How Can I Discharge a 5vp-p, 1khz in 1 Micro Sec

What value is your hold capacitor?

it depends...

U take any values and explain me plz soon....

50pF-1000pF

Short it out with a really short wire (= low enough resistance)?

I can't get if i do like that.....

I'm also kinda surprised the thread didn't take off.

Have a GA for your effort so far .

You've presented a nice peak hold circuit; your mission (should you choose to accept it) is to turn it into a peak detector - i.e. it should also provide a signal indicating that a peak has been found (and the input signal is on the way back down).

Such a circuit (as you probably know) is the basis of a pulse height analyser. Add a windowing function, and it can be used (for example) to discriminate pulses from a radiation detector to count only pulses of a specific incident photon/particle energy (I'm sure there are many more applications, but this is the one I know best).

Not putting you down - lovely job so far - just needs a bit more to turn it into a peak detector.

[No idea whether this was what the OP was looking for - but since you raised the issue ... ]

Well, after five hours of poking through my computer files, multitudinous Google Scholar searches, and reading a few dozen articles in Nuclear Instruments and Methods and IEEE Transactions in Nuclear Science, I have to say, I don't know what to say. It was lots of fun though, and I saved the articles, ranging from the 1956 to 2001 time frame.

First, much of the non-nuclear community seems to use the name Peak Detector to mean what we were talking about, although the nuclear community seems to prefer the name Peak Detect and Hold, although then the articles talking about this fail to carry out the peak-detect function as you described. If you have a reference or two, that'd be great!

I remember being interested in this in the mid 70s, and I recall a differentiator circuit was involved to accurately locate the peak, but I could only find one article using one, and it was simply to get dV/dt for a leading-edge trigger. It would be a differentiator with zero-crossing detector that we'd be looking for. Somehow my search terms and reference and citation checks failed to locate any of the articles I remember.

As far as my rant about analog circuits being preferable to ADC+digital is concerned, for any serious nuclear-sensor application, to do a proper job I'd go with a really fast ADC, like one of those GHz folding-amplifier parts from NSC, and programmable FPGA processing, rather than stand on my head with a small pile of analog tricks. Sheesh! But I suppose that still leaves room for a simple analog circuit that tells us when a new peak is reached, a peak detector. :-)

BTW, for anyone interested, the polarity of D2 needs to be reversed in my drawing.

Well, if you're really interested, here's the circuit of a nucleonics amplifier/single channel analyser/EHT control board I designed back in '97. It worked (and AFAIK still does!).

I realize that you won't be able to make out much (if anything) of what's going on here, but if you're still interested, I'll send you a pdf of the circuit, and try to do a few notes to explain what goes on. PM me your e-mail address. May take a while - don't hold your breath!

You're quite right about the use of fast ADC's in state-of-the-art nucleonics systems. Personally, I don't like them, because they have little (if any) input filtering, and the poor processor running the system tends to get overloaded by any little bit of pick-up (local radio station etc), and can't cope with the signal it's supposed to be looking for.

Ty for ur circuit sir. But i want a simple circuits with automatic switch which can hold the peak automatically for a particular time....

What do you mean, automatically? What do you mean, a particular time? It's hard to be helpful when the question is so vague, we're bound to dissatisfy, aren't we?

What did you think of the circuit I posted trying to help you? It should at least be worth a G.A., shouldn't it? If it lacked some feature, what was it, the new-amplitude-has-arrived pulse concept John mentioned? If so, how would you define it, what rule would you propose? Did you want some kind of a timer function added to the circuit? If you'd really like help and are unhappy with the answers, then give us some specific feedback.

There are at least two serious parts of engineering involved here. The first is creating a detailed specification (or in the case of this forum, a detailed question). The second is coming up with a circuit or solution to meet the specification (or in the case of this forum, to provide advice toward that goal). Both parts are equally important to spend serious effort on. But you have to do the first part. (And, I should add, after we provide some help and teaching, you can also do the second part. :-)

Sry sir if i have told anything wrong. I hav designed a above circuit. But i also wanted to design a circuit which will hold the peak for say 1msec and then discharge and then again hold the peak like that. I hav understood ur circuit. But iam giving inputs as 5Vp, 1MHz sinewave and 5Vp, 1MHz pulse wave as input then i was not getting the output as u shown in ur outout waveforms. To get the outputs as ur waveforms what i hav to give inputs can u tell me plz....

Is that an OP467 you're using in your spice circuit, it's hard to read. The '467 is one of those fast-slewing CFB-in-VFB-clothing opamps. It has an npn+pnp input stage, and usually you don't know from part-to-part what polarity the bias current will be. I mention this because your circuit has nothing to define the current in D2, and to be sure the op-amp has a proper source for its input bias current. The OP467 spice model takes a Boyle-model shortcut and assumes npn transistors only, helping bias D2, but with a rather small sinking current. It's not a good scene and can't be relied upon to work properly on the bench, even if it might in spice. If anything, the pnp transistors in the real part likely have lower beta, and the real op-amp would have a sourcing bias current. That would back-bias D2.

It appears you are running the peak-tracking capacitor open loop, that is, you hope the op-amp will bias D1 a diode-drop above the cap voltage that you'd like to see, by virtue of a diode drop across D2 in the feedback. The flaw with this idea is that you don't know what the current will be in either diode, leading to as much as 500mV of uncertainty. You could define a decent current in D2, 1mA, etc., with a resistor to Vee, but that still leaves D1 with widely-varying currents and forward voltage drops when it's on or partially on. That's why my circuit has the diode inside the feedback loop.

As for your discharging circuit, it looks like you have a 200-ohm resistor to ground - that's pretty small, and would give you 0.2us discharge time constant, how can that be helpful?

You need to think through what you'd like to happen with respect to your automatic-discharge idea. What triggers it? What's happening with your signals in the meantime? Are they continuous or pulsed, etc.?

BTW, were you going to award me a G.A. for my first answer, with drawing? :-)

By usiing the above circuit i discharged the the given input 5Vp, 1Mhz in 1microsec

I hav another question that if i want to hold a peal for 10 usec and then i want to discharge what i can do modification to this circuit....

Win has already given you the answer in #6 - why not study it (paying particular attention to the transistor)?

_{And while you're at it - give #6 the GA it deserves! Use the button.}

Yep, believe it or not, it is possible to make out the details and understand the circuit you posted (after looking at a better copy and reading your guided tour, that is!). Amplifier stage(s) with programmable gain, followed by a pair of comparators making up a pulse-height window, with levels set by DACs programmed from a uP. A oneshot and an inhibit gate to create the output pulses.

You've got it. But I reckon it would be tricky without reference to the 'better copy' .

[Have changed the title, cos I've had enough of looking at the word "Cischarge". The rest of the title still doesn't make any sense, but at least I've got rid of that word ].

[Edit: I see that Admin have changed the thread title, but the miss-spelling still propagates through the thread].

Sir can u plz explain why we used 5.6k, 1k infront op2(U1B)

You have extra parts in there that make no sense to me.

can u say what r those?

In order to make the high input impedance of op-amp.