Yes, I know you are not abusing them---I just wanted to bring up the uncertainties for reverse breakdown voltage, if someone was intending to use that as an intended mode of operation.
Also a mention: in bipolars, reverse breakdown of the base-emitter junction is more of a no-no as it degrades the noise performance. Whether a comparable degradation mechanism exists for LEDs I don't know.
Help help, we are in the matrix. Obvious and yet not obvious. The old 16 frames cinema wasn't fast enough. US people say they could see 50 Hz TV. Can't say I saw it strobe. I often though the great inventer must have thought of TV when letting us loose on the planet. It would be so very diffficult if we saw the frames change.
I hold fast to what I said before. 90% of music has no phase preference. A transient should be postive. My instinct is to say your phase detector will be correct regardless of how it was recorded. There will be special cases when wrong is right.
I have searched and searched for that and never found the transistor data. As someone said about LED's. Surely if the current can not pass the zener + region it is protected. One thing you can try is fit a limiting resistor and measure current if reverses biased. A LED at 24 V reversed showed zero mA. I doubt it was a micro amp when using 10K. Two weeks later still working. If set up for 30 mA I suspect it will fail quickly. I don't know about LED's. I suspect they zener. I am told they can work as LRD ( receiving ) if reversed.
Motchenbacher and Fitchen, Low Noise Electronic Design, or the more recent Motchenbacher and Connelly, Low Noise Electronic System Design, have discussions. The latter reference describes the degradation of beta and increase of noise pp. 133-136, and the effect of annealing using a high forward base current.
There are also several references in that M & C chapter about the effect.
LEDs are indeed photosensitive, particularly red ones. They act like photodiodes, not light-dependent resistors. One guy I know keeps objecting to their use as bias sources for fear that light will cause excess noise and hum if the light source is modulated by the drive current---but it's easy to shield them from such, and the effects are rather small compared to the normally substantial forward currents. Since the red GaAsP parts have a temperature coefficient fairly close to a silicon base-emitter junction, but a larger voltage, one can make o.k. current sinks and sources with BJTs and an emitter resistor. You have about 900mV to play with across the resistor.
You've been in the trade for a looong time, so you know how equipment in the flesh sounds - you have good benchmarks for what you're after in this regard. I'm coming from a different angle, which is that I don't want to hear audible artifacts damaging the low level information which is critical to allowing a good quality illusion to form - that level of reproduction which allows the acoustic captured in the recording to dominate the listening space is of primary importance to me. Whether a particular instrument is acoustically polarity correct doesn't bother me, so long as it occupies its space convincingly, and is fully revealed and distinct within the playback arena.
I get a special pleasure from making "impossible" recordings "work" - in particular, a $1 Gene Pitney pirate CD which has been atrociously butchered by some idiot for the mastering - noise reduction was applied in the most grotesque, brain dead fashion ... on a conventional hifi this sounds totally hideous, torture of the highest order, 😀. Yet, on a finely tuned setup all this diabolical mess fades into the background, doesn't register strongly unless one deliberately tries to listen specifically to what was so badly done ...
LEDs in my case are for indication of something only. NEVER a part of any meaningful ciruit. I am absolutely disgusted with the current crop of LEDs.
I remember when they cropped up on the mass market at the beginning of the 70ies. The average expected kife of a LED was - hold on, now - 80 years. I don't know how true that was, but I do know that I have never seen a dead LED beofre about 2000. Now, I see them dropping like flies, ever since the Chinese muscled into that market. Now it's routine to have LED die on you ever two weeks or so.
So, my reasoning is - if they could mess them up that much, how can I believe their other properties, such as low noise?
In current sources, I believe transistors will work better if their base is driven by higher voltages, such as say zeners rated at 5.6 or 6.2 V.
I remember when they cropped up on the mass market at the beginning of the 70ies. The average expected kife of a LED was - hold on, now - 80 years. I don't know how true that was, but I do know that I have never seen a dead LED beofre about 2000. Now, I see them dropping like flies, ever since the Chinese muscled into that market. Now it's routine to have LED die on you ever two weeks or so.
So, my reasoning is - if they could mess them up that much, how can I believe their other properties, such as low noise?
