pulse integrator 1

[JTreglio]

I need to develop a circuit that will integrate a pulsed voltage signal and provide an analog output to be fed to a PLC module. The pulses are from one to five milliseconds long, with a pulse rate maximum of around 30 Hz. I'd like to be able to reset the integrator between pulses. With current circuitry, the input voltage is around 1 V, coming from a LEM.

I've got a schematic of an integrator using a LM101A, but I'm not an EE and am more than a little unsure about how exactly to put this together. We don't have enough work for an EE full time, and have not been able to find anyone locally (we're in Gardena, CA) to help us on this.

Any help on this would be greatly appreciated.

Jim Treglio
Molecular Metallurgy, Inc.

 

[nbucska]

Hi Jim:
If you reset it, why to integrate?
Read my FAQ, please !!!!
What is the purpose and requirements?



Plesae read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[nbucska]

And what is "LEM" ?

Plesae read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[cbarn24050]

It sounds like you dont want an integrator. The best way is to use a micro to measure the pulse width each cycle.

 

[Skogsgurra]

LEM = Liaison Electro Mecanique. They produce Hall devices to measure DC current.

As cbarn says. Use a simple micro (there are those in 8 pin housings nowadays) to measure pulse width and a PWM output to produce a DC voltage.

You will probably (for sure) need a comparator to sharpen up your pulse and a smooting circuit plus amplifier to get the (probably) 0 - 10 V needed by the PLC.

I think that you can contact any competent consultant via the net and have this done in a couple of days.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[JTreglio]

In the current configuration, the system will pulse any of six spark-plug like arc devices in some sequence, typically 1-3-5 for ten minutes, then 2-4-6 for ten minutes, and so on. These are used to form a pulsed plasma for an ion implanter. The LEM's measure the arc current during each pulse and the beam current. It is important that we monitor the performace of each of the six arc devices, hence our need to read and reset the integrator between pulses. We have tried the averaging methodology in the past, but it did not work very well. I'm not sure what the problem was, though. It might have been timing.

Jim Treglio
Molecular Metallurgy, Inc.

 

[Skogsgurra]

OK. The width measurement doesn't work for you then. You do need an integrator, as you said. And you do need to reset the integrator, as you said, too.

The best way seems to be using a suitably fast integrator and transfer the analoge value to a second hold capacitor at the end of the pulse. Use a buffer between the hold capacitor and the output.

A good quad opamp, an analogue switch (possibly a 4066 style) and some logic is all you need. Plus, of course, a few capacitors and resistors.

Depending on how you define "the end of the pulse", you will get a small and constant offset error, which can be adjusted for. You can then reset the integrator and wait for the next pulse.


Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[nbucska]

Before we design, start with the specs.
What kind of accuracy do you need per pulse and per
10 min. cycles? What is the speed of the PLC?

If I understand, You use a HALL device to measure the
current of the arc. Why not optocoupler? the HALL is less
linear.




Plesae read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[cbarn24050]

You can still use a micro, it takes a numer of samples during the pulse. Even a cheap PIC would get at least 30 samples.

 

[itsmoked]

Depending on what your triggering is and whether or not information is available as to when the pulses will occur (or not) can change the design a lot. From what I perceive from the above comments and what I've experienced in semi-equipment designs I've done, there are often many tweaks along the way or recipe differences that can render fixed values problematic.

I usually reach for a micro to do it all or to control the analog stuff so changes can be implemented in software in lue of hardware. Micros are just so much more nimble in tuning and for the inevitable modifications that come along.

Your LEMs are horrifically expensive any chance of using shunts?

Is this a 1-off or something you're going to make many of?

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[nbucska]

What is the voltage and current range of the pulses ?
Please specify required accuracy etc.


Plesae read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[JTreglio]

We control the pulse rate and pulse length by computer -- these are inputs. The voltage range is from 0 to 5 volts representing 0 to 2 A on one line, and 0 to 200 A on the other. The LEM's are already in place, and were used in the previous design -- this is an upgrade. This is the only one of its kind in the world, but at some time we may expect to make more. The accuracy does not have to be too great. I'd be happy with 1%, but can live with 5%.

Jim Treglio
Molecular Metallurgy, Inc.

 

[ScottyUK]

Arcing and HV? Stick with the LEMs!

You might want to consider using some metallic foil as a shield connected to frame earth inside the bore of the LEM, especially if your conductor voltage changes rapidly relaltive to ground. Many LEMs are sensitive to breakthrough when high dv/dt on the conductor appears on the LEM output as a false current signal. If you can sense current on a conductor at or near earth potential the effect does not occur. The shield must not form a shorted turn: I used a Chomerics self adhesive foil to line the bore and cover the end faces only and soldered a drain wire to the foil.


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[Skogsgurra]

Hear! Hear!

That is real good advice. Arcing inevitably produces high dv/dt. And it couples into anything nearby.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[nbucska]

I suggest using a PC compatible computer card (e.g. from
bagotronix or jkmicro (

http://www -.com)

from $69.-

Use an 8 bit ADC and just integrate it in software.


Plesae read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[DHambley]

Use one high-impdedance opamp (like OP2177) and an analog switch (like ADG442). The integrator consists of a resistor to Vin- and a cap C1 from Vout1 to Vin1- in parallel with SW1. Vout1 drives another cap C2 through a resistor (for stability) and SW2. Set up your logic pulses so that SW1 turns off during the integration and SW2 turns on shortly near the end of your integration to hold the signal. Buffer C2 voltage with the second half of the opamp (simply Vout2 tied to Vin2-) and there you have it...an accurate integrate-sample-hold-buffer for less than $5.00.