MPY634KP error

Hi Ronie,

I apologize because I do not quite understand what you are describing regarding random fluctuations. If a fluctuation takes place in an MPY634 X-input level the denominator term is affected, and if it occurs on a Z-input the numerator is affected. Either change should be reflected in the divider transfer function and the output voltage would change relative to that function.

You haven't mentioned how you constructed your MPY634 divider circuit. It may be that the random fluctuations are something that are best avoided. The MYP634 should be assembled on a PC board, with proper power supply decoupling and if necessary to keep noise out placed in a shielded enclosure. Also, do pay attention to the Frequency Response as a Divider information graph on data sheet page 4.

Trimming is often applied to the divider's Y1 input. Its effect simply sums with the divider term The multiplier's optional offset trim circuit shown in data sheet figure 2 is a suitable arrangement for the trim circuit. This same trimming circuit can be used for an X, or Z, input but then affects the denominator or numerator term, respectively, as described by the divider transfer function.

Regards, Thomas

PA - Linear Applications Engineering   

Actually what is happening that suppose 2/4 is 0.5 volt but lets say its coming 0.54 volt means 0.04 volts error. Now if voltage is increasing in both num and den say 3/6 then output is 0.59 volts means 0.09 volts error. Thats what I am saying that this error is linearly increasing with voltage increasing. Note that MPY634 is driven by + and - 12 volts supply. I have used datasheet divider circuit with Y1 connected to ground.

Hi Ronnie,

Okay, I comprehend what you are describing. The MPY634 divider operation section does discuss the fundamental accuracy in that mode, and it is evident from the Electrical Specifications that the typical error can be rather high. The Electrical Specifications for the divider mode indicates the total errors are dependent on the applied and X and Z voltage levels. The linear error response you are observing may be the natural errors. Do note that the MPY634 is trimmed with +/-15 V supplies, and operating with a different level supplies although okay may result in higher errors.

Please provide your circuit schematic and details of your MPY634 circuit board layout. Also, what precision sources are you using for numerator and denominator input voltages?

Regards, Thomas

PA - Linear Applications Engineering

I am having X1 DC input from OP27 opamp within 5 volts, X2 is ground for denominator......Z2 from another OP27 and Z1 is ground.....Y2 is connected to o/p as feedback and Y1 is ground.....Scale factor 4 and 3,5,9,13 are NC

Hi Ronnie,

We obtained an MPY634 and set it up in a bench test circuit as a divider with a scale factor of 4. The input voltages were stepped through various levels, but the numerator/denominator ratio was maintained at 0.5. Precision power supplies were used to provide the input levels and set were with a DVM.

The results indicate that the output error deviated from ideal up and down and didn't follow any particular pattern. You can see the results in the attached MPY634 Excel file. No two devices are going to exhibit the same output for the same input levels. It appears the device you tested behaved in its own individual manner and produced the output result you observed.

We found that setting the scale factor using the calculated R_SF resistor resulted in an error of about 17 %. The applications section information states that the equation will put the resistor within about 25 % of the expected SF. Adjusting the R_SF value can bring the SF substantially closer to the desired value. This is evident in the Excel file graph were R_SF was set for an SF of 4 using a 5 k-ohm potentiometer.

Regards, Thomas

PA- Linear Applications Engineering