noise problem in voltage booster circuit

[john111smith]

hi all

my voltage booster circuit {

http://john111smith.googlepages.com/vbstr.png}

shows the following noise levels:
A: <+-50mv
B:~ +-200mv at 67v output
D:~ 0

OpAmp: OP07 or LM741

How can I reduce noise level at point B to about +-50mv?

 

[Skogsgurra]

First, I think that your circuit is designed a bit off normal beaten tracks. Using an opamp as a controller without any stabilizing feedback is usually not possible.

I also have a problem with your supply being +100 V and your maximum output the same voltage. That will not be possible with a 150 kohms source resistance that is loaded with around 1 Mohm.

I would change topology and resistor values. Use an opamp with PI or PID feed-back. Use a PNP transistor with level-shifting.

The reason for your ripple is probably either internal noise in opamp and transistor or marginal stability (most likely) or hum pick-up in long conductors going to inverting and/or non-inverting inputs of opamp. Possibly a little of each.

Scrap the whole thing, get a good reference design and go from there.

Gunnar Englund

http://www.gke.org

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

 

[john111smith]

Thanks Gunnar

+100V supply is a nominal label and it is about 126V

Could you please send me a good reference design.

 

[OperaHouse]

You have an inductor and a capacitor in the output, got something against a capacitor on the output? You would normally sample the voltage after a filter. This is a basic shunt regulator, but I can not begin to figure out what use a power supply is for that can supply less than a quarter milliamp. This can barely supply enough current to a cheap DVM with a 1 megohm input.

 

[john111smith]

it drives a piezo - no current need

 

[OperaHouse]

I wouldn't say that, this is a capacitor. So what is your noise.... line frequency from the raw 100V, pickup on the input signal, or op amp (transistor) noise. Is it oscillating at one frequency? What frequency does this amplifier have to work up to

 

[OperaHouse]

Try a .001 to .01 uF across the gain resistance of the pot and 330K to get rid of the noise.

 

[logbook]

This is an interesting circuit. I am assuming that it is not supposed to be self-oscillating and that you are just using it as a linear amplifier for the signal applied at R43. All that follows is based on this assumption:

The forward path gain is very much larger than the opamp alone. The output stage simulates as having a gain of 40dB at LF. Your feedback factor is variable, but can be quite high (closed loop gain of 6.5 to the effective inverting input on R49). The circuit simulates as being unstable. If you reduce the loop gain the "noise" (instability) will reduce. This is most easily done by putting a series RC network across the base-emitter of Q2. 1K in series with 1µF should do the trick.

I think if you do what OperaHouse has suggested the noise/oscillation will increase.

 

[john111smith]

Thanks All

Also before there was a RC(1K,2n2) between collector and base that force circuit to oscilate, other values for C (20p to 100n) didnt produce good results.

because of low output impedance of +100V supply source, I have to use this resistances values, but if you think change of resistors, produce considerable change in noise, I will replace +100V supply.

I will try suggested configurations.

J.Smith

 

[OperaHouse]

Posting a question about how to use a fuse and a diode and there are 40 responses. An op amp and a transistor and this board falls flat on its face! First question in a month that even gets the brain cells to twitch.

Hanging a 330K on the base of a high voltage transistor is just asking for trouble. You haven't told us what your noise is, but if you disconnect the end of that resistor at the op amp and tie it to ground, that noise might still be there. Do this to define the noise source. Replace this with a base to ground resistor of 4.7K and base to op amp of 33K. If I was designing this, I would be grounding the base and driving the emitter. Well actually, using a single supply op amp like a LM324, base tied to V/3 supply with cap to ground and drive the emitter with the op amp.

I'm all for E-Star, but if you are driving a reactive load you should waste a little power. Change the 150K collector resistor to a lower value like 68K or lower. Does this circuit actually control the voltage now? My calculation indicates you can't get much over 70V as designed.

 

[john111smith]

Thanks
today I changed configuration as below but the result was not good.

remove diode, connect emiter with R10K to -15VEE, chnage 330K to 10K, connect base to -15VEE with R10K,
change R150K to 47K

only a 1nF cap between opamp output and its negative input reduced noise to 50mv (equal supply noise)

Now, I try to reduce noise of +100V supply(by the way, it is a nominal label and actualy is about 126V).
parallel 220uF, 100nF, 10nF are after bridge diode and LRC [680uH,20K,parallel 100nF,10nF,1nF] are before load.
FFT view of noise shows 100Hz peak noise harmonies.

 

[Skogsgurra]

There are many ways to skin a cat and I still think that your approach isn't one of the best.

But, you seem to be on your way to something that works, so - why not?

The fact that you have a 100 Hz ripple (Europe? Not US?) hints that you have some ripple from your rectifier left. Your "controller" has been slowed down so it can't reduce the ripple - or the ripple is entering some of the high-impedance nodes.

How much 100 Hz do you have? And, excuse me for asking that, where have you put your ground clip? How much ripple does your 126 V (the "100 V") have?

Gunnar Englund

http://www.gke.org

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

 

[OperaHouse]

By connecting the emitter to the -15 supply, you could be introducing some supply noise. I'm fond of localized feedback, but would rather see a 1K emitter to ground. It really looks like you are introducing noise into a circuit since the cap on the op amp works. Less than 20 mV of power supply noise at this gain would give 500mV on the output. Does the 1nF cap neg feedback give acceptable results? You have not indicated if the circuit has to have any frequency response. Would a cap on the collector to ground cause adverse performance in your application?

 

[john111smith]

Thank you.

 

[MikeHalloran]

Given a steady, quiet, DC input, a piezo does ... nothing, which is not always a useful behavior. So we have to assume that input D varies with time, and that you want the output to track it, pushing charge into the piezo and pulling it out again.

But some of the things that you could do to the circuit in order to make it quiet also cause it to suppress higher frequencies in the signal you are trying to transduce here.

The members who are trying so hard to help you could do a much better job if you revealed the range of frequencies that will be presented at input D.



Mike Halloran
Pembroke Pines, FL, USA