Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Why output oscillates at low volatage?

Status
Not open for further replies.
R

rishad babu

Member level 2
Joined
Oct 6, 2015
Messages
42
Helped
0
Reputation
0
Reaction score
0
Trophy points
6
Activity points
323
Am here with a problem regarding high and low ac mains voltage cut off circuit.i have attached the circuit diagram.the high voltage part is working almost good.but the low voltage part isnt working good.if i set the low voltage cut off point as,for example 190,the output starts oscillating when the voltage is lesser than 190.i mean the output is on and off alternatively so fast..the output is cut off only when the voltage falls almost below 170v.what could be the reason for that...?

The output from the high and low voltage cut off circuit is fed to the RD7 of the pic as an input which switch off and on its output(RD5) according to the RD7 input value.Under normal condition RD7 would be 1 and under low or high voltage conditions RD7 would be 0. IMG-20151217-WA0000.jpg
 

You are looking to design a Window Comparator which requires two Comparators with wired OR output low for detecting UV or OV.
This requires low ripple or noise inputs, hysteresis, supply filtering and a good low noise reference voltage for comparing the scaled inputs.

Define your spec first using RMS converted to peak.

  1. UV detect level
  2. UV recovery or hysteresis and response time. ( i.e. 1 cycle. or 20ms or more)
  3. OV detect level
  4. OV recovery level or hysteresis and response time.
  5. Latched condition , option.

However you have a bunch of transistors with potential positive feedback from switching currents and probably poor supply decoupling and stray wire coupling in a breadboard with lots of opportunity for metastability or oscillation.

Use good supply decoupling, good layout and use two comparators with hysteresis instead.
Ensure the input signal has good rejection of 100 Hz ripple and harmonics and adequate hysteresis,

Normally Load R and bulk filter C give RC=8 for 10% ripple at full load. Additional filtering is needed for the comparators.
 
Last edited:

1. The UV detect level is 180v
2. UV recovery is 190v
3.OV detect level is 260v
4. OV recovery is 250v
5. Am sorry i didnt get what is the latched condition.

If the supply decoupling is poor,how should i improve it...?
And if my layout is poor,will using comparators make it better...?
Is using bunch of transistors an inefficient method...?
 

Cascaded transistors with Re=0 have high gain and crosstalk can cause instability with layout without good ground plane and RC filtering on inputs and supply.
Forget this approach, the switching thresholds are far too inaccurate and hard to control precise.thresholds and can drift >>5% with temperature and you expect ~1% accuracy, I assume. ( 0.65 +/-0.05 for example swing on Vbe = +/- 8%)

Comparators are high gain high impedance, precision differential threshold with excellent supply noise rejection.

Some of these examples will give you an idea but may not be perfect for your application.
https://www.google.ca/search?q=wind...w=1467&bih=1062&dpr=0.9#imgrc=YUqfU-aiHZ4EJM:

  1. You must convert RMS to Peak using sq.rt(2) multiplier for AC.
  2. Then scale down to low voltage with a precision resistors, < 1%.
  3. You need a precision bandgap reference voltage for the comparators
  4. e.g. 255Vac (mean OV) is ~360Vdc and 185Vac (mean UV) ~261Vdc, so if we scale that /100 we get 3.60V and 1.85V to detect with 4% input hysteresis desired.
  5. You can also use the Vref to subtract an offset from the input voltage to improve accuracy. (optional)


There are many ways.to do this.

But you must expand your spec to define AC inputs , DC outputs then intermediate scaled levels,to perform the task as I have suggested above.

Since I do not know your DC supply options and Vref options, you can do this to meet other requirements.


With Hysteresis of 10V at 255V average , the hysteresis and is only 4%. which is tight.
If the input and output range the same then the % positive feedback would also be 4%, but if not then it must be scaled. e.g. if input is 0 to 3.6Vdc for 0~255Vac and output comparator is 0~5V then 4% input hysteresis translates to about 3.6 / 5V *4%= 2.9% feedback ratio to non-inverting input..
But of 130~260Vac input translates to 0 to 5Vdc input then you need an offset included and may require an Op Amp to condition the signal for gain and offset with a Vref. This is just an idea, maybe not what you had in mind. but certainly more precise and stable.
 
Last edited:

Hi,

Now:
* you measure AC peak voltage. Noise on the AC line cases large measurement errors. The usual flat top of a main AC sine causes errors.
* you have no temperature compensation, therefore I expect a drift of about -1V / °C.
* the AC filtering is slow and poor
* there are no hysteresis
* high part count

-->
I recommend:
* use a more "rectified average" measurement method for reliable measurement
* use higher order filters to attenuate mains frequencies (and multiples), while the same time get faster response. Adjust cutoff frequency.
* like SunnySkyguy said: use a reliable reference and use comparators. This gives lower part count and better precision.
* use logic to combine both outputs, add delays to adjust activate and release times. ( 2 x R, 1 x C, 1 x D and a Schmitt trigger)

Klaus
 

For 260Vac am getting a rectified DC output as 19.8Vdc and for 180Vac am getting 13.3Vdc..am using a 15v stepdown transformer.then why should i use a /100 scaling...???

