I recently read through a math article by Steve regarding using an inline 0.1Ohm resistor instead of using a current clamp to avoid probe drift etc.
Firstly, fantastic idea, and so I started to look for panel mount resistors with a high wattage. The highest I could find would be 300W 0.1Ohm, but this requires to be mounted to a panel 5000cm2 for good thermal conductivity! This 300W resistor would be able to handle about 20-25amps.
Then I got to thinking why not use the ground strap to the battery as our resistor? If we can get an accurate resistance measurement of the ground strap, can we not use this as our inline resistor? This way we don't need to purchase any parts, and also no worries about overloading resistors. The other option would be to use an inline maxi fuse, as these resistors are highly accurate (cheap too). If the current flow during an overnight recording is too much - the fuse blows.
Appreciate your advice on this topic.
In order to use the battery ground lead/strap as your resister (and why not) then we need to calculate the ground lead/strap resistance
To do this we can use starter motor current as our load and measure the volt drop across the ground lead
Using channel A connected across the ground lead and channel B to measure current, we can apply the math channel A/B to calculate the resistance (Voltage/Current = Resistance)
Here I have concentrated on the volt drop with the lead under cranking load (between the rulers)
Once we have this value (0.001329 ohms above) we can incorporate this value into our parasitic drain math channel. The following link will help here
Your connection for a parasitic drain measurement would therefore be:
Channel A across your battery ground lead (capturing voltage drop)
Math channel A/0.001329 (Voltage/Resistance = current)
It may be necessary to add low pass filtering and use Bandwidth limit (if using the 4425 or 4225 scopes)
More information can be found here
I am not sure of how accurate this would be but I will try to check and feedback ASAP
I think a custom probe could be useful here too as we would not need to wait for the scope screen buffer to complete to view the data
I hope this helps, take care......Steve
Below we have the resistance calculation for the vehicle ground lead that travels from the battery negative terminal, to the chassis and onto the engine. (One single lead)
Channel A records the voltage drop in the entire vehicle ground lead from the engine to the battery negative terminal. As you can see A/B = Voltage / Current = Resistance at 1.52 mΩ
However, if we use this resistance value during our parasitic drain measurement, we discover an error! Below we measure the vehicle parasitic drain during the shutdown stage (prior to sleep) with a TA234 current clamp (Channel B) and the math channel formula A/0.00152 (Voltage drop / resistance)
Whilst the values appear close, they are not close enough when looking at parasitic drain. If our target drain is below 80 mA (sleep mode) then the error we have here between the current clamp measurement (Channel B) and the math channel calculation is excessive!
Current clamp 168.1 mA – Math channel 147.3 mA = 20 mA error. (13% error)
20 mA could be the difference between a pass or fail result with a parasitic drain test!
The error above becomes more dramatic when the vehicle moves from shutdown to sleep
Here the current clamp suggests our parasitic drain is acceptable (43.16 mA) yet the math channel is claiming a failure (96.4 mA) The error here has increased to 55%
With the vehicle still in sleep mode, the current clamp and channel A test leads are removed from the vehicle to confirm their zero offset.
Interesting here how the current clamp has drifted up to 16.66 mA and channel A indicates 23.79 micro volts (disconnected from the vehicle)
Note how the math channel also indicates a parasitic drain value based on these residual values from the current clamp and channel A! (Between the rulers)
Incidentally, whenever measuring parasitic drain with a current clamp, it is essential to capture the zero offset (drift) at the end of your measurement by removing the clamp from the vehicle and recording the displayed value with no current flow through the clamp jaws.
Here you can qualify the final parasitic value taking into account current clamp drift. More information on parasitic drain testing and zero offset can be found here
If we now go back to image 2 above, we have to make a number of amendments to obtain the correct parasitic values based on the discovered zero offset values
The math channel displayed 96.4 mA based on channel A being at zero volts when disconnected from the vehicle.
If we recalculate, channel A displayed 146.5 microvolts - 23.79 microvolts (offset) = 122.71 microvolts
0.00012271 V / 0.00152 Ohms = 81 mA
The current clamp (Channel B) displayed 43.16 mA, assuming no drift, however the clamp had drifted to 16.66 mA, therefore our true parasitic drain is 43.16 – 16.66 = 26.5 mA
Even with these errors corrected, the accuracy of our parasitic measurement across the vehicle ground lead is incorrect by over 60%!
The math channel cannot be wrong as the value displayed is simply the result of dividing the voltage drop (channel A) by our calculated resistance value of 1.52 mΩ
What is wrong in my case was using 1.52 mΩ as the resistance value in the math channel!
