Designing and Simulating Shaping Amplifiers for Detector Applications

[pretty_fly]

Hello everybody, nowadays I am working on shaping amlifiers and need guidiance about how to design one and do its simulations before pcb design and manufacturing. When googling about shaping amplifiers i found documents mostly about filters (active and passive), which i know works for rejecting some frequencies. So how can i convert a filter to a shaping amplifier? For example designing a CR-(RC)^4 filter which sems a gaussian output shaper.

Looking forward to see your feedback...
thanks in advance

 

[Baluncore]

Do you need to bandwidth limit a digitally modulated signal before it is transmitted over a narrow channel?

Pulse shaping amplifiers change the shape of pulses. They do that by changing the relative amplitude and phase of the frequency components of the pulse. http://en.wikipedia.org/wiki/Pulse_shaping

Op-amp filters can be used for baseband pulse shaping of low data rate signals. Higher data rates usually employ passive LCR circuits.

 

[pretty_fly]

Do you need to bandwidth limit a digitally modulated signal before it is transmitted over a narrow channel?

Pulse shaping amplifiers change the shape of pulses. They do that by changing the relative amplitude and phase of the frequency components of the pulse. http://en.wikipedia.org/wiki/Pulse_shaping

Op-amp filters can be used for baseband pulse shaping of low data rate signals. Higher data rates usually employ passive LCR circuits.

Hi Baluncore

my purpose is to get a gaussian signal, symmetrical shape is best, at the end of my "shaping amplifier". I found the maxim integrated app note "gamma ray detector" circuit. However it gives a bipolar output. So probably i need to learn first the theory of active filters and types, then do some simulations and at the end manufacture it.

I need help for each steps and advices...

1. How can I learn the theories on my self, a good detailed source (I am exploring to find on websites, books, app. notes etc...)

2. Simulation softwares (tried proteus but it gave me 10 kV output for 10mV input on a simple op-amp amplifier ! )

3. Is there any special point for designing the pcb card. (I know the paths must as short as possible)

etc.

btw. I am a physics student not an electronics and I like to work on this subject and want to learn more about hoping to succeed :)

thank you

 

[Baluncore]

my purpose is to get a gaussian signal, symmetrical shape is best, at the end of my "shaping amplifier".

Is that a single uni-polar pulse with a gaussian shape in the time domain? What will you do with the gaussian pulse?
How wide in time will the pulse be? How often must pulses be produced? What decides when it will produce a pulse?
What is your starting pulse shape that you must convert to gaussian? What output voltage amplitude do you require?
How many of these pulse shaping amplifiers will you build?
1. Describe your requirements better so we can identify the field you are working in. We can then give you references.
2. The open loop gain of an op-amp is about 1,000,000. To use op-amps for signal processing you will need negative feedback.
3. PCB layout and design can be very involved. When we know better what your pulse characteristics are we can better advise you.

 

[jim hardy]

a good detailed source

any help here?
http://www.ti.com/lit/an/snoa224a/snoa224a.pdf

 

[pretty_fly]

...

Maybe this picture can be helpful.
t1 ~10 us and A1~10 mV for a sample source.
t2 = 10us (means 100kHz signal/radiation detection --> each rad. particle gives one signal)
A2 = 1V

Gaussian output is for spectrum (identifying different energies of different particles.) Also planning to convert it to a digital signal to count the number of signals. I want to build one to learn the steps of designing a detector amplifier.

About proteus... I mean it is meaningless to get a kv output from an op-amp. So I think proteus is not a good idea to concentrate on for simulating my current issue. Which software you are using??

Thank you jim.

 

[Baluncore]

If I understand it correctly, you want to convert the impulse with an exponential tail into a Gaussian pulse so you can use a pulse height analyser to differentiate between pulses of different energies.You then want to select pulses of a certain narrow range and count those events only.

About proteus... I mean it is meaningless to get a kv output from an op-amp. So I think proteus is not a good idea to concentrate on for simulating my current issue. Which software you are using??

I think proteus is OK but you must simulate a circuit and not just an open loop op-amp.

 

[pretty_fly]

If I understand it correctly ...

exactly :) a different particle with different energy produce a signal with (linearly) different amplitude.
for example: 500 keV particle --> 1V , 250 keV particle --> 0.5 V

(btw not a general rule but mostly detector amplification systems work like this)

You then want to select pulses of a certain narrow range and count those events only.

For example I want to count particles with energies more than 250 keV which produce signal bigger than 0.5 V

This part is out of my plans for now...

I think proteus is OK but you must simulate a circuit and not just an open loop op-amp.

What do you think about LTSpice of Linear technology ??

 

[Baluncore]

You appear to be working in a field that needs similar circuits to those used in XRF analysis. It seems that analogue circuits are now being superseded by digital processors.

http://www-physics.lbl.gov/~spieler/Heidelberg_Notes/
See; All sections, VII is a good introduction to reality and PCB layout requirements.

http://www.niu.edu/~mfortner/PHYS690B/Pulse%20Shaping.pdf

Google “XRF shaping circuit”

What do you think about LTSpice of Linear technology ??

Try them both once you have a minimum circuit module to test.

 

[analogdesign]

You want the CR-RC shaper discussed in Helmuth Spieler's notes that Baluncore linked to.

The preamp can be simple but you need to follow the design procedure in Spieler's notes carefully because it will set your noise performance.

LTSpice is an excellent program for investigating this system. Start with idealized components and go from there.

 

[pretty_fly]

Baluncore and analogdesign thank you for advices.

Currently i got some other advices too and need to read, think and work on all of those. Hope to see your further advices soon

 

[Lojius]

Maybe this picture can be helpful.
t1 ~10 us and A1~10 mV for a sample source.
t2 = 10us (means 100kHz signal/radiation detection --> each rad. particle gives one signal)
A2 = 1V

View attachment 78648

Gaussian output is for spectrum (identifying different energies of different particles.) Also planning to convert it to a digital signal to count the number of signals. I want to build one to learn the steps of designing a detector amplifier.

About proteus... I mean it is meaningless to get a kv output from an op-amp. So I think proteus is not a good idea to concentrate on for simulating my current issue. Which software you are using??

Thank you jim.

Forget the CR-(RC)^4. Pls take a look at Ohkawa et.al. 1978 (http://www.sciencedirect.com/science/article/pii/0029554X76901531). This paper describe the design procedure for " Direct synthesis of Gaussian Filter for nuclear pulse amplifiers" and also https://www.researchgate.net/publication/267338240_Low_cost_front-end_readout_electronic_for_instrumentation_used_in_neutron_experiments - similar method with Salen-Key Low-pass fillter topology. Hope this infos will help you.