- - - Updated - - -

I had also tried with LM393.but the result was same.

- - - Updated - - -

If the AC filtering is poor,how could i make it better..?will it be ok by using a high value capacitor instead of 470uf....?

- - - Updated - - -

I dont know if u guys got what exactly mu problem is..let me make it a little more clear.
If i set the UV cut off as 190Vac,then when the voltage is between 190 and 170 the output from this circuit(which is fed to RD7) oscillates from 0 to 5Vdc,which in turn makes the pic output(RD5) also oscillates from 0 to 5Vdc.the oscillation is very fast.
 

Hi,

Yes, i think we understood.

As said before: It as a problem of filtering the AC voltage and a problem of missing hysteresis.

the oscillation is very fast.
Click to expand...
Two cases:
1) it is double the line frequency, then it is caused by poor AC filtering
2) it is even higher frequency, then it is caused by local oscillation in your circuit.

***

If you give the value instead of "very fast", then we don´t need to "assume". Values are more informative and easy for us, the thread will be shorter and yout get faster and more precise response.

Klaus
 

I cant give the value..it oscillates in such a speed..
1. If its due to poor ac filtering,help me how to clear it.
2.if its due to oscillation of my circuit,how could i clear it.
I guess the Problem of hysterisis can be sovled by using a comparator instead..
 

Actually,i couldnt find the value. When a relay is connected to it, it almost oscillates 50 times in a second. Or even more than it.
 

Hi,

I wonder why 50 Hz is difficult to measure.. A scope, a frequency counter, a soundcard, a dim with frequency measurement or similar should do.
If you don't have any, then you should consider to buy something like this. You will need this to build electronics.

Did you add hysteresis?

Use two 1M resistors
One 1M resistor from relay_out back to the point where the 10k pot meets the 470R resistor in the under voltage detect circuit.
And the other from relay_out back to the same place in the overvoltage circuit.

****

We don't see how, and from where you generate the 5V.
Also we don't see where and how the relay is connected.
Is there anything else connected at the relay contacts?

Klaus
 

We do have a plan to buy such equipments to measure frequency.we are looking for that.but cant be made soon..
I havent added hysterisis.let me work on it as u recommended now.
But let me ask.is it better to use this circuit or should i go for comparators..??

- - - Updated - - -

5v is generated using 7805.
Relay is supplied from 7812 and its driven by a transistor whose base is supplied from the pic RD5.
Just a flywheel diode is connected to the relay coil.nothing else is connected to the relay
 

Hi,

But let me ask.is it better to use this circuit or should i go for comparators..??
Click to expand...
Replying to your posts is not usefull if you don´t read it.

Please read SunnySkyguy´s replies and also my post#5 with it´s recommendations. --> Both recommend the use of comparators.

5v is generated using 7805.
Relay is supplied from 7812 and its driven by a transistor whose base is supplied from the pic RD5.
Just a flywheel diode is connected to the relay coil.nothing else is connected to the relay
Click to expand...
If you want help, --> please show us your complete circuit.

Klaus

- - - Updated - - -

Added:

Now you tell us the relay is driven by a PIC.
I always thought the schematic "RO" is the relay drive output.

Doesn´t the PIC has a built in ADC or comparator? So you don´t need the whole circuit.


***

Please give complete information. Not piece by piece. This is really annoying.

Klaus
 

Am reading every posts so clearly.i asked again about comparators just because i needed a clear cut conclusion..anyhow,am sorry for that.
In the beginning itself i had told that the output from this schematic is supplied to RD7 of the pic which accordingly controls its output RD5..n its not RO..its RD7..i had told the same thing again also assuming that you guys didnt get exactly what my problem is..
Am giving informations piece by piece because i doesnt actually know what all are the required informations here..am sorry for that too..i thought info about 5v generation and relay are not significant here since oscillations is present even without relay.oscillation is there in the output...
The pic do has ADC port..but i never thought of that.thats a very usefull info..thanks a lot..

- - - Updated - - -

I will post my whole circuit

- - - Updated - - -

Posting photos is a lil difficult one since am using mobile.no options in mobile version.i will post it soon..
 

I have tried with the pic adc pins...but the problem is still present..i searched n found that BOR is something related with low voltage..but i couldnt understand it properly..
 

Hi,

BOR is for power supply monitoring and has nothing to do with ADC.

What problem is still present?
Still the piece by piece information.

Please concentrate on the problem.
Where exactely do you see the problem? Describe the problem very detailed.
Give pictures, schematic, PCB layout, code....

Klaus
 

remember output of mains rectifier has ripple, so that needs to be sorted.

