Wrong Zeners from factory?

Erik Friesen

I have attached my circuit to the mcp3208. I have verified that these zeners are clamping the voltage around 2.8 - 2.9 volts. The zeners are 3.9v

Zeners D11-D13 www.fairchildsemi.com/ds/MM/MM5Z3V9.pdf

Diode D6, D9, D10 www.semiconductors.com/acrobat/datasheets/PMEG3002AEB_1.pdf

This seems odd. I have verifed the number 55 printed on the zeners.

Phil Pilgrim (PhiPi)

You probably aren't driving enough reverse current through the zeners to bring them up to their rated voltage. The ones you cited are spec'd at 5mA.

-Phil

Post corrected: "aren't", not "are".

Post Edited (Phil Pilgrim (PhiPi)) : 6/23/2009 3:37:44 PM GMT

Erik Friesen

I am not sure I am following you Phil. If they were up to their rated voltage, it should show that upon testing. The max output from my isolation amps is 15V, which calculates to 5 ma through my 3K resistor. Are you suggesting that I will get a 1 volt change at the extreme ends of the rating? It doesn't seem like I am at the extreme here, but I will admit I don't know the meaning of every piece of data on the datasheet.

Leon

What are the zeners for, exactly? What value are the resistors, and what are the input voltages?

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Phil Pilgrim (PhiPi)

Here is a typical set of I/V curves for zeners of various reverse voltage ratings:



A "perfect" zener diode would have a vertical line at each rated voltage. But, as you can see, real devices are far from perfect.

-Phil

Tracy Allen

Erik,

As you see from the graphs Phil just posted, low voltage zeners have a particularly "soft" knee, so at microamp current levels the voltage will be the 2.8 volts you observed, and conversely, even when the input voltage is low, these diodes will be drawing a small current and dropping a voltage (=error) across the resistors R27-R29.

You are using these for input protection. A more usual scheme is to tie the (+) side diodes to the Vdd power supply, so that fault currents are shunted there when the input voltage exceeds Vdd by one diode drop, same as you have on your schematic for the (-) side in relation to Vss.

I use a protection array like the SP721. It is a single chip that has six protection circuits for both (+) and (-) fault protection.

How are the isolation amplifiers working out?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Erik Friesen

These are not so much for input protection as to keep the 15v possible voltage from hitting the internal protection diodes of the mcp3208 and skewing the signals. That is part of the reason why I did not go the diode to vdd route in the beginning. I suppose that some schottkeys would do the trick though.

@leon .The resistor coming from the isolation amp is a 3k resistor. I am expecting input voltages from 0 - 3 volts.

@Tracey . I am happy with these op amps. My first test was with these at 100 gain, and they are quite stable. So far it looks like they are going to work out well with those galvanic sensors, and they seem to be more stable than the competitors boxes input.

Thanks all, I had not realized low v zeners had that kind of knee.

Tracy Allen

It looks like the ADC is powered from 5 volts, so a 4.7 or 5.1 volt zener might provide adequate shunting without loading the 0 to 3 volt signal.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Erik Friesen

In the back of my mind, I guess I was worried about the fact that the reference is 4.096V. I suppose that is of no consequence?

Tracy Allen

No, the reference does not matter. The substrate diodes will start conducting at about 5.3 V when the chip powered with Vdd=5 V.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com