Best way to reduce noise when reducing gain, triode pre-amp stage?

[mooreamps]

Sorry for being such a troglodite, but *why* Led biasing?
I see no special advantage, and it has certainly created some social trouble in the last days.

Well, if led biasing was popular in tube amps, I would think Kevin would have captured that in one of his TUT's ; but he did not...
As for the equipment that I design and build ; it produces a certain character of chime to the tone ; less dull sounding than using resistor bias...
Not trying to say it's better or worse ; just a new type of tone which is more natural sounding with respect to the instrument..
PS. Bypass cap's not required... :|

-g

[jimboyogi]

OK, thanks again, I think it is extremely cool that there are so many different ideas and solutions to a fairly simple circuit. No wonder we all love to play with this stuff!

J M Fahey, maybe led's could help light up that cave! Seriously I like led biasing, it's neat, simple, fully bypassed, and I do like the tone, as you pointed out mooreamps it has a certain sound of it's own. However this tone is not so critical on a reverb recovery stage. From reading MerlinB's pre-amp book, where he points out the importance of fully bypassing the first stage for hum rejection, I used it here to try and minimise hum, especially because the input signal is so small (~ 20mV at the grid). Also because I have been playing around with different Rp's and even different tubes, led biasing keeps the bias voltage constant despite all of these experiments.

Also JM Fahey, why do you refer to the 4n7 cap being wasted on grid to ground, versus being plate to ground? Can you please explain why plate to ground is a superior position?

I am pretty happy with using a small grid leak for the input, due to the inductance/treble attenuating effect that I discussed above. Signal loss at 1KHz for the 22K should be ~ 10.6%, at 2KHz ~ 19%, at 3KHz ~ 26%, etc. For a 47K it should be at 1KHz ~ 5%, at 2KHz ~ 10%, at 3KHz ~ 14% etc.
I have put in a small (10K) grid stopper, I like to keep something here just for RF stability, especially as the tank and lead are running exterior to the chassis.

Enzo I tried covering the led to block all incident light, listening carefully I couldn't hear any change.

I did have a very long shielded lead between the chassis and the reverb tank, I shortened it to ~ 0.5m and there is a small decrease in the noise. All in all I am very happy with the sound now, and I am also happy to have replaced the 12DW7, a fairly obscure tube, with a 12AU7, which are far more widespread and obtainable.

[Enzo]

Moore, do you ever even think about what people are saying?

I never suggested taping the LED permanently, nor did I mention playing on a gig.

The man is building an amp, and has a noise issue in the amp. Not playing it on stage, he is BUILDING it. Presumably he is not BUILDING the amp while playing it at a gig. Presumably he is building it on a workbench. Further, if his bench is like most, it has a lot of lighting right directly above the work. It turns out that it doesn;t contribute any noise. Fine. And even if it did, and in the manner I suggested, I have no doubt that once assembled and on said gig, it probably wouldn;t be doing it anyway. The point was to solve the noise issue he is having NOW, while BUILDING the amp. If the amp were to stop making the noise when boxed up, great, but while open on his bench, he needs to know if the noise is continuous or if it is contingent upon some environmental factor.

DO you understand that distinction, gary? Is that over your head? I know you are invested in throwing LED biasing into every conversation you can, more power to you. A suggestion that an LED could be involved in a problem is not an attack on your fragile integrity. And I hope you don;t find out the hard way that those seemingly opaque black plastic transistors we use all over are actually relatively transparent to infrared, and if you are working on a sensitively balanced circuit it can indeed throw the operation off.

[mooreamps]

I can tell you a lamp shining on a led does not produce any more noise than what is already there...

-g

[J M Fahey]

why do you refer to the 4n7 cap being wasted on grid to ground, versus being plate to ground? Can you please explain why plate to ground is a superior position?

