# 0-30 Hz White noise stimulation device (College project)

#### wayneh

Joined Sep 9, 2010
16,398
Im not sure I understand how to do this so I think I may need to simplify my requirements further.
You're doing it without knowing you're doing it. You're mapping every possible input voltage to an output current.

One way to process the input signal would be to use a dual op-amp, one in inverting mode and one non-inverting. Both could be set to the gain required to boost the input noise signal from whatever it is, say 1V p-p, up to the rail voltage, say 9V p-p. Bias the noise input and opamp reference voltages to one half the rail voltage. The two outputs would then drive the constant current circuit.

#### Raymond Genovese

Joined Mar 5, 2016
1,658
I have been following this thread and find it interesting. Please, experts, don't jump all over me, but instead, try to explain, if you will...

Why couldn't he use a digital white noise source?

Something like this? I remember transferring that code for a 202 with no problem at all, although I was only interested in masking noise.

It would seem (by my limited knowledge) to offer some advantages - easily adjust the current output, and adjust to 0-30Hz by varying the frequency of bit spitting.

#### wayneh

Joined Sep 9, 2010
16,398
Why couldn't he use a digital white noise source?
As my college project, I have to design and build a prototype of an electrical stimulation device.
Being homework, I suppose his off-the-shelf options are limited. Otherwise I have no answer or input on the noise generation aspect of the project.

#### Kellin

Joined Feb 17, 2019
20
You're doing it without knowing you're doing it. You're mapping every possible input voltage to an output current.

One way to process the input signal would be to use a dual op-amp, one in inverting mode and one non-inverting. Both could be set to the gain required to boost the input noise signal from whatever it is, say 1V p-p, up to the rail voltage, say 9V p-p. Bias the noise input and opamp reference voltages to one half the rail voltage. The two outputs would then drive the constant current circuit.
Would the inverting op amp and non-inverting op amp produce constant opposite-polarity signals or do they have the ability to switch polarities cyclically? I have designs for including an inverting op amp and non-inverting into my circuit, both taking input from the output of my 8th order filter (outputting 0-30 Hz white-noise), but I cannot find information on whether or not they have the ability to switch polarities.

If I used both the inverting op-amp and non-inverting op-amp to produce two opposite signal and output those to signals through two electrodes, would the positive signal travel toward the inverted signal through whatever is between them? (Eg. a solution with 1k Ohm impedance?)

#### wayneh

Joined Sep 9, 2010
16,398
Would the inverting op amp and non-inverting op amp produce constant opposite-polarity signals or do they have the ability to switch polarities cyclically?
In the scheme I described the two outputs would always be equal magnitude and opposite each other, with opposite polarity. They follow the noise signal as it goes above and below the center reference.
I have designs for including an inverting op amp and non-inverting into my circuit, both taking input from the output of my 8th order filter (outputting 0-30 Hz white-noise), but I cannot find information on whether or not they have the ability to switch polarities.
I don't really understand the question. Can you post a schematic?
If I used both the inverting op-amp and non-inverting op-amp to produce two opposite signal and output those to signals through two electrodes, would the positive signal travel toward the inverted signal through whatever is between them? (Eg. a solution with 1k Ohm impedance?)
Yes, but the arrangement I provided ensures that the voltage is converted to a defined current. Just presenting a voltage will not do that unless you are certain of the impedance. If you're talking about electrodes attached to a subject, there's no way to know the impedance.

#### Kellin

Joined Feb 17, 2019
20
In the scheme I described the two outputs would always be equal magnitude and opposite each other, with opposite polarity. They follow the noise signal as it goes above and below the center reference.
Would the attached schematic work as you described? Take my 0-30 Hz white noise signal and pass it through two op amps, one inverting and one non-inverting, and output the resulting signal through two electrodes so that the positive signal will travel through a steady impedance of 1k Ohms to the negative electrode (created with the inverting op-amp)?

