What you say is certainly true of the output's ability to sink current (negative half-cycles). But, as far as I can see, there is nothing to limit the current the circuit can source: the output current on positive half-cycles doesn't appear to be limited by anything other than than the output transistor's saturation resistance, and maybe available gate drive voltage in the case of the FET output version of the circuit.
That's what I do with my home hi-fi amp and loudspeakers. But that setup is not capable of generating enough SPL to instantly and permanently damage my hearing in case there happens to be an accidental large transient.
Large transients like that - as well as wildly unpredictable audio levels - have become increasingly common these days, if any part of your listening involves streaming digital audio, such as from Youtube.
There is also the fact that headphones, by their nature, are more likely to suffer from problems like bad contacts. If you hear nothing, turn up the volume in response, and then try pushing the headphone plug to make sure it's fully seated, for example, you can be exposed to a full-volume blast for a split second. It's a mistake, but an easy mistake to make.
I try to make it a habit to actually pull my headphones off my ears, leaving them sitting loosely on my temples, while any initial volume adjustments are being made; I only move them back over my ears once I'm sure I can barely hear the sound.
Even so, I have been hit with sudden unpredictable loud transients, for example, when listening to a poorly recorded Youtube clip, or when Youtube's autoplay feature started a second clip recorded at a much higher level than the one I'd been listening to.
Worldwide statistics since Sony invented the Walkman back in the 1980s show that headphones are dangerous to the health of your hearing - I think it's worth taking what extra precautions we can.
-Gnobuddy
112 dB transients are perfectly acceptable for "loud" listening. This leaves plenty of dynamic range for listening at 90 dB,which seems to be a level where a lot of musical programs sound full and loud without being deafening.
A lot of musical programs don't have anywhere near that dynamic range. There are quite a few that do, though.
Re your last statement there has been a proliferation in the use of portable devices such as cellphones and tablets, and of matching high sensitivity headphone sets.
Looking at a range of product in the Jaycar electronics chain store range operating in Australia and New Zealand headphones range in maximum power ratings from 10 mW to 100 mW with sensitivities between 100 and 114 dB.
All are fitted with a 3.5 mm plug and supplied with an adaptor to mate with 6.5 mm sockets on many stereo amplifiers. The use of an adaptor adds an extra connection and makes the plug more at risk of physical damage.
I no longer listen through headphones however I have owned sets of Sennheiser headphones - both high impedance types with 6.5 mm plugs and similar to models in the company's current "Audiophile" range. The better ones now cover 6.5mm and 3.5 mm plugs but with some exceptions these are high impedance types.
My preference is for quality and I would be prepared to pay for this rather than suffer a cheaper product which is loud.
With regard to your observation about sinking and sourcing current the emitter of Q4 will accept current from the constant current source connected to the negative supply rail and via the 32 ohm impedance of the headphone from earth. The amount of current drawn through the 32 ohms depends on the magnitude of the voltage swings in the output stage. At a 100 m.W. peak output that is not going to be very much and I think the headphones would be more at risk of failure than the output stage.
I had envisaged some current limiting with pairs of voltage regulators for each channel. 100 ma types 78L15 and 79L15 have a peak current of 140 m.a. Changing the value of the 5 watt resistor in the constant current source from 6R8 to 8R2 would reduce the dissipation in the output stage and the regulators. Additionally a 100R resistor between Q3 collector and Q4 base would limit the current under momentary fault conditions and improve any overload recovery times. There should also be a zobel network at each amplifier output.
For youtube material you might consider a separate input with divider resistors to attenuate the level to match other inputs and a mute switch to increase the attenuation between video sound clips. Some visual means such as a panel meter or display might be a useful prelude if the take off point was ahead of the muting switch.
I think gorge make have gotten of the bus with the discussion branching into the area of headphones - this being incidental to his intention to learn more about the internal workings of op.amps. There is a range of choice to consider in the Passdiy article.
...and, should a sufficiently large input signal be present, positive voltage swings can rise to within two or three volts of the +15V rail. A 12 volt peak swing into a 32 ohm load is 4.5 watts peak, 2.25 watts RMS (RMS being calculated over a single half-cycle in this case). That was my concern.
The various streaming audio businesses must be using some sort of level-normalizing software on their music servers to avoid this problem. With the ability to look ahead in the digital audio file to see what loudness is about to show up, and some knowledge of psychoacoustics and perceived loudness, this seems feasible.
I wonder if there is any open source software out there that can be inserted in the audio playback chain of the computer to do this job.
Gorge wasn't exactly treated with the greatest of courtesy when he started this thread, and there were recurring bursts of hostility aimed at him after that. I'm guessing he had enough, and went elsewhere.
A pity. We collectively lost the chance to help someone who was interested in audio electronics.
-Gnobuddy
Hi Guys
Have any of you actually MEASURED a 90dB sound field? It is intolerably loud.
The OSHA recommendations for exposure to specific sound levels is a total joke that is economically driven and has nothing to do with safety. Their listings are a way to allow American industry to continue as it has, with only the mildest attention paid to hearing protection.
I've measured SPLs in my listening space and I cannot stand to be in the room with true 90dB-C (flat) levels. That is 85dB-A, which is the standard and legally-mandated way to measure it.
A-weighting measure "ear loudness" and ignores three-quarters of the actual sound energy present.
C-weighting is a flat response and tells you the total sound energy present and the physiological impact on you.
The difference is only 5dB between A-weighting and C-weighting but the sound difference is almost four times difference. The logarithmic dB scale hides this large difference.
I designed many head phone drivers and in the ones that realistically only go to about 100dB (for ESLs), no one has complained of a lack of headroom or of clipping. The fascination with big macho numbers takes a hit once real measurements are made.
