Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Voltage Clamp

Status
Not open for further replies.

ThirtyTwo

New Member
sry if this is in the wrong forum, i dont know much about electronics


I got this schematic off a site that was last updated about 6 years ago, so i couldnt e-mail the author the question. its for my car, my computer on my car cuts the fuel if boost from the turbo goes beyond 11psi, the MAP sensor tells the computer how much boost the engine is taking in. So if you let the MAP sensor send its normal signal until the voltage for 11psi(around 4.7v) is reached and then make it stay at that voltage the computer wont know you are getting higher boost than 11psi(or whatever you set the clamp to). So this device i posted allows current to pass through normally until the voltage set by R3 is reached and then clamps the voltage at that point.

**broken link removed**
This voltage clamp is designed to limit the input voltage based on the voltage set by the trimmer at R3 (Vref is 5v input), and then send it to output. But it doesnt limit the input voltage until the set voltage is reached.

Sorry if thats hard to understand cause i suck at explaining things.

But my question is, i dont have a 12v on my application so i really dont need it on the voltage clamp. And i figured since it has a diode that blocks the 12v to the system anyway i could just take it off completely without affecting the performance of the device. So do you guys think it would affect anything?

And also it says "Noise suppression is achieved via the use of a low pass filter on the input, formed by R2 and C1. High frequency roll-off is at about 100 Hz, allowing the circuit to be responsive but effectively suppressing RF noise.", so does the D2 diode have to stay in the schematics for this to work still? Or can it be taken out since the 12v line and diode was taken out?

Here is what i would like to have it set up as
**broken link removed**

think it will affect it any?
 
"32", that 12 volt line has to stay there. The diodes are series-connected between +12v and ground so that the voltage on pin 3 of the IC will not be damaged by spikes on the input that exceed the supply voltages on the chip. Your original schematic does not show the ground (pin 4) and Vcc or +12v connections (pin 8) to the IC. So you still need the +12v even if you were to eliminate the voltage clamp at the input. The IC has to have the +12v and ground connections just like your TV set needs to be plugged in to get a TV signal to the tube.

So, leave the original schematic as is and don't forget to add power and ground to the chip connections which are not shown on your skem.

Dean
 
so something like this???
**broken link removed**

also, im really new to electronics but when reading a schematic and two lines intersect, how do you know if they are supposed to be connected or not???

like between D1 and D2, do those intersecting wires connect to each other? When i added the line from pin 8 to 12v from the IC on the right i put a gap in the line when intersecting the output line to show that it doesnt connect, but thats probably not how you are supposed to do it, so how would you draw that?
 
Last edited by a moderator:
LM358 is a dual op-amp, so the supply pin 4 and 8.
 

Attachments

  • fcdschem2.gif
    fcdschem2.gif
    4.8 KB · Views: 1,643
I think Sebi was trying to say that since the op amp is a dual op amp, you only need to show the power connections on one of the halves, not both. And traditionally, we put power at the top of the triangle, groung (or negative) at the bottom and leave that left side for the input pins.

(pin 8), that is "pin 8" enclosed in parentheses in my first post came out as a happy face. Really weird.

Dean
 
Inveterate proofreader that I am, I have to point out that 10k and 0.01uF form a lowpass filter with a corner frequency of about 1.6kHz, not 100Hz as the original designer apparently stated.
Also, if your output needs to be zero when the input is zero, you should add a 4.7k resistor from pin 1 to GND.
 
Hi Dean,
You must "disable smiles" when you post if you use low numbers in your text or numbers up to sixteen look like smiles and things.
The number 8 followed by a closed bracket looks like a smiley: 5) 6) 7) 8) 9) 10) 11) 12) etc.
 
ThirtyTwo said:
so something like this???
**broken link removed**

also, im really new to electronics but when reading a schematic and two lines intersect, how do you know if they are supposed to be connected or not???

like between D1 and D2, do those intersecting wires connect to each other? When i added the line from pin 8 to 12v from the IC on the right i put a gap in the line when intersecting the output line to show that it doesnt connect, but thats probably not how you are supposed to do it, so how would you draw that?