In current sources, I believe transistors will work better if their base is driven by higher voltages, such as say zeners rated at 5.6 or 6.2 V.
I remember reading that testing transitors is a way of degrading their noise performance. I forget the details. These days I just install them and hope for the best if the input stage. This is a bit daft as if BC560C someone has tasted them. Perhaps there is a safe test ? I guess a gain test as long as the polarity of wires is correct is OK. On something I built recently I will sometimes reverse bias the base of a BC327-40 by 5 volts. It is a switching situation. I am using a LM324 as it's driver. If I had used LM339 comparator the problem would not have come about. Usually I used LM324 as it can both source and sink. If LM339 it can only sink. Thinking about it a mildy bad choice. The LM324 limits the current to about 20 mA should the base zener/avalanche. The BC327 is a tough device. The intersting thing is hard reverse biasing make this transitor shunt behave as if a very high value resistor. Whilst this is not critical it is preferable. It is a crowbar protection circuit. Without the reverse bias the crowbar changes the DC conditions akin to DC offset.
I would love someone to say why LED's make good voltage references. I use them and like the light saying things are working. I never really found them better in any way. I might have seen very low noise from them. This was nullified when a simple filter cap of 10 uF was used to make all versions quiet.
The high power LED's are a different animal. I suspect the LED costing 5 cents at 1 mA will last 10 000 years or at least 100. Kingbright seem OK. Opto couplers are wonderful. Not only do they make is safe, they also level shift.
I would love someone to say why LED's make good voltage references. I use them and like the light saying things are working. I never really found them better in any way. I might have seen very low noise from them. This was nullified when a simple filter cap of 10 uF was used to make all versions quiet.
The high power LED's are a different animal. I suspect the LED costing 5 cents at 1 mA will last 10 000 years or at least 100. Kingbright seem OK. Opto couplers are wonderful. Not only do they make is safe, they also level shift.
We see almost nothing but reflected light.
Everything you see is because you see the light reflected from it.
Unless it's the sun, a light bulb or on fire.
Exactly so. It is the frequency that the object won't absorb. The idea I wanted to put across was phase distortion in light as in defraction. To simplifiy I am sure the eye is allowed more bandwith. Someone made a good point ( I think it might have been me? ). Light stimulus has always been much as now if we just say how much light and how many objects. Sound has changed. Maybe we are not evolved enough for modern sounds ? If we were hi fi would be a bit of silly nonsense.
I helped design some small speakers of a very standard type. I have been given a pair as a thank you. I have not heard them for some months and have to say my doubts are far less now. On first hearing they are much like my very big speakers. Due to box effects they seem almost better, that is bass when perhaps a harmonic. On analytical grounds they are 10 ot of 10 verses price. The problem is I have to listen harder than usual. I am very happy both ways. Now it will be better visiting Colleen. She loves music even more than me and we met this way. She has been hoping this day would come and I get the speakers. Much as she likes the Dynaco speaker these will sound better and look better. If I have a doubt about the small ones is that my brain has to work harder.
It is very interesting how condutors often have a silver colour. I read it up once. It was very obvious stuff.
I helped design some small speakers of a very standard type. I have been given a pair as a thank you. I have not heard them for some months and have to say my doubts are far less now. On first hearing they are much like my very big speakers. Due to box effects they seem almost better, that is bass when perhaps a harmonic. On analytical grounds they are 10 ot of 10 verses price. The problem is I have to listen harder than usual. I am very happy both ways. Now it will be better visiting Colleen. She loves music even more than me and we met this way. She has been hoping this day would come and I get the speakers. Much as she likes the Dynaco speaker these will sound better and look better. If I have a doubt about the small ones is that my brain has to work harder.
It is very interesting how condutors often have a silver colour. I read it up once. It was very obvious stuff.
Yes, more voltage is always better (larger emitter R with lower current noise and with e sub n of the bipolar a smaller effect) unless the zener (say) is itself noisy. And the near-5V parts are usually thus.
As far as reliability of current crop parts, it may be a poor process used to attach the bond wire. But the Avago (formerly HP) LEDs are quite reliable and available---the HLMP-6000s are very consistent. But the temperature profile is crucial, as the encapsulating plastic will soften.