Sure.
1) Hiss attenuation:
Grid to Ground: it will attenuate hiss coming from an earlier stage.
Problem is, there is none such stage. The reverb tank is a passive transducer.
Only possible "good" effect: it will lower the generator impedance seen by that grid, lowering a little *one* of the noise factors (also killing all signal highs).
Any noise produced in the tube itself, such as electrons hitting the plate at random speeds and angles will not be affected, yet you are already killing signal highs for sure.
Sacrificing something for little or no return is one definition of waste.
Plate to ground: it will attenuate *all* hiss , both that produced at the grid *and* that produced at the plate.
For the same amount of "medicine", you are "curing" more.
Yes, you are killing highs as before, but for a good cause.
2) Reverb tank "brightening" : if you were using the tank directly feeding the cap, the resonant peak could be used to further extend the tank's frequency response, it's an old classic trick used by many manufacturers for ages.
Using a high value "grid stopper" between tank and capacitor negates that, killing the "Q".
3) *You* are having a hiss problem, in a classic circuit not noted for it.
Forgive us for having our attention drawn to the unusual element there, it's the first suspect.
This is called "good troubleshooting practice"
Besides, *all* junctions are noise generators when passing current (as here).
Some examples :
Zener diodes have been used for ages as white noise generators; so noisy as to be used in analog drum machines; regular diodes (meaning *all* diodes) are the classic noise source to calibrate sensitive receivers; the spectrum extending well into UHF frequencies.
.
Dear Gary, may I suggest you reading a little more before answering?
Or at least using that wonderful "search" function found in Google.
LEDs *are* photodiodes too, generating voltage and current when hit by light.
Searching "led as photo diode" offers 666000 results, including:
Sensor Workshop at ITP | Reoprts / LED As Photodiode
30 Apr 2006 ... Allow the LED to operate in the solar cell mode, quadrant #4 (operating voltage is positive, current is negative), or in the photodiode mode ...
itp.nyu.edu/physcomp/sensors/Reoprts/LEDAsPhotodiode
.
LED circuit - Wikipedia, the free encyclopedia
An LED can be used as a photodiode used for light detection as well as emission. This capability has been demonstrated and used in a variety of applications ...
en.wikipedia.org/wiki/LED_circuit - Cached - Similar
.
LED as photodiode
13 Dec 2005 ... It gives you a few clues about what to look for: "The LED is a photodiode that is sensitive to light at and above the wavelength at which it ...
Microchip Technology Inc. is a Leading Provider of Microcontroller and Analog Semiconductors, providing low-risk product development, lower total system cost and faster time to market for thousands of diverse customer applications worldwide. › ... › [Microcontroller Discussion Group] - etc. etc.

mooreamps
I can tell you a lamp shining on a led does not produce any more noise than what is already there...
-g

Turning a bulb lamp on a Led will only add a little hum, easily masked by already present hum, but hitting it with the funky light pulses produced by a fluorescent lamp is something else.
Oh well.
*This* also fits my definition of wasted time.
Oh well [2]

[mooreamps]

Waste of my time as well.. I find no data sheets on any led's showing me any spec's on how to design a circuit using a Led as a photo-detector... -g

[jimboyogi]

Gentlemen Please

Unfortunately this thread, as for so many others, has degenerated into antagonism instead of respectful discussion.

mooreamps, you made some interesting points early on in the thread, and for that I thank you. If you could stick purely to that kind of respectful discussion then this forum would function much better for all.

JM Fahey, thanks for the explanation. Please remember that I meant no offence with my questions, and the LED in the cave joke was just that.

Respect.

James

[Enzo]

With respect, LED data sheets won't generally discuss using them as sensors for several reasons. One, that is not their INTENDED purpose, the same companies usually make semicondcutor sensors as well as LEDs. Two, though they will function as sensors, if it goes on the data sheet, then the LED maker has to guarantee that sensor performance. SInce the LED was not designed with this application in mind, they are not about to characterize their part in that way.

In a similar vein, transistors are not usually characterized on data sheets wired as diodes. But engineers use them that way all over electronics. COnnor or RG or someone could explain it better than I, but as I understand it, a diode wired transistor is faster than a regular diode.

There are many non-standard applications for electronic components. They won;t be on the data sheets because the part maker is not interested in guaranteeing those particular specs. And that applies elsewhere. Various people have used woofers as microphones for kick drums. You won't find ElectroVoice, or EMinence or any of the others putting microphone response patterns on their speaker data sheets, but the speaker will work as an acoustic sensor - microphone - and people do use them that way for recording and live sound.

[mooreamps]

Turning a bulb lamp on a Led will only add a little hum, easily masked by already present hum, but hitting it with the funky light pulses produced by a fluorescent lamp is something else.