*Apologies, saw a mistake in the first circuit, second one now uploaded*

#### Attachments

• 13.3 KB Views: 6

#### wayneh

Joined Sep 9, 2010
16,398
Would the attached schematic work as you described? Take my 0-30 Hz white noise signal and pass it through two op amps, one inverting and one non-inverting, and output the resulting signal through two electrodes so that the positive signal will travel through a steady impedance of 1k Ohms to the negative electrode (created with the inverting op-amp)?

*Apologies, saw a mistake in the first circuit, second one now uploaded*
How is the input signal biased? In other words, what does the voltage waveform look like?

Your arrangement would work fine if you have a dual supply, and the noise is biased at ground level, meaning it varies ±V about ground. The gain needs to be chosen to keep the signals within the power rails of the opamps, unless you intentionally want clipping.

#### Audioguru

Joined Dec 20, 2007
11,249
The opamps must be powered by plus and minus batteries and be biased at 0V so that their outputs can swing positive and negative.

Two seried 9V batteries will make the voltage high enough for the simple reverse-biased emitter-base noise generator. The junction of the two batteries will be 0V. Then the preamp, lowpass filter and output opamps must also be powered with plus and minus and be biased at 0V.

#### Kellin

Joined Feb 17, 2019
20
How is the input signal biased? In other words, what does the voltage waveform look like?
The voltage waveform for the 0-30 Hz signal is attached below as 'filtered' and the pre-filter signal is attached as 'pre-filter'.

Your arrangement would work fine if you have a dual supply, and the noise is biased at ground level, meaning it varies ±V about ground. The gain needs to be chosen to keep the signals within the power rails of the opamps, unless you intentionally want clipping.
I had a dual supply, one 15 V to power the white-noise circuit and a separate 5 V supply to power my MAX292 filter, but I added a 5 V voltage regulator to power the filter and remove the second power supply as I would like to stick with the one 15 V supply if possible. Are you suggesting the dual supply to power the 2 Op amps separately from the rest of the circuit?

I'm not sure how to bias my noise signal at ground level, could you advise me?

The opamps must be powered by plus and minus batteries and be biased at 0V so that their outputs can swing positive and negative.

Two seried 9V batteries will make the voltage high enough for the simple reverse-biased emitter-base noise generator. The junction of the two batteries will be 0V. Then the preamp, lowpass filter and output opamps must also be powered with plus and minus and be biased at 0V.
I will try power my circuit with two 9V battery in series tomorrow and post the results, thank you

#### Attachments

• 161.2 KB Views: 4
• 154.7 KB Views: 4

#### Kellin

Joined Feb 17, 2019
20
V2 and V3 represent your input signal(s). For this simulation I've set them to be square waves opposite each other. This causes the load to see an alternating square wave, current one way and then switching to the opposite polarity. If instead you allow the inputs to be both high or both low at the same time, the load will see periods of zero current.
View attachment 170844
I have added this circuit from #28 to my current circuit that generated the 0-30 Hz white-noise signal with the single 15 V power supply. As previously mentioned, I have just one output rail on my breadboard for the output signal and in my schematic attached below I am using this as an input into the current circuit you have provided in #28. I see that in your original circuit you have both a positive and negative aspect to your input signals and have just the one output. How do I work around this? Do I need to put the inverting and non-inverting op-amps between your current circuit from #28 and my 0-30Hz output to generate the two different inputs? Once I figure this part out I will be very close to finishing the project, thank you very much for your help.

(The second op-amp should be TLV271 as well, my bad)

#### Attachments

• 193.4 KB Views: 11

#### wayneh

Joined Sep 9, 2010
16,398
That's the one thing that won’t work. If both inputs are set to the same level, there will always be zero current.

I can say more when I return to my laptop. At the park getting some vitamin D!

