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.

Problem with measured output voltage( 0 to 3 KV )

Status
Not open for further replies.

tictac

Full Member level 5
Joined
Oct 22, 2006
Messages
297
Helped
3
Reputation
6
Reaction score
3
Trophy points
1,298
Location
Polland
Activity points
3,572
Hi
I want to measuring 0v to 3000 volt with atmega32. I use divider resistor for adjusting the high voltage betwean 0 and 5 volt.
Rup=5 M-ohm , 30 watt
I can easily measuring the voltage below than 2400 volt with 3 volt accuracy. but for the voltage that higher than 2400 volt I have some problems.
for the first minute of showing the voltage with the mega32 it has a good accuracy but after a while the measured voltage goes up slowly !!! I think its because of resistors that is warm. what can I do to prevent the measured voltage goes up ?

Regards
 

Do you know a resistor T.C. specification? Did you check the divider output without connecting the uP? Can you know, that the HV supply is actually constant?
 
  • Like
Reactions: tictac

    tictac

    Points: 2
    Helpful Answer Positive Rating
One possible reason is the input resistance & capacitance of the Atmega ADC - if the HV divider has too high resistance you can experience unstable behaviour
 
  • Like
Reactions: tictac

    tictac

    Points: 2
    Helpful Answer Positive Rating
Hi FvM
I know about T.C a little . I dont use it. but I use different coefficient for converting 0-5 volt to 0 to 3000v.
for example (0<vin<5)
C code:
if (vin>0 && vin<=1)
vout =vin*630;
else if(vin>1 && vin<=1.5)
vout=vin*628;
...
...
...

so I can display the voltage with good accuracy. but after a while it increases slowly for output voltage higher than 2400 volt.do you mean that I can solve this problem with using T.C of resistors?

luben111 ,

if the HV divider has too high resistance you can experience unstable behaviour
I dont think that the resistance is too high .its about 5400k-ohm. because I can measure the output voltage that below 2400 volts without changing at the value. and it has a good accuracy.

Regards
 

if the HV divider has too high resistance you can experience unstable behaviour
I agree with this. Maybe the 5400k is too high for ADC reading. In one of the books, I had read that max allowable resistance of PIC ADC to ground was around 20k, so, it should be similar in the case of AVR. I think you can decrease the operating current by using higher value resistors (use 2W or higher power resistors for voltage safety) to avoid overheating of resistors, then use an op-amp based voltage follower or voltage buffer circuit. This might help if the behaviour is due to the divider resistance.

Before that, check that the high voltage supply is stable and actually not changing - the change shown by your AVR on the display. Also, as FvM has stated, check the output of the voltage divider with a voltmeter without the uC connected.

Hope this helps.
Tahmid.
 
  • Like
Reactions: tictac

    tictac

    Points: 2
    Helpful Answer Positive Rating
Hi,
Belive me, in normal conditions resistors above 1M should be treated with extreme care. Small contaminations with little moisture yield resistance under 1M so for me it looks quite possible you get swings of the measured signals! Of cource you can put 10M or 20M resistors but by paying extra care for the design (cascaded resistors, cuts in PCB under the resistors, ultrasonic cleaning of the PCB, usage of bigger footprint components, covering the resistors with epoxy, etc.)
 
  • Like
Reactions: tictac

    tictac

    Points: 2
    Helpful Answer Positive Rating
Tahmid,
So,I will reduce the value of resistor to the 3M-ohm. the total power of those resistors is about 30 watt.
at the first of my test,I used (2X2.2 M-ohm,7watt). As I say before at the voltage that higher than 2400 volt the voltage that voltmeter shows goes down (just 5 volt) but the sampled voltage goes up after a while. I check the resistors temperature.they were warm. when I use one Fan for cooling the resistors ,the sampled voltage goes down and reach to the real value.so I decided to increase the resistors power.
now I will do proper your saying.

luben111 ,

Small contaminations with little moisture yield resistance under 1M so for me it looks quite possible you get swings of the measured signals!
at the output voltage below than 2400 volts the sampled voltage is fix and didnt change .can you describe it a little more,please?
 

so I decided to increase the resistors power
According to your post, it has been more than factor 15 above the actual load before. So it seems to be mainly a problem of a huge resistor T.C. (temperature coefficient). Oversizing the resistor is basically good in precision measurement, but selecting a precise resistor type seems more important in this case. Because good high resistance values are hard to get, you may want to use multiple 1 Mohm metal film resistors instead.
 
  • Like
Reactions: tictac

    tictac

    Points: 2
    Helpful Answer Positive Rating
Hi
You said that your resistive divider is >4 MOhm. This means that small contaminations could result in effective change of this resistance (despite that the resistor value will stay constant the leackages on PCB will result is change of whole resistance).
Once you apply the high voltage (voltage does not change and is stable) on the divider and if there is contamination an ionic process will occur. The ionic density will increase with the time so if you experience that the measured voltage increases with the time (high resistor in the divider decreases) one possible explanation is the contaminations.
In case the high voltage has also high frequency things will become even worse with contaminations. One simple solution to decrease the effect of the contaminations is to use cascaded resistors. For example if you have 4 MOhm resistor you should change it with 4x 1MOhm resistors in serial - the effect of contaminations will be reduced 4 times.
 
  • Like
Reactions: tictac

    tictac

    Points: 2
    Helpful Answer Positive Rating
Dear FvM and luben111
I apply your good instruction and will report my experience.

Regards
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top