If we recalculate the resistance of the vehicle ground based on the parasitic current flow through the vehicle ground lead (via the TA234) we discover a different resistance value!
Looking at channel A / channel B we have 0.0001465 V / 0.04316 A = 3.4 mΩ (between the rulers) not accounting for the zero offsets mentioned above!
We can use this value for our parasitic drain calculation rather than 1.52 mΩ
Using the math channel, A/0.003407 we now have a realistic parasitic drain value using the ground lead as our known resister.
As you can see the current clamp (Channel B) indicates 43.16 mA whilst our black math channel (A/0.003407) displays 43.01 mA.
This is much better but remember:
Our current clamp zero offset is 43.16 – 16.66 = 26.5 mA
Our Channel A zero offset is 146.5 microvolts - 23.79 microvolts = 122.71 microvolts
0.00012271 V / 0.0265 A = 4.6 mΩ accounting for zero offsets
0.00012271 V / 0.0046 Ω = 27 mA
The error now is less than 2% between the math channel and the current clamp
Now look at the vehicle ground cable resistance value when we return to the shutdown phase (between the rulers below) it appears to have decreased to 1.383 mΩ which now requires another math channel (A/0.001383) to calculate parasitic drain!
This is a real nuisance as parasitic drain will often increase/decrease over time (especially where parasitic drain is responsible for discharging our battery)
In order to graph parasitic drain whilst compensating for tiny fluctuations in ground cable resistance I have used the formulae Voltage/ (Voltage / Current).
This can be seen as the orange math channel in images 5 & 6 “A / (A / B)”
Whilst this serves as a “get out of jail” it does not work in the real world as the whole reason we are using ground lead resistance to calculate parasitic drain is to remove the need for a current clamp (other than to measure current to calculate resistance)
Given the vehicle ground lead displays minuscule changes in resistance in response to fluctuations in current flow, it appears that parasitic drain measurements using this technique are going to be challenging!
If we look back at image 1, we can see how the ground lead resistance remains stable at 1.52 mΩ throughout cranking with varying degrees of current and voltage (Ohms law)
We can therefore conclude that during periods of high current flow through the ground lead, resistance appears to stabilize with minimal fluctuations. (With a good secure ground lead)
However, during periods of tiny current flow (parasitic drain) the ground lead & chassis resistance fluctuates in proportion due to numerous thermal effects.
I think we need to get this in perspective when we talk about resistance fluctuations; we are not talking about ohms, we are talking milli and micro ohms that are typically negligible but never the less have an adverse effect on our parasitic drain measurement calculation.
As if this wasn’t tricky enough, using the current clamp during this experiment introduced another variable in terms of accuracy at very low current flow thanks to drift and inherent environmental noise.
To summarize, we have a varying cable resistance, drifting current clamps and zero offsets in which to accurately calculate parasitic drain!
For this reason, I want to confirm the performance of the TA234 current clamp using our known fixed 0.1 Ω resister inserted in series with the vehicle ground lead.
More information can be discovered here
Below we have the resister inserted as above with channel D measuring the volt drop across the
0.1 Ω resister alone
Channel A is measuring the entire volt drop across the 0.1 Ω resister and ground lead to engine
Total resistance = 0.1 + 0.00152 = 0.10152 Ω
Channel B is measuring the volt drop across the 0.1 Ω resister and ground lead to the chassis.
Total resistance 0.1 + 473 micro ohms = 0.100473 Ω
The 473 micro ohms value was calculated during a previous ground lead calculation test as per Image 1
Remember this is a single vehicle ground lead that utilizes a crimped eyelet midway through the lead to provide a fixing point in which to bolt the lead to the chassis on route to the engine.
As we can see above the 3 math channels indicate very similar parasitic drain values, with the most accurate being channel D recording 2.341 mV drop across the 0.1 Ω resister (math channel D/0.1 = 23.4 mA)
The remaining math channels incorporate the 0.1 Ω resister into their total resistance value and now appear more repeatable thanks to the increased resistance and stability of the known 0.1 Ω resister. This appears to have removed the inherent instability of the vehicle ground lead at such low current flow
Looking at the current clamp (Channel C) we display 25.6 mA (within the ruler table) but our clamp has drifted up to 11.25 mA, therefore the current clamp returns 25.6 – 11.25 = 14.35 mA
The error between the 0.1 Ω resister at 23.4 mA and the current clamp at 14.35 mA equates to 9.05 mA (39% error)
Both of the above values conclude that the parasitic drain is below the target 80 mA and therefore in specification. Given the 0.1 Ω resister is not subjected to the variables of the current clamp and reduces the effects of resistance fluctuations, I would most certainly choose the 0.1 Ω resister method for parasitic measurement accuracy over the current clamp
Whilst it is not impossible to measure parasitic drain across a good ground lead it is not without challenges
• Prior to measuring I would carry out zero offset to the channels capturing the voltage drop
• When measuring current (in order to calculate cable resistance) ensure the clamp has been switch on for 10 minutes
• Ensure the clamp is used away from sources of environmental noise.