YOU are your own hearing protection board. Turn it down and enjoy music for a lifetime.
Have fun
Have any of you actually MEASURED a 90dB sound field? It is intolerably loud.
The OSHA recommendations for exposure to specific sound levels is a total joke that is economically driven and has nothing to do with safety. Their listings are a way to allow American industry to continue as it has, with only the mildest attention paid to hearing protection.
I've measured SPLs in my listening space and I cannot stand to be in the room with true 90dB-C (flat) levels. That is 85dB-A, which is the standard and legally-mandated way to measure it.
A-weighting measure "ear loudness" and ignores three-quarters of the actual sound energy present.
C-weighting is a flat response and tells you the total sound energy present and the physiological impact on you.
The difference is only 5dB between A-weighting and C-weighting but the sound difference is almost four times difference. The logarithmic dB scale hides this large difference.
I designed many head phone drivers and in the ones that realistically only go to about 100dB (for ESLs), no one has complained of a lack of headroom or of clipping. The fascination with big macho numbers takes a hit once real measurements are made.
YOU are your own hearing protection board. Turn it down and enjoy music for a lifetime.
Have fun
Over the weekend I am expecting to meet with two brothers who work in IT who keep up with software developments. One of them does work in writing computer programs.
I have a three page article in Electronics World and Wireless World in November 1995 by John Linsley-Hood describing an audio processor capable of expanding and compressing by around 26 dB - with very low THD. This is a 130 kHz operating speed chopper attenuator using a J111 FET a CD4069 and a few TL052 and LM833 ICs. The TL052s are used for 2 third order low pass filters to obviate the switching artifacts.
Possibly the offending noise at the start of a new video clip is due to high levels both of compression and volume. If most material is compressed a little expansion might allow a lower volume setting where the clip intro drama is not so bad as to be tolerable.
A friend of mine once observed the irony of a situation where his teenage son could spend time listening to music at an intolerable level indoors, then go outside with ear protection to use the lawn mower.
Wearing ear protection in the workplace is apparently taken a little more seriously than to the levels of sound of PA type systems at a sports or concert stadium. These other money making ventures are magnet for young people to dress up to attend for the party atmosphere enhanced by consumption of alcohol.
I went to a sports event 3 years ago the sound system was overbearingly loud and the young people paid no respect to older sports patrons in the audience. I believe these young people have been conditioned to hearing loud music - this infringement of personal liberty was a one-off experience for me.
Do musicians and producers try to emulate this in recording studios. If so this is exploiting "bullet- proof" attitudes which is a stage of life for a good many people.
Last edited:
I'm curious - was the chopper switching the FETs channel resistance between "on" and "off", using PWM to establish an average value? And does this produce lower distortion than using the same JFET as a voltage variable resistor?
I'm doing some initial thinking about a project that would require a microcontroller to vary the value of a resistance, and using PWM plus a CMOS bilateral switch (4066, etc) on a fixed resistor was one idea that crossed my mind. A digipot is another, but that idea comes with headaches of it's own, too.
-Gnobuddy
Switching duration is through a control voltage derived from the audio signal.
The audio "conditioning" circuit block - an inverting op.amp sums audio of both channels - output is via a series capacitor into a half wave rectifier format with a small capacitor load. The positive voltage end of the capacitor load connects to a buffer amplifier outputting to CD4069 Hex inverter.
The square wave from this circuit block feeds J111 gate. Going back a step, divider resistors at the input of the buffer op.amp allow a dc voltage to be applied the junction point for the expansion function.
Within the range 100Hz to 10kHz, the chopper section is claimed to have a THD of 0.005 per cent for a 1 volt rms output.
There was an earlier article about the latter in the April issue of the same journal. From memory the author discussed the principles adopted for I.C.volume controls then on the market. One of the possible benefits with this approach was better matching of channels than in affordable volume pots of the day.
Last edited:
I think the difficulty in finding good stereo pots, and even log pots, has actually become much worse in recent years.
There are two electronics stores within an hour's driving distance of me, and neither one stocks even a single stereo (ganged) potentiometer. Both carry a small selection of linear pots - and, apparently, no log pots whatsoever.
We all know there are fewer electronic hobbyists today than there used to be in the past. Not only that, the few remaining hobbyists are almost certainly not interested in audio electronics - so they have little use for log pots, or stereo pots.
-Gnobuddy
There are two electronics stores within an hour's driving distance of me, and neither one stocks even a single stereo (ganged) potentiometer. Both carry a small selection of linear pots - and, apparently, no log pots whatsoever.
We all know there are fewer electronic hobbyists today than there used to be in the past. Not only that, the few remaining hobbyists are almost certainly not interested in audio electronics - so they have little use for log pots, or stereo pots.
-Gnobuddy
Last year I bought some dual log pots from futurlec. These cost a mere $0.80 US apiece but the tracking error in the one sample I measured is nearly 10 % at half rotation. I use a balance control so I can match the sound levels with this - albeit necessary when the volume setting is altered.
Reputedly linear pots have better tracking - these might be better with a law faking resistor between the wiper of each gang to ground. The Baxandall active volume control with the pot in the feedback loop of an op.amp is another option.
For the sort of circuit you are thinking about there was a noise blanker circuit by Linsley-Hood published in Wireless World in January and February 1983 that may be of interest. These are available on line the link to the latter of the two is here http://www.americanradiohistory.com/Archive-Wireless-World/80s/Wireless-World-1983-02.pdf. The circuit is a noise blanker for vinyl input however adding a capacitor allowed a delay of the signal preceding the blip to pass in the blank space.
There is sound processor software like Audacity which is available free. You would have to read the legal conditions for using this.
There are other online sources to which one can subscribe to where the material has already been processed and is available on demand.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.