Help: White - Pink Noise Generator

[Tedyp]

Hello all.
I need to build a "pink noise generator" or a "white noise generator", using as final stage an LM386 and a speaker. It has to work from a 9V battery
I've tried 2 configurations for the "white":
1. The "open collector transistor" as a noise source (base = gnd, emitter via a 470k to VCC), followed by an amp. transistor and then the LM386. It works very well, but only from 9V and up. Below 8.5V it stops working at all. I've tried several transistors including RFs, the same results. I have only one type of transistor, not identified (named 668CB) that works ok, from 9V to 6V, but without becoming heated - they all drop dead, without any explanations.
2. I found on "Red" (http://www.redcircuits.com/Page75.htm) another design using TL062 as noice source. It works OK at 3V, but the LM386 is not very happy and getting full oscillations.
Both designs are build on PCBs, and have on the same PCB the noise source and the LM386, and both of them are not doing the job.
I need this device very badly and very urgent. I've spent over a week trying to find a design, circuit, schematics and doing "experiments" with those 2 options, without any result.
I'll be very thankful if any member may help me with this issue.
Thanks in advance,
Tedy

 

[audioguru2]

I've tried 2 configurations for the "white":
1. The "open collector transistor" as a noise source (base = gnd, emitter via a 470k to VCC), followed by an amp. transistor and then the LM386. It works very well, but only from 9V and up. Below 8.5V it stops working at all. I've tried several transistors including RFs, the same results. I have only one type of transistor, not identified (named 668CB) that works ok, from 9V to 6V, but without becoming heated - they all drop dead, without any explanations.

The transistor's reverse-biased base-emitter is acting like a 7V zener diode. Have you tried 3.3V or 4.7V zener diodes instead?

2. I found on "Red" (http://www.redcircuits.com/Page75.htm) another design using TL062 as noice source. It works OK at 3V, but the LM386 is not very happy and getting full oscillations.

Have you tried the TL062 noise source circuit with the battery's full 6V to 9V voltage?
Have you tried operating the TL062 noise source circuit at 3V and the LM386 amp at the battery's full 6V to 9V?

 

[Tedyp]

Thank you audioguru,

The transistor's reverse-biased base-emitter is acting like a 7V zener diode. Have you tried 3.3V or 4.7V zener diodes instead?

Yes, I've tried Zener diodes from 2.7 to 12V - no noise at all.

Have you tried the TL062 noise source circuit with the battery's full 6V to 9V voltage?

I've tried it with a very stabilized P.S. Dead as a rock !!! Between 2.7V to 4 V works like a charm. The interesting part is that at 9V, I could hear and see on my ocilloscope the noise the reversed bias transitor's noise.
At 3V I could hear the TL062 but could not see it on my scope even at the highest amplification (Tektronix 120MHZ)

Have you tried operating the TL062 noise source circuit at 3V and the LM386 amp at the battery's full 6V to 9V?

I've done some experiments with this configuration, but somehow I've got a "motor boat effect" - all the circuit tended to oscillate. As well, even I use very good quality self made PCBs, single sided, and even I use decoupling caps all over the power rails - I cannot understand why it oscillates. I've checked the LM386 data at National, it's exactly the same circuit I use.
Something at my circuit sucks  The PA doesn't match the first stage at all

If you - or other members - have any ideas, I'll be very thankful for help - as I'm spending so many efforts and so much time without results, I really need your help !!!

TIA
Tedy

 

[Tedyp]

I've found that the main problem is that the LM386 oscillates at 9V - no matter which and how many decouping CAPs I put and if there is an input signal or not. Again, it's built on a PCB, not on VERO board. Any idea what to do to stop it oscillating? I'm getting a very nice sinus at output, but as I don't build an oscillator - how can I get rid of it?
TIA FOR HINTS,
Tedy

 

[audioguru2]

Hi Tedy,
I have used many LM386 power amp ICs, followed the instructions in the datasheet and have never had a problem. The datasheet shows a 0.05uF capacitor in series with a 10 ohm resistor from the output pin to ground to stop a very high frequency oscillation, and advises to keep the gain above 9 for stability.

What frequency does your LM385 oscillate at? The motorboating problem that you mentioned before is not caused by the LM386, it is a power supply problem that is revealed by the extremely high voltage gain of your circuits.
Please attach a schematic of one of your oscillating circuits.

