Another 1kHz filter and autoranger

[Koifarm]

😱

 

[jan.didden]

Have you checked the operation of the autoranger section through a simulator?

Jan

 

[Koifarm]

No I did not, I just draw it.

This week I will make a test version and optimize it.
I know I need to do something to not let switch the relais on the 1kHz freq.

An extra transistor and diode before the relais can help.

 

[jan.didden]

Let us say that the signal is a 10V peak sine wave and one of the comparators is set at 5V.
So, when the signal sine goes up through 5V, does the relay switch and lower the signal?
Then on the other side of the sine peak, when it goes down from 10V to less than 5V, does the relay switch back?
If this happens, how does the output voltage look?

You really should run this through LTspice before spending time and money on it.

Jan

 

[Koifarm]

Thanks Jan,

using LTspice is not my thing. Just breadboard it also works for me.

 

[AIM65]

Not sure TL072 will be ok driving such a big cap then a relay.

For a basic autoranger you'll need first a peak detector with integration then two comparators in order to drive the relays. Adding hysteresis to the comparators may be required.
TL072 for the peak detector is ok, for the comparator you can use LM393.

In addition, it’s safer to add two Zener back to back at the output in order to protect the sound card.

Chris

 

[Elvee]

The project will need in-depth revisions: two examples, the 100µ directly connected to the outputs of the opamp, and C13 connected to the NI inputs of the opamps without DC bias.
I didn't look further, but that alone is sufficient to make the project stillborn.

 

[jan.didden]

Thanks Jan,

using LTspice is not my thing. Just breadboard it also works for me.

Use a manual simulator. Draw out the signals and voltages and the actions.

Jan

 

[Koifarm]

I had a simulator(iCircuit) on my Ipad.

This works.
Now I must decide to switch on Vmax, Vrms or V p-p.

 

[jan.didden]

So what is the top yellow curve?

Jan

 

[as_audio]

Post 1 :

no dc path for pins 3 and 5 of IC 4, as suggested by Elvee, but this is easily cured if the rest is ok,
but it is not, pin 2 is on 10V and pin 6 at about 1 V, so 3 and 5 can not be connected as shown.

But nice drawing program - which one is it ?

 

[Koifarm]

Yellow curve is 1kHz 990mV input signal.(2.8Vmax). signal is above 1.3V setpoint, therefore output opamp is high at 1kHz.
Diode, capacitor and resistor at output opamp stabilize the output voltage at 9.76V+/- 8mV so transistor can switch on the relais.

I need also to ad input overvoltage protection. Because when applying 100V input signal the 1V path is chosen at the start until the right relais is switched on.(about 100mSec). Maybe I can use BAV99 clamp diode at signal input

As_audio I use designspark pcb with my own Elektor like library.
I will change the schematic tomorrow with all suggestions from this evening.

 

[as_audio]

I like the elektor look, but was not able to find out it works with Linux.

 

[jan.didden]

Yellow curve is 1kHz 990mV input signal.

You do realize that that signal is heavily distorted due to the gain switching?
Not useable for distortion analysis.

Jan

 

[jan.didden]

@Koifarm, before you start breadboarding, carefully review what @AIM65 said in post # 4.
His 2nd statement is crucial for a working design and is not a part of your design.

Jan

 

[Koifarm]

Updated schematic post 1.

• Added input and output protection.
• Added DC reference for inputs TL072.
• Adjusted and tested driving relais(10mA).
• Corrected reference for comparators by factor 10.

 

[jan.didden]

As was mentioned before, the essence of an autoranger is that you develop first a representation of the actual input level, for instance with a peak detector/integrator.
You use that to drive the comparators.
You do not use the actual signal to drive the comparators because then the comparators will switch on the signal shape as is shown in your yellow curce in the screen shot.
Are you aware of that?
I am not asking because I don't know.
I have sold several 100 autorangers; I am asking because I want to save you time, money and disappointments.

A good example is in the AP Systen One schematics available from AP.com; you only have to register.

Jan

 

[Koifarm]

I get your point but I intergrate the signal with the output diode and capacitor at the output of the comparator. Now I am switching on the precies max amplitude of the inputsignal with minimal delay.

 

[AIM65]

Hi Koifarm,

Rather than directly working on the schematic, it's useful to start working on a functional view. Based on what I understood of your project, here's a suggestion of block diagram. If you populate each box with the relevant circuit, and handle properly the box interconnections, this should work. Other solutions exist, but this one is very close to your design.
Upper row is signal path.
Lower row is detection and range management.
Some boxes could be added for signaling : range indication, max detection, etc...

You'll notice that box order matter, that's what Jan says in previous post.
Hope this may help

Chris

 

[1audio]

The input protection need some current limiter. A series resistor or incandescent light bulb are the usual solutions. I think you may need a buffer into the notch.
20 dB steps are really too much to get good accuracy from an ADC. I would like much finer steps but at least 6 dB to keep the ADC in its optimum range.