:|


Test setup :

1. Fluke 179 multimeter. Input impedance = 10 meg ohm
2. one standard led
3. one 100 k ohm resistor.

A. Measurement one. Multimeter on A/C setting. Test leads open. Measurement varies from 100 mv to 225 mv.
B. Measurement two. Multimeter on A/C setting. Test leads across led covered from light. Measurement approx 10 mv.
C. Measurement three. Multimeter on A/c setting. Test leads across led held directly under lamp. Measurement approx 10 mv.
D. Measurement four. Multimeter on A/C setting with 100 K resistor in shunt with the led. Measurement approx 10 mv.

Oh well.
*This* also fits my definition of wasted time.

If it is a waste of your time, then what do you care about it ?

-g

[raiken]

A video is worth a thousand words:

http://www.aikenamps.com/led_photosensitivity.avi

Test setup: Fluke 87V, red LED, one curly fluorescent bulb, one black plastic anti-stat bag to block the light.

Every engineer knows about the photovoltaic effect of semiconductors...if not, they learn it the hard way.

I ran into a particularly nasty variety of it while doing a contract engineering job for Microtime in CT back in 1991, designing their FS-10 PAL video frame synchronizer (10-bit digital video was a very big deal back in 1991!). The output video had noticeable chroma noise on the vectorscope display, and I spent days trying to track it down. They had me put up in a plush video editing suite area that had dimmable lights. While pondering the problem, I turned the lights up brighter and noticed the noise problem got better. It turned out the Atmel UV-erasable PLD I was using (I think it was the ATV2500) had a problem with noise that would get better when light was shining into the erase window.

If it had only worked better in the dark, I would have been okay, because I could have stuck an opaque label on it, but it got better with light shining on it. so I was out of luck. Finally, after several weeks, Atmel released a new version of the part that didn't have the problem...

RA

[R.G.]

:|
Test setup :
1. Fluke 179 multimeter. Input impedance = 10 meg ohm
2. one standard led
3. one 100 k ohm resistor.
A. Measurement one. Multimeter on A/C setting. Test leads open. Measurement varies from 100 mv to 225 mv.
B. Measurement two. Multimeter on A/C setting. Test leads across led covered from light. Measurement approx 10 mv.
C. Measurement three. Multimeter on A/c setting. Test leads across led held directly under lamp. Measurement approx 10 mv.
D. Measurement four. Multimeter on A/C setting with 100 K resistor in shunt with the led. Measurement approx 10 mv.

LED photovoltaic effect

"...using green LEDs as photovoltaic light sensors"

Shine light on an LED, and it will generate a miniscule amount of electrical current!

Light emitting diode fault detection using p-n junction photovoltaic effect.

Maybe you need a different experimental setup.

[mooreamps]

Mr. Raiken. Difficult to see the display, but it shows a voltage across a 10 meg ohm impedance. What is the voltage produced across a 100 K ohm impedance ?

Mr. R.G. That's fine. I just don't care what they do under a lamp. They otherwise suit my purposes as expected..

-respectfully;
-gary

[raiken]

Mr. Raiken. Difficult to see the display, but it shows a voltage across a 10 meg ohm impedance. What is the voltage produced across a 100 K ohm impedance ?

-respectfully;
-gary

If I hold the light close to the LED, I can get up to 140mV into 100K, at the distance shown in the video, it is around 22mV into 100K. Normal ambient light far away, only a couple mV into 100K.

Amazingly enough, into the high impedance of the meter, I can get up to a whopping 1.2V with the light right on the LED, and it drops to zero when the black plastic is put between the light and the LED...you could make a nice ambient light detector with an LED.


RA

[R.G.]

Mr. R.G. That's fine. I just don't care what they do under a lamp. They otherwise suit my purposes as expected..

Well, that's OK then. If you like them, and they work for you, good.

If you use them in audio circuits, you might want to paint them or pot them so that ambient light never can get in to them. Otherwise, one day, it might introduce odd and unexpected noises.

[Steve Conner]

Gee, I really must try this LED biasing that everyone is talking about in every single thread on this forum. Not.

I've seen a cool trick where you connect a green LED to a red LED, and then shine a green laser on the green LED. It acts as a solar cell and makes enough current to light the red LED very brightly.

Also, the BEAM Bot community like to use LEDs for their bugs' "eyes".

Having said that, people have used LED biasing in solid-state power amps for years, with no noise issues. The LED is bypassed by a capacitor, and the biased stage is inside a feedback loop.

Randall: I'd bet you (or Atmel...) had a race condition or metastability issue in that design, and the extra leakage current from ambient light was skewing the thresholds of your gates enough to resolve it.