Michael: Please do rain on my parade. I live here in sunny southern California and and any rain, parade or no is a gift of the gods. It keeps us from reverting back to the dessert. So go ahead and downpour if you would-please. Joking aside, I do appreciate some skepticism for what I am attempting to do. Now that is what science is all about, isn't it. I have put my entire working and World War 2 experience in the electronics field and saw the beginnings of radar and TV. Worked as an electronics maintenance technician at Chicago's Midway and O'Hare airports. Following which I returned to the University where I received bachelor and master degrees from the University of Southern California. After graduating, worked in the defense industry building the latest airborne radars for aircraft such as the F14 and F15. Does all this mean I know anything? Not really. As I approached 85 years, as is typical, my hearing got poor, designation is 'hard-of-hearing' (HOH). After using several brands of hearing aids, I too was complaining "I hear, but I don't understand". And quickly deduced that I was not hearing consonants, for reasons unknown.. Went shopping on the Internet to find a hearing aid that advertised the two precious words "amplified consonants" because that is where the information is contained. No such device was to be found. Thus as I now approach my 90th birthday I set to work to figure what this confusion (in my mind, at least) was all about.......And now I have some certainty that I have solved the enigma and believe I am on the proper road to fixing the 'I hear but I don't understand' syndrome. The very finest of today's hearing aids including those with DSP) do not resolve that problem. If you feel that the digital signal processor can assist in this matter, I am an anxious listener. My particular solution does not introduce frequency as a key parameter, but uses a new type of modulation as it's basis ....If you wish to exchange ideas on the subject that would be very advantageous to me and hopefully also to you. .Best Regards, Nate Almond
I picked up from your earlier posts that you are planning some signal processing involving ADC and EEPROM and that you don't have any hardware yet. If it were my project I would design and make a custom board but it doesn't look as if you have the resource to do that. So my suggestion was that the cheapest way for you to experiment is to use a ready made DSP development kit - that way you can do almost any imaginable signal processing - frequency domain, time domain, linear, non-linear etc. But the great thing about DSP is its all in software so you can change things cheaply and quickly at the start of the project. In the end you might decide to dump the DSP and use pure analogue for a production design (although if you want a small hearing aid the investment cost in using your software on an existing DSP willl be much lower than a custom analogue solution).
II am looking to build a headphone amplifier with the following specs.
1. 2 1/4 inputs - Channel 1 and Channel 2
2. Each input channel would have 3 controls
a. Master output to headphone
b. channel 1 volume control
c. channel 2 volume control
3. Individual headphone output for each channel
Any ideas of where I can get some information on how to build this.
I am a novice.....
Ken: This is the old engineer guy who you have helped on several past occasions. Once again asking for some sage advice from you. Would you have any clues for me of a simple circuit design for a digital audio amplifier that I might construct in order to test my unique concept for an improved understandable hearing aid? It would have to include as a minimum an A to D converter and a EEPROM. Thank you for any help you can provide. Nate Almond
I don't know of a single device of the type you describe. If one existed, it would have to have separate ability to program the EEPROM.
TI has some digital input Class D amps, but high power ~ 20W. The digital input is in the I2S format, commonly used in consumer electronics audio equipment. This type of part does not have any internal EEPROM that could be programmed.
Maybe the best bet to prove the concept would be to use a D-A after the external A-D and EEPROM, then feed the new HOH analog input a part like the National LM48310 or other analog input Class D amp at http://www.national.com/cat/index.cgi?i=i//277 or some similar parts at TI (TPA2001) at http://focus.ti.com/paramsearch/docs/parametricsearch.tsp?family=analog&familyId=923&uiTemplateId=NODE_STRY_PGE_T. You could hear the result on an 8 Ohm speaker, or thru a headphone if you chose a headphone amplifier from either of these 2 suppliers.
From the brief description it sounds like you want to build a very high quality "audiophile" type product.
National Semiconductor builds some high performance amplifiers (see list at http://www.national.com/analog/audio/high_performance) .
On page 10/11 of the datasheet for the LME49600, there is a description and schematic of a headphone amplifier
( http://www.national.com/ds/LM/LME49600.pdf ). It has separate input jacks and volume controls for each channel, and single output jack for the two channels. But you could change that to separate output jacks.