Bs. Most of the recordings out there have one or more mics ( or direct inputs) phase inverted (electronically or acoustically) while others are not. Switching phase at reproduction will put some in phase and others out of phase. The engineer / producer has decides what's right, go with that not with your ego.
Completely disagreed on zener noise - my experience is totally the opposite. I find that zeners up to 6.2V rating are the most quiet of the lot, trouble starts above that, when they go into avalnche mode. Besides, I can't remember ever having used a zener diode without proper capacitor filtering. A long time ago, I would also add a singnal diode, like 1N4148, to thermally compensate the zener, but these days it's no longer really necessary unless somebody slipped me some dud zener diodes. They are much better made these days, and despite being sold with the usual +/-5% tolerance, I find many sit on the nominal voltage actually +/- 1 %. Also, their zener noise levels have been reduced and then some.
Frankly, I'd say these days the low noise chase is mostly related to the transistor used. In my book, the old BC 109 C metal can is still the most quiet transitir I know of, even if it's still limited in voltage.
Actually, there was one better still, Motorola's BC 651, but it is now out of production, like so many truly excellent parts of yore.
I agree. There is a limit to the anticipation torture I am willing to put myself to for better playback, so if anyone is worried about it, they should build in a phase inverting switch and hop to it every time they change the LP.
In my experience the worst zeners for noise are 4.7V or 5.1V, right at the transition between zener and avalanche, which unfortunately is also at about the zero tempco operating point. There was an old Moto databook which showed curves of typical noise and confirmed this, at least for those days.
A lot of the zener noise is associated with surface states, hence the practice of "burying" the zener, but I think this is not common for single isolated parts. It was used in the National temperature-stabilized part, the LM199/299/399 if memory serves.
The tempco of zeners follows the voltage, negative for low-voltage devices showing the true zener action, positive for avalanching. This has not changed since their invention afaik. It is not a matter of making the parts "better". There are of course the old temperature-compensated ones with the junction diode built in, although they suffer from higher impedance.
As far as low noise transistors, the now-scarce low rbb' Toshiba parts are the best I've found (for bipolars), the 2SA1316 and 2SC3329. The highest beta varieties are of course preferred to have lowest current noise. You have to run enough current for the referred collector shot noise to be comparable or less than the thermal noise from the amazing 2 ohm rbb'. If the external impedances are too high the base current noise will spoil the noise performance.
But for any voltage reference part commercially available, transistors should not be the limitation. I've had to resort to more complex circuits, typically using JFETs biased at the zero tempco point, to get things down to really impressive levels.
Another approach although cumbersome: ensemble averaging of multiple references. I actually tested this recently with a series string of the bog-standard TL431, each connected with cathode to reference for about 2.5V, then stacked. The square root of n dependence for percentage of noise versus d.c. voltage was followed, once the additional improvement of no reference terminal current noise (no voltage divider) was accounted for.
Time permitting I may start another thread about references, although there are a few floating around in here, including experimental results comparing zeners and LEDs. The problem with LEDs: they lack a sharp knee. They appeal however because of the matching tempcos to bipolar Vbe's.
Brad
Here is a dubious datasheet that suggests an impossibly abrupt transition occurs for noise versus zener voltage: http://www.taitroncomponents.com/catalog/Datasheet/1N4614.pdf
If it were true, in terms of "signal-to-noise ratio" the 100V part would almost be as good as the 4.7V part.
Might be worth getting hold of some of these parts around that transition.
If it were true, in terms of "signal-to-noise ratio" the 100V part would almost be as good as the 4.7V part.
Might be worth getting hold of some of these parts around that transition.
A couple of years ago, when I heard this objection, I used a strobe light to really bang on the LEDs in a cathode bias arrangement. Nothing, nada, zip, bupkis effect. I think it's posted somewhere here on the forum. Another urban legend.
I remember when flat earth manufacturers wouldn't put an 'on' LED on a product for that reason. Thank <insert deity> they are no longer in business.
I've used red LEDs as low-budget phototdiodes, but they aren't very sensitive. Blues are practically solar-blind, although much of that may be due to the plastic limited UV transmission.
But yeah---look at the nanoamp photocurrents and compare to the few-mA or more operating currents when used as bias sources. Just not much of an effect possible.
- Status
- Not open for further replies.