DMM use on 3.3V logic circuits

[sma]

I have been given a task to evaluate a circuit card which uses 3.3Vdc and lower powered devices.  How can you perform a resistance test or similar on the supply rails of these devices to detect shorts to ground before applying power.

Given the output voltage of most DMM exceeds the device rail limits on high range measurements how can I prevent damage to the devices.

[JustinP]

Hello,

Depending on your input impedance of the device you may need to add a low reverse leakage current zener diode pair across the output of the DMM.  Each zener diode's reverse breakdown voltage should be the desired maximum voltage developed across the device and should not be less than the maximum voltage of the range.  If you look at page 4 of this manual, for the NI-4070/4072 DMM then as long as your input impedance is lower than 100 kOhm then you will not exceed this 3.3 V limit.  To minimize additional accuracy error, NI recommends using zener diodes with reverse leakage currents less than 5uA for the 100 and 1000 Ohm range, less than 500nA for the 10K range, and less than 50nA for the 100k range. The attached picture shows the arrangement of the zener diodes. 

[Now_With_Underscores]

Hey SMA!

For resistance measurements, our DMM sources a small test current and measures the resultant voltage.  You can find the specific sourced current in the DMM specification sheet, but for example, in the 1,000 ohm range we source 1mA.  Assuming your DUT's input resistance is less than the range (1,000 ohms), the maximum voltage developed across your DUT is given by the 'Max Test Voltage' specification (1V in 1,000 Ohm range).  However, if your DUT's input resistance exceeds the range (1,000 Ohms), then the voltage developed across your DUT will exceed the 'Max Test Voltage' specification linearly until we level off around 12V.  Hence Justin's recommendation to add an external snubber circuit.  The Zener diode circuit is a simple method to do this, but as Justin mentioned you'll need to make sure the reverse leakage current is really low or you'll affect the accuracy of the measurement.

[sma]

The device in question is a circuit card assembly powered by a sigle 3.3V power supply.  The circuit card contains 3 additional regulators to provide 2.5Vdc, 1.8Vdc and 1.5Vdc.

The measurement points will be on the output of each respective requlator so you are suggesting that I will need to ensure that the DMM output voltage does not exceed these levels to prevent damge to the devices connected on these rails or will the limited output current of the DMM prevent this from occuring.

Given this is a standard technique on TTL and CMOS (5V) logic circuits I really need to confirm if we can inadvertently damge these new technology devices.

[JustinP]

Hello,

So it sounds like at most you want 3.3 V present on the components.  The regulators reduce that 3.3 V to 2.5 V, 1.8 V, and 1.5 V.  So you can select a Zener diode pair to prevent voltages above 3.3 V from being introduced to the circuit.  As mentioned before, the voltage will only exceed the maximum test voltage if the resistance exceed the largest value for that given range. 

Thank you,

Justin

[Now_With_Underscores · Accepted solution]

Hey SMA,

What is the output impedance of the three regulators?  Typically, output impedance from power supplies is quite low (<<100 Ohms), so we shouldn't ever reach the condition where our resistance exceeds the range while the DMM is connected to the DUT, and thus the voltage should never exceed TTL spec. When connected to the power supply, I imagine you intend to perform an offset compensated ohms measurement?  If not, let us know what you're measuring more specifically and we can look into that use case.

If you're switching the signals and there's a chance the DMM will float while in the resistance mode, then when you switch to your DUTs, the excess charge held on the output capacitance of the DMM will temporarily cause a voltage spike up to the ~12V until your circuit has absorbed the charge.  I doubt this will affect power supply rails due to the small capacitance and low resistance of the DUTs, but if you are concerned that the DMM will float between tests - causing charge injection on connect - then adding a snubber circuit to the DMM leads will prevent damage to your components. 

[Now_With_Underscores]

Hey SMA,

I see you marked my previous post as solved. To satisfy my own curiosity, can you let us know what your plan of action is?

[sma]

My plan is to do the following:

1.  The test platform "LM-STAR" performance details are a bit scarce at this stage so I will include suitable snubber circuits within the interface device to prevent the ouput of the DMM exceeding the required 3.3, 1.8 and 1.5V limits.  To allow for instances where switching may occur and the DMM is configured for resistance measurements the snubber will be connected before the paths to the DMM inputs and will remain in place untill all measurements are complete in this section of the test.  The prime aim of these tests are to look for a short, however the specification of an upper limit will ensure correct path connection during testing so again the snubber is needed.

2.  You will be interested to know that I contacted Zilog technical support and they reported that they have no record of this being an issue, however this assumes the testing of a single device that you know is correctly connected for testing.