about any of their DSPs. SigmaDSP was on my mind but the lack of a second I2S bus on the chips priced around my set price point made me turn to TI. I found a purepath miniDSP chip that worked just fine and wasn’t too costly.