Anyone have profound understanding about ICL7106/7107?

[WalkingBeijing]

icl7106 power supply

My ICL7106 based LCD voltage-meter has 1 count jitter problem.
This meter is mounted inside a bentchtop DC 30V/3A power supply and is used to measure the output voltage of this power supply. The actual output of this power supply is stable and this problem is from the voltage-meter. This linear style bentchtop power supply has a hulking transformer. And this power supply has a closed METAL housing that have connection to the 'Protective Earth'.
But when I connect the 'output Negative' and the 'PE', display will turn very stable....
Anyone could give me some advices.
thanks....

 

[laktronics]

all about icl7106

Hi,
1. If possible use a 0.1uF,high voltage capacitor between -ve lead and power earth.

Do you really need the leads to be floating?, it may be sufficient to ground any of the two leads to remove the noise.

2. If the ADC clock has an adjustment, adjust it to make the conversion time an exact multiple ( as closely as possible) of mains period. But when mains frequency varies this problem may reappear.

To me one count variation is well within the specification of the meter. It appears your transformer has some leakage to the secondary.

Regards,
Laktronics

 

[Audioguru]

7107 common power supply

I think the ICL7106, ICL7107 cannot measure their own power supply voltage due to a common mode input spec. The ICL7107 also needs a negative supply separate from what it is measuring.

 

[FvM]

probleme icl7106

You didn't tell how the 7106 is supplied and how the aux supply is connected to measurement signal. Thus I agree, it may be an issue of exceeding´7106 common mode voltage range.

 

[Audioguru]

floating input + icl7107

The ICL7106 is normally powered from a little 9V battery.
The ICL7107 needs to light up to 28 LED segments at 8mA each which is 224mA which is too high for a battery so use a power supply.

The datasheet shows how to connect the inputs.

 

[FvM]

7107 datasheet

The ICL7106 is normally powered from a little 9V battery

Yes, in a portable instrument, but very unlikely in a desktop power supply!

 

[WalkingBeijing]

how to power icl7106

Thanks all of you...
the power to this voltage-meter is independent to the Measured voltage. It's from a LM7809 regulator driven by a secondary set of the transformer. It means that one transformer has 2 secondary set, so one set is for the bentchtop DC power supply and another is for voltage-meter.
Circuit is a typical one. DIP-40 7106's pin 30(in -), pin 32(anolog common) and pin 35(ref -) are tied together. Pin 30(in -) is tied to "output negative" of bentchtop power supply. And I use 2V full scale.

with best regards.

Added after 46 minutes:

Hi,
1. If possible use a 0.1uF,high voltage capacitor between -ve lead and power earth.

Do you really need the leads to be floating?, it may be sufficient to ground any of the two leads to remove the noise.

2. If the ADC clock has an adjustment, adjust it to make the conversion time an exact multiple ( as closely as possible) of mains period. But when mains frequency varies this problem may reappear.

To me one count variation is well within the specification of the meter. It appears your transformer has some leakage to the secondary.

Regards,
Laktronics

Thanks your opinion.

About the exact multiple of the mains period I still have many questions.
According to the Intersil ,Maxim or Microchip's 7106/7107 datasheet, they will tell you that 100K resistor and 100pF capacitor is good choice to achieve 48KHz clock frequency which could have a strong rejection to the 60Hz noise. But in fact these simple 2 passive components together with 7106's internal circuit couldn't insure a exact multiple of mains frequency at all. Just little tolerance in component's value will make the clock frequency depart from the multiple of the mains frequency.
Oh!my god, I think even though use potentiometer and adjustable capacitor, It's hard to insure the exact multiple of the mains frequency in long time operation. How could they achieve 48KHz with the equation f=0.45/RC ?

and another question Is 47.940 KHz or 48.060 KHz the exact multiple of 60Hz mains frequency?

 

[FvM]

how measure capacitors with 7106

A basic problem when supplying unipolar 9V to 7106 could be in common mode currents, that may be coupled to the supply and possibly overload the input circuit. So it may be a solution to add a bypass capacitor between 9V- and Vin- (if not already present in the circuit). But there may be other interference pathes in the design. Regarding clock frequency for maximum 60 Hz supression, the integration interval is fc/4000, so 48 kHz would generate 1/12 s which exact mupltiply of 60 Hz period. I think, 40 kHz is another common value for simultaneous 50/60 Hz supression.