If you want to talk to someone who is an expert on this type of product and on these particular National high performance parts, may I recommend Mark Brasfield (email@example.com ). He may even sell you one pre-built. Definitely worth a conversation, and probably your quickest source of definitive information about headphone amplifiers.
All the class-D ICs I have seen the datasheets of seem to give outlandish figures for power, only to find out at the back of the datasheet in on of their charts that distortion is in the double figures at anything approaching moderate. Do you see Class D as the future of audio amplification?, or do you see some other class/hybrid class becoming the standard in future audio amplification?.
John, Farnell Technical Team.
I am wanting to get into Digital Audio, I have a good background in electronics from uni, but I did not take the Z-Transform option on my course!, good basic information is sadly lacking on the internet, I would like to start from theory, like how I would go about replicating say a Butterworth Filter Digitally, I am also looking at dsPICs as an introductory way to get into DSP.
Any suggestions would be appreciated.
John. Farnell Technical.
You have asked an excellent question, but the answer is not so simple.
One of the principle reasons to use Class D audio is to generate large output power without having the power dissipation and heat generated by Class A, B, AB, G, and H audio amplifiers. So Class D is a very attractive option.
The often forgotten spec in audio is the infamous 10% distortion figure for output power. This is an industry standard and is a common (almost required) spec point in all audio amplifiers. You will find it in data sheets for practically all audio amps. In case you didn't know, the 10% distortion figure is THD + N, or Third Harmonic Distortion + Noise, and is the level at which most people will clearly notice that the sound is distorted somewhat. This is because the 10% level is also the level where the output signal waveform tends to start clipping.
Unless something else is radically wrong with the amplifier, the THD+N drops off as you lower the output power. So the industry has a 1% distortion spec as well on amplifiers.
I have heard some very good Class D amps but all were operating at 1% distortion level output power, or even less.
For audiophile quality audio, most people will not choose Class D since they are sure (1) they can "hear" the distortion even at lower output powers, and (2) somehow Class D amplification will modify the sound vs. pure linear amplification, and some other reasons that are hard to prove technically. Kinda like gold plated audio cables sound better :>)
As equipment becomes smaller and lighter, choosing linear amplification for audio is a hard sell due to lower efficiency and heat dissipation. So.. increasingly, Class D is being chosen. A good example would be the "bookshelf" type amplifier systems in the home versus the older larger equipment.
A good example would be audio amps in thin TV's. It is well known that if heat generating linear audio amps are close to the LCD screen, the color near that location can be slightly modified and noticeable. Also, power supply requirements are bigger if those are linear, and they also generate heat. Class D allows TV makers to put acceptable sound in a thin package.
Cell phones and items of that type also benefit greatly by using Class D amps, even at 1W (1% THD+N) to 2W(10% THD+N). In today's environments with large amounts of ambient noise, having higher output levels of sound is a good thing. Some of the Class D IC amps today have 10X the output power (10% THD+N) in the same die size as older linear amps at 0.5W.
To avoid using Class D amps, designs have appeared using variations of linear amps - Class G, and H. These types of amplifiers attempt (and succeed) to reduce power losses and heat generation by varying in some way the supply voltage to the amplifier based on the input signal level. The higher the input, the higher the supply voltage is allowed to go, and vice versa. But they do not match Class D in efficiency or heat reduction. But they are linear and can have better power output at low THD+N. Class G switches the supply voltage based on some input level or levels in multi-level Class G, while Class H tracks the input signal and changes the supply voltage accordingly.
The other thing to remember is that Class D is not that beneficial at very low output powers. An example would be headphone only amplifiers. Most of these are linear, either capacitor output coupled, or direct coupled (+ and - supply voltages internally generated by a charge pump circuit). Since the headphone or earbud is very close to the ear canal, any distortion and Class D switching noise would be easier to hear and be downright annoying.
By the way, if you think datasheet high output power ratings are somehow misleading, try figuring out what Peak Music Power Output (PMPO), Peak Envelope Power, RMS power, Sine power (the most reliable measure of the output power capabilities of an amplifier), DIN power (European standard) means when you see them on some consumer audio products. An example would be a small bookshelf type speaker amp that offers 100W PMP when the amps clearly can only put out less than 10W Sine Power.
I hope this answer helps you more than add confusion to the mix.