 

[Tedyp]

I'm ashamed to tell.... :-[ :-[ :-[
At least I found the reason of the oscillations:

The LM386 works OK, but as I'm using as a speaker a small tweeter, I haven't noticed that as soon as I put it near the PCB, no matter where – I got some kind of "microphony" feedback. This tweeter is enclosed on a plastic enclosure, and it's dimensions are: diameter about 3.5 cm and height about 3 cm.  On it's package is written 240W, but of course it's about 2-3W. It is a PIEZO module, with a crossover built inside, and they claim it's 4 Ohms. Anyway, as I have no choice I have to use it, and I can hear the white noise loud and clear – when the tweeter is about 5 – 10 cm from the main unit, but as soon as I bring it closer – the current begin to rise (at maximum volume, without oscillations, it takes about 180mA) – and it goes up to 250mA with the oscillations. I've tried to put an aluminum sheet between the device and the tweeter, it makes some changes, but as I intend to put the device inside a cigarette sized box from ABS (plastic) – and the tweeter glued outside the box, but very close to the PCB, - I cannot find any idea how can I fix it. I know there is some kind of Graphite Spray, maybe I'll try to use it inside the box, but I cannot believe it may help.

I'll be very thankful if you may have another idea, or if you "met" such a problem.

I'm "grinding" my brain to find a solution for this problem.

I've tried to find a similar metal tweeter – but could not find a cheap one anywhere.

If you may have any idea – please help me

Thanks again for all your good will,

Tedy

 

[Staigen1]

Hi

I got some kind of "microphony" feedback

Sometimes ceramic caps can act as a "microphone", can it be this behaviour of
ceramics that cause your problem? Do you have any ceramic coupling caps in your circuit?

By using a 3.6V zener, properly biased, i belive you can hear its noise clearly by
using only a 386 Amp at 200 times amplification!

//Staigen

 

[Tedyp]

Hi Staigen,
All my decoupling caps are NP0 - do you thing that changing them to disk caps may have such an influence on the circuit that it may stop to oscillate ?
I'll give it a try, no problem, if you think so.
Regards,
Tedy

 

[Staigen1]

No no no! Not your decoupling caps, your coupling caps! ???
A little schematic would be fine too! ;D

//Staigen

 

[Tedyp]

Thank you Staigen,
I've changed the coupling caps first to disk caps, and noticed no improvement.
I've changed them to Tantalum caps - almost no change. I don't know if there is an acoustic or magnetic feedback (I cannot put a resistor load instead of the speaker, as then I'll put off both magnetic AND acoustic), I've tried to use short screened wires to the potentiometer, but now, again, the acoustic power DECREASES as I bring the speaker near the PCB, (the current decreases as well) - and when they are about 1 cm one from the other - I have no audio at all. I've twisted the short speaker wires as well.
You've asked for a schematic. As this project is a relative very simple project, I've designed the PCB with Protel directly from National's data sheet, including the generator. In principle they work OK, but this feedback - from about 5 cm and down drive me crazy, as I have no idea how can I find where it comes from. I have a nice lab, with almost all the best measuring gear etc, but it's the first time I have no idea where to look...
Best regards,
Tedy

 

[Staigen1]

Okay, its your project, good luck

//Staigen

 

[Tedyp]

LOL,
No Staigen, it's nothing secret here , it's just I'm too lazy to turn on my third PC where I have installed the Orcad (yes, legal )and begin to "draw". I use it on more complex projects, but this one, just to spite, small and simple, I cannot find the root of the problem ...  >
Basically it works perfect, just the PCB and Speaker "matching" - the feedback is the main problem. I've wraped the PCB with Aluminum foil - now it has not effect, so I think you are right about Acoustical feedback, but cannot find where from it "enters" - I get this feedbak all over the PCB... Here it looks more as a mechanical (acoustic) problem... as you've suggested, - but I cannot "cure" it...
Regards,
Tedy

 

[audioguru2]

I cannot find the root of the problem ... - the feedback is the main problem.
I think you are right about Acoustical feedback, but cannot find where from it "enters" - I get this feedbak all over the PCB... Here it looks more as a mechanical (acoustic) problem... as you've suggested, - but I cannot "cure" it...

Of course you can fix the problem: just replace any ceramic "microphonic" coupling capacitor with a metallized-plastic-film capacitor.
Connect a full-range speaker to the circuit and tap on the coupling capacitors. Then you will hear which capacitor is microphonic.