Since nobody answered your question about schematics & connection points..

Generally,

1) a net or (line) that ends in a "T" implicitly IS connected.

2) nets that cross are NOT connected. sometimes you see one make a (jump) over the others sometimes not.

3) any net that has a connection "dot" at the points of ambiguity ARE connected. In otherwords, place a DOT to explicitly show connection.

With that said, using the dot method is out of date according to IEEE but my experience is that it is still in wide use. Sometimes the schematic is so garbled looking you can't make out those connection dots. Ever seen a copy of a copy of a copy of a copy? In those cases, understanding of the circuit context is required to assertain connection points that are ambiguous. With printed paper designs declining in popularity, this is less of an issue.
 
wow you guys really know what your talking about lol :)

heres the page i got the schematics from if anybody wants to look at it http://www.teamfc3s.org/info/articles/fcd/diy_fcd.htm

thank you sebi for redrawing it with all the corrections, and dots

i thought there were actually 2 LM358s on the board, thats why i put the power to both. good thing i came here for help cause i would have never gotten this thing to work :(

Ron H said:
Inveterate proofreader that I am, I have to point out that 10k and 0.01uF form a lowpass filter with a corner frequency of about 1.6kHz, not 100Hz as the original designer apparently stated.
Also, if your output needs to be zero when the input is zero, you should add a 4.7k resistor from pin 1 to GND.
is the higher lowpass freq a bad thing??? I dont think its going to be at 0v much on my application but i guess it wont hurt anything to put that resistor in. thanks

I ordered the parts for it and they should be in soon. But in the mean time do you guys think this will work for what i want it to do (output follow the input until the voltage set by the trimmer is reached and then clamp output to that voltage)?
 
ThirtyTwo said:
<snip>is the higher lowpass freq a bad thing???<snip>
Probably not.
<snip>do you guys think this will work for what i want it to do (output follow the input until the voltage set by the trimmer is reached and then clamp output to that voltage)?
I do. I can't speak for those other guys. :D
 
audiogruru, I'll bet if I do (pin 8) as ( pin 8 ), it'll work out OK.

With that said, using the dot method is out of date according to IEEE

That means that there's no way to show a connection of two crossed lines unless one assumes that crossed lines are always connected unless a "hop" is made. "Hops" were popular in 1945 and they made for nasty-looking, difficult-to-read schematics. Another way of showing non-connections was to run one "wire" up to another, stop, and continue the line on the other side, as though it dove through the paper underneath the cross-wire and popped up on the other side. Trouble with that was that printing errors often made that a mess, too.

After working for Tektronix, I've pretty much worked to their B.C. drawing standards (B.C. = before CAD). Most Tek skems of that era are works of art. I find most skems drawn by simulators and skem capture programs to be hard to read compared to hand-drawn schematics prepared by a trained draftsperson. Another source of really pretty schematic diagrams are in the Popular Electronics magazines of the 1960s. The older Radio Amateur's Handbooks were another example of pretty skems.

"32", a close look at the pin numbering of the LM358s would show that the two sides used different pins, usually a sure sign that it's a multiple-circuit chip. The other indicator is the chip designators, "U1-1" and "U1-2", indicating that these are two halves ("-1" and "-2") of the same chip (U1). You'll also sometimes see them marked as U1a and U1b or IC1a and IC1b. It all depends upon the person doing the drawing or the CAD program or both.

