TIDA-01513: TIDA-01513: Automotive High-Voltage and Isolation Leakage Measurements

Hello Vedran,

Greetings,

TIDA-01513 has an flowchart to identify the type of error and support the error calculations for different type of errors.

Calculations of leakage current and isolation resistance are fairly estimated and measured if we have an single point of failure.

Is 10Meg ohm considered as failure or normal condition? I think is more of normal condition, but i am waiting for your feedback.

If we consider 500K as single point failure then calculation of the same will become fairly easy with accuracy with four different modes of K1 and K2.

If we consider the dual point failures: 500k at positive terminal, 10Meg at negative terminal, Accurate estimation of the isolation resistance will be complex.

Thanks

Regards

Rama

Hello Rama,

Thanks for your prompt response.

Yes, you are right, 10Meg ohm is considered as a normal condition and 500K is considered as a failure. But the problem is that this 10Meg resistance will influence the accurate measurements of isolation resistance failure because we will always have some resistance between High voltage terminals and chassis and it can range from 1Meg, 10Meg or bigger.

I put the equivalent circuit of the situation with 500K and 10Meg resistance in the picture below. With 10Meg resistance unconnected accuracy is quite good, but when I connect 10Meg resistor it starts to influence the accuracy of isolation failure resistance measurement.

Please correct me if I’m doing something wrong.

Thanks.

Best regards,

Vedran

Hello Vedran,

Your understanding is correct.

This issue can be solved by measuring and calculations systematically. As per TIDA-01513 we have 3 measurements.
1) Pack Voltage
2) ISO POS
3) ISO NEG
We have two switches S1 and S2. Pack voltage might not change in all combinations. But ISO POS and ISO NEG will change in every combination of switch based on type of error.

Residual error due parasitic components can also be observed/calculated by doing step by step analysis. Residual Isolation resistance 10Meg on Battery Positive terminal can be measured at ISO NEG when only S2 is closed. Same from Battery Negative terminal can be measured when only S1 is closed.

If we use precision OPAMPS like OPA2348-Q1 we can measure ISO POS and ISO NEG very accurately, it will enable in solving equations of parasitic resistances first and measure the internal errors. In Case Parasitic resistance is too low, it is a sign that we have already seen an isolation error in the system.

Thanks in Advance...
Best Regards
Rama Kambham

Hello Rama,

If I understand you correctly, if I know the resistance between negative terminal and chassis (in this case it is 10MEG), I can get the accurate measurements of isolation breakage on the positive terminal (in this case it is 500k).

But what if this resistance between negative terminal and chassis is not a constant (it our case it can be between 1Meg, 10Meg or bigger and it shouldn't be considered as a fault), would it be possible to do accurate calculations of isolation resistance between positive terminal and chassis, and if it would, can you please tell me what equations you are using to calculate this?

I’m using this equations in the example for calculating isolation breakage at the positive terminal (schematic in the picture below):

K2 is closed, K1 is open, VBAT=400V, VREF=2.5V, R34+R37+R38+R39+R40+R41=1.207Mohm, R33=5.11kohm

I_leak=(ISO_NEG-VREF)/R33

Riso_pos=(VBAT+2.5-I_leak*(R34+R37+R38+R39+R40+R41))/I_leak

When using the OPA2348-Q1 for measuring ISO_NEG if there is an isolation resistance of 500kohms between positive terminal and chassis and the isolation resistance between negative terminal and chassis is neglected the accuracy of isolation resistance between positive terminal and chassis will be 0.00616% with Riso_pos=499969 Ohms.

In case if there is an isolation resistance of 500kohms between positive terminal and chassis and isolation resistance of 10Mohms between negative terminal and chassis, the accuracy with the equations above will be 11.95% with Riso_pos=559767 ohms.

So if this resistance between negative terminal and chassis is lower than 10Mohms, the accuracy error of isolation resistance between positive terminal and chassis would be even bigger. In our case everything above 1Meg should not be considered as fault, but the problem is that it is influencing the measurements.

Is this the right method for calculating the isolation resistance or I am doing something wrong?

I appreciate your help with this.

Thanks.

Best regards,

Vedran Ivković