• Be aware of the minimum operating temperatures of the current clamp (0° C).
• Ensure the clamps internal battery is prepared for testing
• Take advantage of the Hardware Filter if using the PicoScope 4425 or 4225 scopes
• Ensure the use of low pass filtering
More information on filtering can be found here
Returning now to Channels C and D in images 2 to 6, I thought it would be interesting to capture both Powertrain and Convenience CAN of the vehicle under test (VW T5 Van)
I wanted to understand the relationship between these networks during shutdown, sleep and wake up. Below we have CAN activity (or lack of) during the transition from shutdown to sleep mode. Note how powertrain CAN returns to zero volts where as Convenience CAN Lo settles at approx. 12 V in sleep mode. The transition to sleep for these networks occurs simultaneously at the point the parasitic drain drops to the indicated 41.38 mA
Looking now at unlocking the vehicle with the key fob (from sleep mode) we can see how the parasitic drain momentarily increases to above 500 mA prior to Convenience CAN waking approx. 80 micro-seconds before Powertrain CAN
Whilst the formation of data “bits” looks correct for the Convenience CAN, Powertrain CAN appears to indicate activity only and data bits that exhibit poor structure. Given the ignition was not turned on during this test (to activate Powertrain CAN “On-line”) I will assume this is characteristic behavior. Perhaps a discussion/experiment for another time?
More information on Convenience CAN can be viewed here
I certainly hope the above will help as it has certainly been a “journey” for me
Thank you for taking the time to deep dive into this topic.
I have one suggestion, why don't we use the manufacturer installed battery sensor as a shunt resistor?
Generally these are 100 micro Ohm shunt resistors with a monitoring circuit. By using this method we have the advantage of not disconnecting the battery, not needing to use a current clamp at all. The only thing we need to know in advance is the defined resistance of the battery sensor. This should negate any accuracy issues with the current clamp method. Generally though, absolute accuracy is less critical, what I really want is a method that doesn't require a current clamp, doesn't rely on batteries and doesn't require disconnection of the battery. This way the picoscope can be left monitoring the vehicle for significant periods of time.
I have attached a data sheet on a battery sensing circuit that is used by Delphi for your perusal.
What do you think?
Using the existing Battery Sensor sounds like all boxes ticked as you say,
No battery disconnection
No current clamp drift/zero
No current clamp internal battery
Defining the resistance of the battery sensor is key and if 100 micro Ohm is accurate/typical, it is worth investigating. I will take a look thorough the attached PDF’s for sure (Thank you)
We do still however have the zero offset to consider for the channel measuring the voltage drop across this battery sensor which introduces another variable as mentioned above.
Moving on to starter cable resistance calculation, you are correct that resistance changes with amperage, but only if the volt drop remains fixed.
12 V drop across a cable carrying 200 mA = 12/0.2 = 60 Ohms
12 V drop across a cable carrying 400 mA = 12/0.4 = 30 Ohms
With the scenario I described in image 1, we can see how the volt drop and current flow are proportional to one another. I.e. as the current increases, so does the voltage drop (The voltage drop is not fixed as in your example above)
The calculated cable resistance remains fixed throughout cranking at 1.52 milli Ohm
I think we can all agree that using the Starter Cable as a fixed value resistance for parasitic drain measurements is floored for the numerous reasons described in my forum post.
One of those reasons was thermal but another not mentioned is the thermocouple effect.
If we consider the construction of our copper ground lead, it is terminated using aluminum eyelet crimps, bolted to a steel chassis and crimped about a lead battery post!
Based on cable temperature and the tiny voltages generated thanks to the thermocouple effect of dissimilar metals, we have a fluctuating total cable resistance (influenced by ambient conditions) in which to determine parasitic drain with mA resolution. (Challenging for sure)
I hope this helps, take care…….Steve
I could well be incorrect but isn't the volt drop going to stay the same in automotive because everything is essentially a 12V circuit (excluding PCM controlled circuits)?
In regards to the original issue here, this all seems very complicated compared to simply volt dropping fuses to find a closed circuit current draw. