Practical & Precision Sinusoidal Peak Detector for PIC18F452

Well this is fairly straight forward. You can also consider this. However, without a schematic to view your configuration or any data there's not much I can do for you. You maybe expecting too much from the circuit design. You may not have actually constructed what you hoped to construct. You may be incorrectly reading the PIC microcontroller. I cannot tell.

This sounds a little familiar...

Do you know a fellow named m mirza? Is this the same project?

May be a simple coincidence.

Any PIC is not a peak detector. When you program one it will sample an input, when it gets around it in the program flow, not when that occurs.

You need an analog peak detector, that holds it value until the next peak, or a defined decay time constant. That is definitely Analog Design.

Other than that, I concur with RedFred.

If you want to make it complex then consider that the time derivative changes sigh at the peak i.e. it crosses zero! So you change the peak detection into a zero crossing detection which is very simple with an Op-amp and output limiters.

That's only true if the waveform can only be a single sinusoid wave with one and only one fundamental frequency. A polytonic waveform, like a bandwidth limited approximation of a square wave, will have multiple instances in time where the derivative crosses zero in each periodic sequence. Only one of these will will occur at the peak amplitude.

I agree with you but he mentions a "sinusoidal" signal with different frequencies he did not mention it is a superposition of several signals.

My suggestion was based on the assumption he analyses a clean signal.

I am still aware of the fact that the derivative of a sum can go through zero several times over the lowest frequency period.

It was good to mention for other readers.

I agree with you I am a bit taken aback how few are those who base their answer ONLY on own words and explanations. I could say most are excellent users of search engines, I can do it as well but my pleasure is to take from my memory and knowledge as much as possible for the answer. I have the feeling that you do the same.

Hello and thanks a lot for your answers.But I meant that I want to measure a

sinousoidal input signal that vary its frequencies from 1HZ to 1KHZ at each specific

time and at each time I want to measure its peak.

Thanks,

Hello.Indeed I want to display a 100KV ac voltage [50HZ] may vary its frequency

on the LCD and I have a 100V as a sample of that 100KV ac voltage.

So,I try to find a way to display that voltage precisely.[using mikroC PRO V5.3

compiler,PIC18F452,XTAL=40MHZ].

Would you help me for that?

Thanks,

First of all - as you are new here - the ground rules need to be repeated. You get free advice, and you can take it for what it is worth. We do not do homework. You do your heavy lifting. With that out of the way:

1,. You picked a way too complicated PIC for such a simple task. PIC is so cheap, even later expansion dreams are never realized. K.I.S.S.. Figure the smallest, then 4x the memory, and that is the ballpark.

2,. Analog Design is coequal, if quality performance is expected. Generally, analog guys can program, but do not like it. Programmers, on the other hand, frequently believe they can cookbook it, and it does not work. Example: how much smoothing, averaging, spike and noise suppression do you want in analog? What about allowable phase delay? I do not care if you make them happen. But - a power bus being a badly noisy environment - drop them with care, if you want something useable.

3,. What is the precision and impedance of the 100V sample? If, later a current transformer is added, the same? Specification creep is normal in any proiect.

3.1,. How do you transport your sample(s) to a quieter environment? Because a PIC will go bonkers near to a power bus.

4,. Did you design for such eventuality?

5,. Spend 1/4th of your time in specmanship, exploring these and such. Then, mostly you know what you want to realize in some detail.

6,. Then you can spend 1/4th time in laying out, and refining hardware and software block diagrams.

7,. Then you do know, how much Analog help is needed. At the same time, you can start programming with confidence. A seaserpent, flow of consciousness programming is a guaranteed path to failure. I took a few bets, never lost one on that.

8,. An example for decision. Peak detector is good for pure sinus, that does not exist on a generator's bus. Far from it. Decide, if the solid harmonic is important to you? Dou you want to go to a multisampling / integrating route? For timing, an Analog massaged Zero Crossing is the way to go.

9,. How precisely do you can measure 50Hz? It is a trick question with measurement theory. Software doing any? Yeah, right.

10,. If this is a thesis work, or some such, pare it down to the minimum to do. Pay due obeisance to issues left out. The go on with life.

Contemplate on it.

Good Answer.

I'm afraid though that our OP has taken a task far larger than they can handle and just wants code thrown at the problem. I also very much worry about their safety in this attempt. 100KV is not an environment for a novice. 100KV at 50 Hz implies to me power line distribution with enormous amounts of lethal instantaneous power being available. I realize that this might be instead be step-up transformer for some other purpose that needs to be monitored. Power distribution or step-up, I'm certain the concerns of corona discharge have not been considered if the dynamics of a simple analog peak detector circuit have not been considered.

Hello.My special thanks to Mr.Leveles & Mr.Redfed for their precise analysis of my project.So, I am still at begining of my project way to realize that.

Thanks,

You are very welcome. I hope that your 100KV is either a typographic error or that your PIC based instrumentation is actually getting a linearly scaled down signal for you to measure. As I said earlier 100KV is a very dangerous voltage to work on and equipment working at this voltage should not be worked on casually. As the video shows, 100KV will easily arc through air.

One complication that I did not mention earlier about using a peak hold circuit to measure the peak voltage that have baffled many engineers. While many function generators successfully generate very clean sine wave signals, very few real world signal sources are this clean. As such any apparent error voltage level produced by a peak hold measuring circuit may actually be due to a real signal plus noise or interference.

when i was a kid . . . (? do i need to tell futher)

so when i was a kid i concidered that if i put the isolated (cleaned ends) wire - the thinnest i got to a ~220AC outlet then it will fail before the mains - the both went - iwas blinded some time by the flash of spark cloud ... (?what) ... when i checked my thumb and pointer fingers - i could admire that isolation had saved my skin just under it but not around - it (the ash) came off with the soap and water mostly

so everything near above ~400 V (0.4kV) tends to set up an air traffic (man is a very good ground - even with no earth around/avail.)

"a 100V as a sample of that 100KV ac voltage."

? down transformed
? Gigaohm resistive voltage divider
? u name it

no experience with 100KV K is for Kelvin k is for kilo- (i almost hate non physicist - they put it even on product labels WTH?)

no eperience with 100kV eighter - but sampling of that range is tricky i suppose (all my digital meters fail near 2kV (as location/close by) metering (something low amps -range) )

... ok ... suppose the stuff is metered by resistive voltage divider - there is some issue other than EM field influence to apparattus ...
(±-.1MV) -[99900]--[100]-(±.1MV) , <-- the problem likely rises with such
(±-.1MV) -[49950]--[100]--[49950]-(±.1MV) ? donno ...

... other words ground or 0 tracking -- as i said no practise with .1MV

It is possible to use a solenoid to induce a curent and work wit it. The field is concentric and the solenoid can be placed with its axis perpendicular to the cable. The obtaind voltage can be a lot less and not as dangerous as the original tension.