This just got more interesting with a much lower price-
EVAL-ADAU1452MINIZ Analog Devices Inc | EVAL-ADAU1452MINIZ-ND | DigiKey
EVAL-ADAU1452MINIZ Analog Devices Inc | EVAL-ADAU1452MINIZ-ND | DigiKey
I did say at the start of this thread that it was never my intention to compete in the minidsp space. If a $50 DAC could offer substantial benefits and improvements over a $5 DAC then I would use it which is exactly what I have done. Similarly for the ADC and much of the other hardware such as DSP, clocks etc.
Also the CODEC on this board seems to be hamstrung to a maximum sampling rate of 48K which kind of limits its performance.
cheers
Not to be rude but, are you not running TI parts? I do not know one engineer who will use anything with a TI on it. Burned to many times by parts not meeting spec or failing prematurely.
So you suppose running 48kHz sample rate is not enough to do audio justice? How was that determined?
I have found DAC and ADC price are moderately unrelated to performance for audio. I recall a time past when so many were drooling over some DAC used by many and extolling the benefits of that part. I got one in and tested it and found exactly what the spec sheet indicated by my interpretation, 24dB signal to noise from the DAC. Even Wadia was claiming fantastic numbers which when tested showed the 24dB figure.
There is a lot of fantasy in the DSP world with big numbers assumed to mean big sound. It simply is not so. Numbers grabbed from thin air and printed like 103dB of dynamic range. Never seen. Never measured either. Using dynamic range to mean signal range is a fallacy commonly used. By that standard my MC phono stage is 138dB which it is not. That is the signal range at 1kHz.
But I digress, modern DSP reference standard is built by Peavey in their Media Matrix division. Peavey was selling DSP when everyone else was talking about what one might do with DSP.
Further, speaking of DAC have seen a lot of faulty implementation of the DAC output stages. A good DAC and a bad op amp or op amp arrangement ends up being really bad. Say one chose to use the antique NE5532 or 5534. Well no matter what else is done it will sound bad because that is the only choice for that part, bad sound. I have a Casio D/A here that is dirt cheap and better than almost any. Then I have good ones also used in professional equipment with excellent op amps which have found no better. Thank you Media Matrix. Then there are good DAC and ADC like in the m-audio products coupled with the aforementioned horrid opamp. Change the op amp and put good caps in the supply and it becomes amazing. Suitable for use with Acorate software for incredible sound.
As for that cheap demo board yes, 48kHz but not for the device or independent implemented system. Those will go to 192kHz but that does nothing but fill up the hard drive unless there is ultrasonic work happening. Then consider any one inch dome without a lens of some sort flips phase about 5500Hz and nothing above that frequency means anything so why be concerned with 22kHz reproduction? Oh yah, reality versus fantasy. Most chose fantasy because fantasy can be anything one wants it to be and never mind the facts. Can I sell you a tapped horn?
In the end this EVAL kit is a great buy for anyone interested in DSP which is why I pointed this product out. If one is spending money freely then Media Matrix is about impossible to beat. There is almost nothing in between these two worth considering. Maybe some day you will have something but, if it says TI on it will tell everyone to avoid, sorry, reality again.
So you suppose running 48kHz sample rate is not enough to do audio justice? How was that determined?
I have found DAC and ADC price are moderately unrelated to performance for audio. I recall a time past when so many were drooling over some DAC used by many and extolling the benefits of that part. I got one in and tested it and found exactly what the spec sheet indicated by my interpretation, 24dB signal to noise from the DAC. Even Wadia was claiming fantastic numbers which when tested showed the 24dB figure.
There is a lot of fantasy in the DSP world with big numbers assumed to mean big sound. It simply is not so. Numbers grabbed from thin air and printed like 103dB of dynamic range. Never seen. Never measured either. Using dynamic range to mean signal range is a fallacy commonly used. By that standard my MC phono stage is 138dB which it is not. That is the signal range at 1kHz.
But I digress, modern DSP reference standard is built by Peavey in their Media Matrix division. Peavey was selling DSP when everyone else was talking about what one might do with DSP.
Further, speaking of DAC have seen a lot of faulty implementation of the DAC output stages. A good DAC and a bad op amp or op amp arrangement ends up being really bad. Say one chose to use the antique NE5532 or 5534. Well no matter what else is done it will sound bad because that is the only choice for that part, bad sound. I have a Casio D/A here that is dirt cheap and better than almost any. Then I have good ones also used in professional equipment with excellent op amps which have found no better. Thank you Media Matrix. Then there are good DAC and ADC like in the m-audio products coupled with the aforementioned horrid opamp. Change the op amp and put good caps in the supply and it becomes amazing. Suitable for use with Acorate software for incredible sound.
As for that cheap demo board yes, 48kHz but not for the device or independent implemented system. Those will go to 192kHz but that does nothing but fill up the hard drive unless there is ultrasonic work happening. Then consider any one inch dome without a lens of some sort flips phase about 5500Hz and nothing above that frequency means anything so why be concerned with 22kHz reproduction? Oh yah, reality versus fantasy. Most chose fantasy because fantasy can be anything one wants it to be and never mind the facts. Can I sell you a tapped horn?