Dean
 
ok i downloaded Circuit Maker Trial and put the schem in it. everything seems to work good but my output is always .002v higher than the clamping voltage (when it is clamping), and a variable gain of around .02v when not clamping???

when i put that 4.7k resistor in from pin1 to ground it wouldnt work at all
 
ThirtyTwo said:
ok i downloaded Circuit Maker Trial and put the schem in it. everything seems to work good but my output is always .002v higher than the clamping voltage (when it is clamping), and a variable gain of around .02v when not clamping???

when i put that 4.7k resistor in from pin1 to ground it wouldnt work at all
I ran the sim on Linear Technology's SwitcherCAD III, and it worked fine with TI's model, with or without the 4.7k load. With National's model, it wouldn't converge unless I made the load 10k or less. I don't know what the cause of your problem is. Can you post the schematic you used, and whose model you used?
The 2mv offset is built into the op amp models for LM358, as a reflection of the real world. It can be as high as 7mv and still be in spec.
 
yeah i just went through and redid the schem on the program and it works now, i think i was using the wrong op-amp (lm358/ns) :oops:

im using the schem that sebi posted, i have no clue what TI or National means lol

thanks for all your guys help though, it looks like its gonna work pretty good :)
 
Hi 32,
TI means Texas Instruments and NS means National Semiconductor, manufacturers of the LM358.

I hope the mixture doesn't end up too lean causing detonation which might blow your engine's head off. (do rotary engines even have a head?)

I had a car with pistons and a turbo. After a repair that pinched the vacuum hose to the waste-gate, it had so much power that it was a rocket!

My son had a car with pistons and a turbo. He got a mechanical gadget that allowed the boost to be increased to ridiculously high amounts and the car's ECU continued to provide the correct mixture until the fuel pump reached its limit. That car was also a rocket (Eagle Talon 4W-drive)!
:lol: :lol:
 
thx

im only making the boost go up about 3 pounds so im not too worried about detonation, plus if you run a higher octane it reduces the chances so im going to run premium which is usually 97.

Im not using it for an RX7 though, its for a chrysler 2.2 motor. But the map sensor has the same voltages so it should still work for me :)

I have mine setup so that my wastegate doesnt get any pressure until a certain point and then it starts regulating, so its kind of like pinchng the line but with more control. but yeah it makes a big difference, spools up quick and gives me tons of boost. 8)
 
well i built it, which was a major pain in the butt using a prototype board, but everything seems to work well. im kind of worried about a couple solder joints that are pretty close to touching on the op-amp but hopefully it will last me for a while until i build another one and make it better.

i put another trimmer from 5v to the input and adjusted it up and when it got to the voltage i set at R3 it stopped right on the money :)

thank you guys all for your help, i couldnt have bult this thing without it. i think im going to get into electronics as a hobby, i had a fun time making the voltage clamp. and i learned alot so thats cool 8)
 
Ok i put it on my car and the 12v keeps zapping the IC. So i ran a simulation with the 12v being 5v instead, and everything worked good in the simulation. But when i built it the output is always around .1v higher than the input :( I think its making the engine run too rich, so is their any way i could add a resistor or something to the output to make it the same as the input???

The clamping action is working great though :)
 
ThirtyTwo said:
Ok i put it on my car and the 12v keeps zapping the IC. So i ran a simulation with the 12v being 5v instead, and everything worked good in the simulation. But when i built it the output is always around .1v higher than the input :( I think its making the engine run too rich, so is their any way i could add a resistor or something to the output to make it the same as the input???

The clamping action is working great though :)
LM358 output won't go close enough to the positive rail to work with vcc=5V. If you are getting it to go to 4.7v, and you are getting +100mv offset, you probably have a bad op amp. LM358 works up to 32V. If your car battery is killing it, it's probably from huge spikes on the +12V wiring. Add a series 470 ohm, 1/8W (or higher wattage) resistor between the battery and the op amp vcc pin. from the op amp vcc pin to GND, add an 8.2V zener in parallel with a 100uF electrolytic cap and a 100nF ceramic cap. This should clamp and filter out the transients, and also prevent oscillations, which op amps are prone to if the power supply is not decoupled.
 
could the op-amp be zapped by too high an amperage? The 12v on my car is at 6AMPS. When i was testing it on a computer power supply everything worked fine, but the 12v amps were really low
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top