Last edited:

#### Audioguru

Joined Dec 20, 2007
11,249
Two 9V batteries are used to make plus 9VDC and minus 9VDC. Then the noise source is powered from the total 18VDC.
Since nobody knows the output DC voltage of the noise source then it should be capacitor coupled into the input of the filter IC.
A MAX291 is a Butterworth type with a sharp corner frequency and a fast dropoff. A MAX292 is a droopy Bessel type with a poor dropoff. Use a MAX291. It can be powered from +6VDC or plus 5V and minus 5V.

The differential opamps have much different attenuations that is wrong.

#### wayneh

Joined Sep 9, 2010
16,398
I have added this circuit from #28 to my current circuit that generated the 0-30 Hz white-noise signal with the single 15 V power supply.
What does the output of the MAX292 look like? I haven't studied that circuit but it seems likely the entire signal is between 0V and 15V?

What happened to the idea of creating an inverted signal to contrast with the non-inverted signal?

#### Audioguru

Joined Dec 20, 2007
11,249
The MAX291 is a much better filter than a MAX292. The output is Cmos with a low maximum output current. The max output into a 20k ohms load is plus and minus 4V when its supply is plus and minus 5V.

Since you are using two 9V batteries then the noise circuit can be powered from regulated +15V and the filter IC and opamps can have regulated +5V and -5V.

The output opamps are differential with one as inverting and the other as non-inverting. But their attenuations do not match. I found this opamps circuit that should work fine:

#### Attachments

• 9.3 KB Views: 4

#### Kellin

Joined Feb 17, 2019
20
Two 9V batteries are used to make plus 9VDC and minus 9VDC. Then the noise source is powered from the total 18VDC.
In a prototype model how would this look, would the prototype require two separate batteries to power it?
Since nobody knows the output DC voltage of the noise source then it should be capacitor coupled into the input of the filter IC.
Use a MAX291.
The voltage of the 0-30 Hz signal coming out of the filter is approx. 200 mV. The oscilloscope reading is attached. I have ordered a MAX291 buttersworth filter and will replace my 292 when it arrives.

I haven't studied that circuit but it seems likely the entire signal is between 0V and 15V?
What happened to the idea of creating an inverted signal to contrast with the non-inverted signal?
The way I have it powered at the moment is a single 15 V supply as the input. I've used 30 mA all the way up to 450 mA input for the current which all yield the same 200 mV signal exiting the MAX292 filter. I tried to measure the amplitude of the signal today with an ammeter and it was flickering between 0 A and 0.02 uA which is pretty much nothing. I'm not sure if I'm meant to be getting a current out of the filter. I'm still confused with the inverting and non-inverting signal. I'm confused between the signal biasing, the circuit you proposed in #28, inverting and non inverting op amps etc and I'm not sure what exactly I need to complete the next step in my circuit.

The MAX291 is a much better filter than a MAX292. The output is Cmos with a low maximum output current. The max output into a 20k ohms load is plus and minus 4V when its supply is plus and minus 5V.

Since you are using two 9V batteries then the noise circuit can be powered from regulated +15V and the filter IC and opamps can have regulated +5V and -5V.

The output opamps are differential with one as inverting and the other as non-inverting. But their attenuations do not match. I found this opamps circuit that should work fine:
Please correct me if I am wrong. So I would then have a +9V rail and a -9V rail instead of +15V and ground, or would it have an 18V rail and 0 V rail?. An 18V signal that I would step down to 15V to power the white noise generator circuit (assuming its 18V and 0 V?). The MAX291 filter would be powered from another 5V regulator and the op amp circuit you attached in #55 can also be powered by this regulator. These two op amps would then output a positive and negative 0-30 Hz signal ready to be outputted through the electrodes? And what amplitude would these two opamps output? I need ~1000uA (ideally adjustable between 0 and 1000 uA but a fixed 1000uA will do for the purpose of finishing the prototype as the deadline is approaching).

If I need both the -9 V and +9V as well as the 18 V, do I include a ground connection between the two 9V batteries or not?