The term digital amplifier is one of those terms that can mean many things. I prefer the following:
All Digital Amp:
This term means that the audio input to the amp is in digital form, usually via a recognized digital audio format like I2S, and that the output is operating in Class D mode where the output signal switches between the voltage rails (+ and - supply, or + supply and GND). But basically it means that the amplifier should be able to take in digitized samples of the audio at some particular sampling rate, and thru internal means modulate (usually pulse width modulation) the output switching waveform in some fashion such that IF you applied a low pass filter between the outputs and the speakers, the combination of low pass filter and speaker inductance/capacitance would convert the switching waveform to analog audio you can hear.
Class D amp:
This primarily means an audio amp where the output stage is operating in Class D (switching mode). Some people refer to this as a digital amp. But it is really a digital (switching) output stage. The input can be analog audio, with the amplifier operating a particular A-D modulation scheme which results in the switching output or the input could also be digital as already described, depending upon the particular amplifier design. Output filtering is general required except in relatively low power cases (generally under 5W, but new designs are pushing the envelope to ~15W) and where the amplifier is physically very close to the speaker (few inches)
Of course there are amplifiers out there that take in digitized audio and do an internal D-A operation and operate the output stage in one of several continuous time analog Classes (A, AB, B) which do not require filterting on the output before the speakers.
The amplifier you need would depend upon what your application is. As I recall, your application was relatively low power, and so the amplifier would be chosen on the basis of what form your audio into will be. If I know that, I might be able to recommend a single monolithic "digital" amplifier for you. If you have a high power need, then that would be a different answer.
Let me know (a) the format of the input audio, (b) the power level of the output, and (3) something about the final application. Again, if I remember correctly, the application was related to hearing aids, but I don't know if it will finally be in-ear type, or some other form. If in-ear, the noise performance of whatever amplifier is chosen will be very important, along with the power consumption.
Ken: This is once again your ancient engineer Al Mond, who celebrated big nine-0 birthday this week, asking your advice. Instead of trying to create and build my idea of a perfect complementary audio hearing aid, let me put the question of you. Do you suppose it is possible to build an audio amplifier which deletes all vowels and amplifies only the consonants? The technical difference is that the vowels are all sinusoidal and the consonants are literally noise spikes. There is a huge theory that goes with this and I am hoping to publish this in the New England Journal of Medicine. But the bottom line is that hard of hearing people typically don't hear the consonants. For one simple but important reason, the consonants are thousands of times quieter than the vowels, and considerably shorter in duration. Give me your thoughts of the possibility of a consonants only audio amplifier.
The HCPL-5530 is commercial grade two channel hermetically sealed optocoupler in a 8-Pin ceramic DIP package with gold plated leads. Solder dipped leads and various lead form options are also available. See datasheet for details.
The product is capable of operation and storage over the full military temp range and may also be purchased with MIL-PRF-38534 Class H or K level testing or from the DSCC SMD 5962-87679. The HCPL-5530 is manufactured and tested on a MIL-PRF-38534 certified line and Class H and K versions are included in the DSCC Qualified Manufacturers List QML-38534 for Hybrid Microcircuits.
Each channel contains a GaAsP LED which is optically coupled to an integrated photon detector. Separate connections for the photodiodes and output transistor collectors improve the speed up to a hundred times that of a conventional phototransistor coupler by reducing the base-collector capacitance. http://www.seekic.com
I hope your Thanksgiving was all it should be....with family and friends. Also, CONGRATULATIONS on your 90th. I hope I am still as inventive and interested as you are when I am 90.
Nate....your question regarding an amplifier that only amplifies consonants and rejects vowels is way out of my league. My natural response would be to say that you would need very sophisticated signal identification and filtering to accomplish the task. I personally don't know any techniques in analog signal processing that could do that job economically or with a high degree of programmability and repeatability (stability).
I think you would have to resort to a digital signal processor programmed to recognize the audio type.
The minimum HW in the DSP scenario is an A-D converter, the DSP, and then a D-A. The processing requirements would dictate the power consumption overall. Because it is mono voice being processed, the sample rates can be kept low, but to recognize the sound type on the fly will still need many calculations on each sample and its relationship to previous samples.
I am sorry I am not much additional help to you at this point. I know a couple of good DSP algorithm people, and next week I will pose the question to them and see what I get back.