You should put a potential divider.
Then heavily filter it with caps
Then you have a dc voltage which represent the value of the post diode bridge dc bus.
then you put comparator with hysteresis on that to indicate if it is over voltage or undervoltage
so that's one comparator for indicate overvoltage
and one comparator for indicate undervoltage
You could use "window comparator", but why complicate things, so as I depict, please use the two comparators as discussed..
 

Hi,

Here's the circuit which i am using right now.

IMG-20151222-WA0015.jpg
 

Hi,

things clearing up.

Now we see that the relay is supplied from exactely the voltage you use for voltage measurement.

So it is quite obvious, that when the relay is activate (current flow) then voltage at the capacitor drops.

--> in the circuit of post#1 there is no other connection to the monitored voltage. So we assumend there is no influence.
Now we see that the whole circuit is supplied by this voltage.

Do you now understand why we ask for complete informations / complete schematics?

****
Back to the problem.
How to avoid this:
1) use seperate transformers. One for voltage measurement, the other for power supply.
2) use a mechanically latching relay, or use a bistable relay...they need only a short pulse to switch, this avoids continous voltage drop. (only short current pulse)
3) measure the current from the bulk capacitros and calculate the voltage drop caused by the current to get a "compensated" voltage information.

Klaus
 
Last edited:

SunnySkyguy said:
Cascaded transistors with Re=0 have high gain and crosstalk can cause instability with layout without good ground plane and RC filtering on inputs and supply.
Forget this approach, the switching thresholds are far too inaccurate and hard to control precise.thresholds and can drift >>5% with temperature and you expect ~1% accuracy, I assume. ( 0.65 +/-0.05 for example swing on Vbe = +/- 8%)

Comparators are high gain high impedance, precision differential threshold with excellent supply noise rejection.

Some of these examples will give you an idea but may not be perfect for your application.
https://www.google.ca/search?q=wind...w=1467&bih=1062&dpr=0.9#imgrc=YUqfU-aiHZ4EJM:

  1. You must convert RMS to Peak using sq.rt(2) multiplier for AC.
  2. Then scale down to low voltage with a precision resistors, < 1%.
  3. You need a precision bandgap reference voltage for the comparators
  4. e.g. 255Vac (mean OV) is ~360Vdc and 185Vac (mean UV) ~261Vdc, so if we scale that /100 we get 3.60V and 1.85V to detect with 4% input hysteresis desired.
  5. You can also use the Vref to subtract an offset from the input voltage to improve accuracy. (optional)


There are many ways.to do this.

But you must expand your spec to define AC inputs , DC outputs then intermediate scaled levels,to perform the task as I have suggested above.

Since I do not know your DC supply options and Vref options, you can do this to meet other requirements.


With Hysteresis of 10V at 255V average , the hysteresis and is only 4%. which is tight.
If the input and output range the same then the % positive feedback would also be 4%, but if not then it must be scaled. e.g. if input is 0 to 3.6Vdc for 0~255Vac and output comparator is 0~5V then 4% input hysteresis translates to about 3.6 / 5V *4%= 2.9% feedback ratio to non-inverting input..
But of 130~260Vac input translates to 0 to 5Vdc input then you need an offset included and may require an Op Amp to condition the signal for gain and offset with a Vref. This is just an idea, maybe not what you had in mind. but certainly more precise and stable.
Click to expand...

I hope you realize none of your AC sensing design meets your requirements for accuracy and stability and that's why it oscillates, as I explained above.

Which part do you not understand?
Step 1 have good input output specs.
Step 2 verify your design against these measurable specs.
Step 3 change either specs or design until specs are met.

Eg.

Vac sense voltage vs Vin vs 12V load current
Vdc tolerance for Vin range and load range with relay on off.

Remember for 10%pp Vdc ripple RC of rectifier must be > 8RC for ripple cap.C and equivalent load R
If too small 100Hz ripple will cause noise which means it oscillates.
If too big, sense delay will be poor and cause instability, which means it oscillates.
If threshold detect changes with load due to DC supply droop, then detection error means unstable result.,
Many other causes of unstable results by poor choice of design.
 
Last edited:

As klausst suggested,i used separate transformers..n the problem is almost rectified..but a little bit remains still..
Previously,if i set the cut off as 200,the output osscilates in the range of 200 and 180 Vac.if i set it as 180Vac,then it osscates in the range of 180 to 160Vac..now the range has decreased..it oscillates only in a very small range,like 200 to 196Vac,or 180 to 176Vac etc..
Am using 230Vac 50Hz supply,12v step down transformer 1A,12v 285ohms relay.the filtering capacitor is 470uf..there is no resistor in that filtering part..so i dont get what is this RC value...???
Please talk to me considering me as a beginner in electronics.
Since am a beginner its hard for me to get what sunnyskyguy tells..am happy that he is willing to help..but i cant get what he suggests me coz am just a beginner..

- - - Updated - - -

The regulators am using are 7805 and 7812..those are high drop regulators right...?will that makes any troubles here...??
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top