#### Attachments

• 130.4 KB Views: 3
Last edited:

#### wayneh

Joined Sep 9, 2010
16,398
The way I have it powered at the moment is a single 15 V supply as the input.
The data sheet I looked at said the max voltage is 12V. Are you sure 15V is OK?
I've used 30 mA all the up to 450 mA input too for the current which all yield the same 200 mV signal exiting the MAX292 filter.
The datasheet says the output should swing to within ~1V of the power rails. So getting only 200mV makes me wonder about the input signal, and of course if the MAX292 is wired and operating properly.
I tried to measure the amplitude of the signal today with an ammeter and it was flickering between 0 A and 0.02 uA which next pretty much nothing. I'm not sure if I'm meant to be getting a current out of the filter.
The datasheet shows the output driving a 20kΩ load. An output of 4V would produce a current of only 200µA. So I wouldn't expect to see more than that. But how did you measure the current? If you placed the leads across the output in ammeter mode, you shorted the output. That might have damaged the chip. If you measure in series with the next stage, then I would expect almost zero current there. I guess you might see <20µA.

I'm still confused with the inverting and non-inverting signal. I'm confused between the signal biasing, the circuit you proposed in #28, inverting and non inverting op amps etc and I'm not sure what exactly I need to complete the next step in my circuit.
This appears to be the source of your confusion from early on. You really need to get a handle this aspect of your project. The circuit in #28 is for a signal contained between the power rails of the op-amps. It makes sense for your application to have the "zero" point biased to half the supply voltage, so that it can swing freely in both directions relative to that point while staying within the power rails.

As we've discussed, I think it would work best for your project to invert the input noise signal to provide the opposing signal. When the noise signal goes positive by 3V above the bias point, that inverted signal would go 3V below the bias point (which would still be above ground and in range). The circuit in #28 then drives a current that is proportional to, and in the polarity of, the noise signal.

You could instead use a single op-amp to accomplish a similar thing. You would set one electrode to a bias voltage of half the supply voltage and also reference one input of the op-amp to half the supply voltage. Then as the signal goes up and down about that reference, the op-amp would drive the floating electrode to any voltage between the power rails. A downside of this approach is that your maximum voltage for driving current is only half the supply voltage. In my project, I was using a single 9V battery (which is only ~7V in reality) and didn't have the luxury of splitting that in half.

#### Kellin

Joined Feb 17, 2019
20
The data sheet I looked at said the max voltage is 12V. Are you sure 15V is OK?
The datasheet says the output should swing to within ~1V of the power rails. So getting only 200mV makes me wonder about the input signal, and of course if the MAX292 is wired and operating properly.
The datasheet shows the output driving a 20kΩ load. An output of 4V would produce a current of only 200µA. So I wouldn't expect to see more than that. But how did you measure the current? If you placed the leads across the output in ammeter mode, you shorted the output. That might have damaged the chip. If you measure in series with the next stage, then I would expect almost zero current there. I guess you might see <20µA.
https://www.experimentalistsanonymous.com/ve3wwg/doku.php?id=noise_generator
This is where I got the schematic, I believed the theory behind it was that you had to give it more than its required voltage to cause the reverse biased transistor to enter avalanche breakdown to produce the white noise. I will reduce it to 12 V tomorrow and see if it has the same effect.

There is nothing on my circuit after the filter so I dont think it was in series with the next stage. I placed the positive ammeter wire at the output of my filter and the negative ammeter wire in ground. It was reading between 0 and 0.02 uA.

"The datasheet says the output should swing to within ~1V of the power rails. So getting only 200mV makes me wonder about the input signal, and of course if the MAX292 is wired and operating properly." Is this for the MAX292 datasheet? The MAX292 datasheet I see states the minimum output DC swing is +-4V. My MAX282 is receiving 5V from a voltage regulator (checked and verified) into pin 5 to power it and a 3.125 V white-noise signal into pin 8 which is in the range requirements on the datasheet too. I will attach my full circuit to this post.

#### Attachments

• 14.1 KB Views: 3
Last edited: