Hi guys, I`m not sure whether this is the right place for this question, but I`ll give it a go. So I was making the first stage of an input signal sensor. It`s just a TL071 with its inverting input grounded and an AC signal fed to the non-inverting terminal. The idea is to pick up even a few mV of input signal and get me a square wave on the output, so I need a considerable gain. The problem is, that whenever I touch the input jack with my finger, the output responds and gives out an amplified signal. This will totally mess with the next part of my schematic, and render it useless to me. Is there a way to attenuate or completely remove this sort of ,,static``. Or maybe another approach if you have something on your mind. Here`s the circuit I had in mind. An input signal should activate the tube amp, and the lack of an input signal should disable it after 10min or so.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
What's the desired input signal? If it's far enough from 60Hz, you might be able
to null out the hum with a tuned filter.
Well it`s an audio source. I guess there will be a lot of frequencies above 50Hz, but bass frequencies are most successful in charging up the cap, I don't think it`s good idea to attenuate them a lot.
1) WHAT is the signal you are trying to detect?
2) where does it come from?
3) why do you attenuate it so much (47:1) at the input?
4) why do you touch the input plug with your finger?
To me it sounds like getting inside a pool and then complaining why are you wet.
EDIT: where did you get the circuit from and what´s it supposed to do?
5) U2 is not biased, it´s being used as a comparator which is possible but a real comparator would work much better and in any case the negative rail (in this case ground) is NOT within the common mode signal permissible range, so I doubt this circuit ever worked as intended.
6) R12 is too low value.
7) an LM7805 can not give you +12V
8) IF you have only 5V , it´s WAY too low for TL07x anyway
Way too many errors so please link to the page that suggested this, I´m curious to read it.
2) where does it come from?
3) why do you attenuate it so much (47:1) at the input?
4) why do you touch the input plug with your finger?
To me it sounds like getting inside a pool and then complaining why are you wet.
EDIT: where did you get the circuit from and what´s it supposed to do?
5) U2 is not biased, it´s being used as a comparator which is possible but a real comparator would work much better and in any case the negative rail (in this case ground) is NOT within the common mode signal permissible range, so I doubt this circuit ever worked as intended.
6) R12 is too low value.
7) an LM7805 can not give you +12V
8) IF you have only 5V , it´s WAY too low for TL07x anyway
Way too many errors so please link to the page that suggested this, I´m curious to read it.
Last edited:
The circuit is mine there`s no link to it. The idea is as It follows. An input signal (audio signal around 10-200mVrms) will start triggering the comparator U2, shaping a continuous square wave at the output. The output will charge the capacitor through the envelope detector and when the voltage rises above 2V it will activate the relay RY2. That will enable the tube filaments. Meanwhile the activation of U3 will disable the transistor Q2. The capacitor C1 will start charging and after 15-16s U1 will trigger the relay RY1 and enable the HV supply for the tubes. The lack of input signal will start discharging the capacitor C3 through the high value bleeding resistor, and after around 9min the capacitor will discharge below 2V and close the filament driving relay RY2. The op amp will also enable the transistor Q2, which will discharge the capacitor C1 and disable the HV supply for the tubes, thus rendering the amplifier inactive. The rest of the circuit is okay, I`m just trying to figure out a better way to convert any signal above 5-10mV to a rail to rail square wave at the output of U2.
P.S. I made some modifications, this is not the final circuit. Basically the circuit looks the same, but all regimes were recalculated for a 12V power supply. I figured 5V is too little for a TL071
P.S. I made some modifications, this is not the final circuit. Basically the circuit looks the same, but all regimes were recalculated for a 12V power supply. I figured 5V is too little for a TL071
Last edited:
Read post 4 carefully. Ask about anything you don't understand. Deal with the points he raised. That will be a good start on getting a working circuit.
By the way, D1 etc. do not form an envelope detector. The time constants are all wrong for that. It is a peak detector, with a slow rise and a very long decay.
By the way, D1 etc. do not form an envelope detector. The time constants are all wrong for that. It is a peak detector, with a slow rise and a very long decay.
An envelope detector is a peak detector, but here the peaks are with the same amplitude, and the time constant is extremely elongated. Anyway, the rest of the circuit I deem ok, it`s just the comparator I`m not sure about.
Hi thebugger 🙂
Glad to help you and congratulations on designing your own .
That said, although the general idea is fine, I guess you want an audio triggered tube amp, so it will turn only on the presence of audio and self turn off if it diasppears for a relatively long time, but as they say, "the devil is in the details".
Not only expected audio voltage is needed, but also "where it comes from" , because its lowish input impedance will load most anything you are getting your signal from.
Also suggest (and trying to keep things with minimal change) that you substitute U2/U3 with a single LM358 which is way better for that duty, specially because its inputs can reach the negative rail .
You should remove R8 which kills sensitivity and messes with levels, and instead apply some positive voltage, from 10 to 100mV to U2 negative input, that way when signal peaks surpass that level U2 output becomes positive and triggers the delayed relay turn on circuit.
You can make that comparator bias adjustable, so you set it to a sensitivity level you like, and at the same time, it works like a noise gate, not being triggered at random by low level noise.
IF you want to make that circuit better, build it as is for now (with corrections) but start studying 555 timers, simple and VERY useful building blocks, and will help you get better defined and stable timing.
Still curious why would you touch the plug with your finger and expect nothing to happen 😉
Glad to help you and congratulations on designing your own .
That said, although the general idea is fine, I guess you want an audio triggered tube amp, so it will turn only on the presence of audio and self turn off if it diasppears for a relatively long time, but as they say, "the devil is in the details".
Not only expected audio voltage is needed, but also "where it comes from" , because its lowish input impedance will load most anything you are getting your signal from.
Also suggest (and trying to keep things with minimal change) that you substitute U2/U3 with a single LM358 which is way better for that duty, specially because its inputs can reach the negative rail .
You should remove R8 which kills sensitivity and messes with levels, and instead apply some positive voltage, from 10 to 100mV to U2 negative input, that way when signal peaks surpass that level U2 output becomes positive and triggers the delayed relay turn on circuit.
You can make that comparator bias adjustable, so you set it to a sensitivity level you like, and at the same time, it works like a noise gate, not being triggered at random by low level noise.
IF you want to make that circuit better, build it as is for now (with corrections) but start studying 555 timers, simple and VERY useful building blocks, and will help you get better defined and stable timing.
Still curious why would you touch the plug with your finger and expect nothing to happen 😉
Thanks man, the signal source is a low impedance one. It`s from my phone so it should handle impedances of as low as 16ohm without a problem. I will look for a better op amp, I chose this one because I had 20 or so at home, and it`s a jfet input, so the input impedance is very high, which is beneficial for my purpose. Yes that`s a good idea, I will try pulling up the voltage on the negative terminal by 10-20mV positive, to act as a threshold. Now is there a way to get rid of the buzz that happens when I touch the input jack. It`s not about audibility, it may false trigger the circuit accidentally.
There is nothing wrong with your opamp. It is the way you are misusing it. Do you understand point 5 in post 4? All it needs is some input bias - on both inputs.
Have you never touched the input of an audio amp and heard hum?
Have you never touched the input of an audio amp and heard hum?
Why does it need biasing. The input constantly swings above and below the threshold. It doesn't matter that the threshold is 0V, the triggering signal is AC. I kinda figured why the opamp is buzzing when I touch the input. My oscilloscope has some pretty heavy transformer inside. It's inducing EM fields in my body, which couple to the preamp circuit. Since the gain is super high, it affects the output. I think an earth ground should fix the problem, also some shielding, and a properly constructed toroidal transformer.
Don't blame your 'scope transformer for hum. Most 'finger hum' comes from capacitive coupling to mains wiring, which will be all around you in a typical dwelling. I don't know why you want to 'fix' it, as it is not a problem but a useful diagnostic tool.
The opamp you have chosen needs biasing because it won't work properly unless both inputs are within the required voltage range (which does not include zero). I would arrange some bias, and add some feedback, so the first opamp simply functions as a high gain amplifier. The threshold is then set by the second opamp. I think this will lead to a more well-behaved circuit.
The opamp you have chosen needs biasing because it won't work properly unless both inputs are within the required voltage range (which does not include zero). I would arrange some bias, and add some feedback, so the first opamp simply functions as a high gain amplifier. The threshold is then set by the second opamp. I think this will lead to a more well-behaved circuit.
Cool 🙂
Oh, the lowly 50 cent LM358 is perfect there: was DESIGNED to do interface work (just what you need) , so inputs can reach the negative rail (in fact 0.6V beyond it) , was optimized for single supply and is HAPPY with 5V ; way more sophisticated Op Amps can not do that.
Yes, whatever is needed so audio triggers it and noise does not.
Short answer: no.
Your body is an excellent almost 2 meters long antenna and you are *submerged* in strong 50Hz fields (wall wiring) so the problem is you
, not the circuit.Just don´t touch the plug tip 😎
OR wear this aluminized suit, just remember to ground it:
An externally hosted image should be here but it was not working when we last tested it.
hey thebugger
are you getting this noise when your touching the ground sleeve/barrel of a 1/4 jack or RCA connector or is it the tip?
if touching what should be case/exterior (assumed also to be ground) creates noise that warrants further examination, but if your touching the tip of an unterminated plug you've succesfully done what i call a "circuit disturbance test"!
try using your thumb to contact the ground on the plug before applying your finger to the tip if the noise is less, rejoice things are normal...
seems that this is an exercise in creating an autodetect on off scheme?
are you getting this noise when your touching the ground sleeve/barrel of a 1/4 jack or RCA connector or is it the tip?
if touching what should be case/exterior (assumed also to be ground) creates noise that warrants further examination, but if your touching the tip of an unterminated plug you've succesfully done what i call a "circuit disturbance test"!
try using your thumb to contact the ground on the plug before applying your finger to the tip if the noise is less, rejoice things are normal...
seems that this is an exercise in creating an autodetect on off scheme?
Last edited:
Yes it only occurs when I touch the active terminal of the op amp. Or maybe I shouldn't ground myself, I should earth ground the circuit itself. What do you guys mean the LM358 input can reach the negative terminal? The negative terminal is grounded.
P.S. Earth Grounding of the circuit completely removes any traces of input noise induced by my finger.
P.S. Earth Grounding of the circuit completely removes any traces of input noise induced by my finger.
Last edited:
For any opamp there is a range of input voltages for which the opamp will work as an opamp. Go outside this range and the opamp will misbehave in some way: clamp its output to one rail, go to some other output voltage, suffer catastrophic damage, oscillate etc.
For many opamps this range does not include the supply rail voltages, so the input terminals (both of them) must be biased to somewhere between the two rails.
Did you not realise this? Have you not understood what we have been trying to tell you?
For many opamps this range does not include the supply rail voltages, so the input terminals (both of them) must be biased to somewhere between the two rails.
Did you not realise this? Have you not understood what we have been trying to tell you?
The circuit's gnd rail is being earthed. And biasing won't work for me. The voltage difference is too little (a few mV) and even a component tolerance of 1% may prove to be too much. Plus, I haven't noticed any misbehavings. I'll substitute the U2 with a better op amp, like LM258 or something, because I noticed the slew rate is kinda low for the TL071, and at higher frequencies, the output looks more like a triangle wave.
P.S. The common voltage mode of TL071 is 4V I think.
P.S. The common voltage mode of TL071 is 4V I think.
Last edited:
Then what is this thread about?Thebugger said:
All you need is a highish gain AC amp feeding a signal detector, fed from a single supply rail. Surely that is not too difficult to design? Who cares what the output looks like when all you want to do is reliably detect the presence of a signal?
I still don't know whether you simply don't understand what we are saying, or you are wedded to trying to make a bad design work by randomly changing components.
- Status
- Not open for further replies.