In the end this EVAL kit is a great buy for anyone interested in DSP which is why I pointed this product out. If one is spending money freely then Media Matrix is about impossible to beat. There is almost nothing in between these two worth considering. Maybe some day you will have something but, if it says TI on it will tell everyone to avoid, sorry, reality again.
"I do not know one engineer who will use anything with a TI on it"
That seems to be a very over the top and unreal statement to say the least. Sounds like you are just propagating more audio mythology here yourself. Show me almost any music made in the last 30 years that doesn't have some 5532 op amps in the audio chain at some point! Go tell D. Self he doesn't know what he is doing since he has used those 5532's for years in his designs.
That seems to be a very over the top and unreal statement to say the least. Sounds like you are just propagating more audio mythology here yourself. Show me almost any music made in the last 30 years that doesn't have some 5532 op amps in the audio chain at some point! Go tell D. Self he doesn't know what he is doing since he has used those 5532's for years in his designs.
Implementation is 50% of the use of any devices and is often ignored. When I see a statement like nobody would use a TI part that makes no sense if you look in so many devices, they are everywhere. If there quality was that bad they would be long gone by now in the electronics industry. Six sigma and lean manufacturing methods pretty much precludes that in any large company these days.
Also esstech have an app note for the Sabre DAC and the NE5534/NE5532 feature in it as one of possible opamps for post DAC processing. Surely the designers of the Sabre DAC would not recommend something that doesn't work and throw all of their good work away with a bad opamp ?
http://www.esstech.com/PDF/Application_Note_Component_Selection_and_PCB_Layout.pdf
cheers
Last edited:
Back in the 1980s TI was the number one part manufacturer in the U.S. yet was number 10 in U.S. sales. Only newbies and the ignorant would purchase TI parts then. Every engineer I know now says exactly the same thing, "No TI parts. There is always another way to make the circuit." Without exception I have never had a TI part work correctly or it has failed prematurely or not worked right to begin with. Every time I have been burned as every engineer I know has been burned by crappy TI parts. I do know quite a few aerospace engineers and none of them will use them. TI is the company that could not quite get it right. When TI took over National Semi I ran out and bought a lot of new old stock of original National Semi parts I used. TI always means inferior design, lacking quality, and poor implementation across the board of every part they make. Use garbage if you want. Not my problem. Nuff said.
Well this is news to me. I've been using TI parts since the 80's and I haven't experienced the problems that you have had with them and that includes both analog and digital parts, nor do I know anyone else who has had similar problems to you. Sure, sometimes you occasionally get components that don't quite meet the specs but this could be from any vendor.
Regarding the NE5532 these are Signetic parts that are now made by numerous other vendors including TI. About the only complaint I have heard about these is from Douglas Self who stated that the TI part has slightly higher distortion than say the same part from OnSemi etc but there was no mention of reliability issues.
Lets put it to the test. Does anyone else have any problems with the TI branded NE5532 parts which are usually labelled N5532 ? Similarly with Nat Semi branded parts such as the LM4562 etc
cheers
I've had plenty of success with TI parts since the late 70s, both digital and analog parts. Never had any fail if used within spec - and I'm thinking about this quite intensely now, I've used hundreds of TI parts and never had one fail when used properly. My experience is that they are as reliable as NS or any other genuine part.
I don't think for a second that all the audio devices that are designed with the NI3886 are now going to be pulled from the market because TI is now producing those parts. This truly does sound like a personal grudge to me, not a realistic situation. There are millions of products out there using TI parts. Now if you are talking about knockoff Chinese counterfeit parts that is another story completely.
Anyone working for any large or medium sized company would understand that 6-sigma standards are the standard of the industry today. That would discount the notion of high or even substantial failure rates on current electronic components from a company following current quality standards. Try and find a company not following Six Sigma and lean practices, it is almost unheard of today. TQM is everywhere today, the days of high rejection rates are over with any company that has to work on the international stage, an engineer allowing that to happen on their watch would soon be fired. I'm not exactly fond of all the mergers and acquisitions as I think it stifles competition and innovation but that has nothing to do with quality issues.
Stereo mode connection of the DAC offers 6dB reduction in uncorrelated noise and little or no improvement in THD so there is really not much to gain over multi-channel mode.
cheers
Funny, I know quite a few aerospace engineers (I lay boards out for them)
and they use Ti parts, I also no a lot of engineers that do medical and other high reliability products and they use Ti parts, in fact its rare to see a design these days without a few Ti devices on there, and we see a lot of designs from many sectors of electronics... It seems strange that every engineer you have met has said no Ti parts, very strange indeed
According to the specs on the ES9018 the DNR for the chip configured as mono output is 135dB and that for 8-channel output is 129 dB so the difference is even more marginal than what I thought !!
cheers