Hi, I have built the circuit shown below to help me test a peice of equipment, it is supposed to pruduce white noise when SW1 is open, and aproximate pink noise when SW1 is closed. It seems to run fine when I want it to produce white noise, giivng a good output level for me to use. However, when I close SW1 the output level drops off dramaticaly (by about 30dB), although, the noise spectrum is shaped in the way I want it.
So, can anyone point out why the signal is being so strongley attenuated when I switch in the capacitor, and give any possible solutions for me to try to solve this problem.
(As an aside, I have built this on matrix board so far using point to point wiring, and I noticed today that if I conect a jump lead to a number of points on the board (notably, at the inverting input of the secon op amp, or after the otuput limiting resistor) I can pick up AM radio with the circuit
.)
So, can anyone point out why the signal is being so strongley attenuated when I switch in the capacitor, and give any possible solutions for me to try to solve this problem.
(As an aside, I have built this on matrix board so far using point to point wiring, and I noticed today that if I conect a jump lead to a number of points on the board (notably, at the inverting input of the secon op amp, or after the otuput limiting resistor) I can pick up AM radio with the circuit
.)Attachments
i tried that one, but it wasn't too happy running from only one supply rail in my case (the breakdown voltage of the transistor I am using is around 7.3V, so I can cope with only one). And I do need to keep this down to one battery, as it will be getting quite heavy usage, and it needs to be quite small and light for its aplication (it also has a small IC power amp stage bassed on an LM386N - 3 driving a samll speaker or shaker).
As far as the filter goes, I am hapy with wht it is and isn't letting through as far as the frequency content of the signal goes, but the output level drops way too much for the filtered setting to be of any use at the moment, and I can't figure out why.
As far as the filter goes, I am hapy with wht it is and isn't letting through as far as the frequency content of the signal goes, but the output level drops way too much for the filtered setting to be of any use at the moment, and I can't figure out why.
I think the drop is to be expected. The conversion of white to pink means that the power per bandwidth unit drops by 20dB/decade. I assume you measure the RMS power (but also when you listen to it) the level necessarily is much lower. Look at it this way, you put in an attenuator by loading the source with the cap.
One solution would be to gang S1 and switch the gain of the output amp. But be carefull, with higher gain in the white setting your LF level may overload whatever you are testing.
BTW, you can delete C3, R6, R7 without problems. Save a bit of power.
Jan Didden
One solution would be to gang S1 and switch the gain of the output amp. But be carefull, with higher gain in the white setting your LF level may overload whatever you are testing.
BTW, you can delete C3, R6, R7 without problems. Save a bit of power.
Jan Didden
That filter isn't close to being pink. 6dB/Oct above 143Hz.
It is pretty-pink from say 50 to 500, though the error isn't trivial.
Real pinkers have several caps:
http://www.spectrum-soft.com/news/fall98/pink.shtm
http://www.users.bigpond.com/vk3jaj/pinkfilt/pinkfj07.html
> the noise spectrum is shaped in the way I want it.
If you like the shape, fine. I use an odd pinkish-noise with less power in the top octave so as not to burn tweeters.
> when I close SW1 the output level drops off dramaticaly (by about 30dB)
Passive White-To-Pink filters always do that. Usually you stick a resistive network into the White signal path so the White mid-band level is similar to the Pink level. Switching C4 from a 1uFd cap to a 50 ohm resistor would be a quick fix.
> the breakdown voltage of the transistor I am using is around 7.3V
They are all around 7V. When your "9V" battery sags to 8V then 7V, the character of the noise will change and then stop. You will get more consistent results if you can sneak a second 9V battry in there. Unlike the main battery, this second one just feeding the noise junction will last forever. Possible alternate is a small 12V "security" battery that sells for $1.
It is pretty-pink from say 50 to 500, though the error isn't trivial.
Real pinkers have several caps:
http://www.spectrum-soft.com/news/fall98/pink.shtm
http://www.users.bigpond.com/vk3jaj/pinkfilt/pinkfj07.html
> the noise spectrum is shaped in the way I want it.
If you like the shape, fine. I use an odd pinkish-noise with less power in the top octave so as not to burn tweeters.
> when I close SW1 the output level drops off dramaticaly (by about 30dB)
Passive White-To-Pink filters always do that. Usually you stick a resistive network into the White signal path so the White mid-band level is similar to the Pink level. Switching C4 from a 1uFd cap to a 50 ohm resistor would be a quick fix.
> the breakdown voltage of the transistor I am using is around 7.3V
They are all around 7V. When your "9V" battery sags to 8V then 7V, the character of the noise will change and then stop. You will get more consistent results if you can sneak a second 9V battry in there. Unlike the main battery, this second one just feeding the noise junction will last forever. Possible alternate is a small 12V "security" battery that sells for $1.
i should have explained what I was using this for a little better I guess. This will be being used to simulate the noise from a burst water mains to try and demonstrate a product the company I'm working for this summer sell, which picks up sound in the range fo arround 100 - 5000 Hz. So, the fact that this only needs to work to 5000Hz somewhat justifies my poor choice of filter, but I should know by the end of the day if this thing works or not. If not, I might give the filters in that first link (thanks PRR) a go and see if that helps out at all.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.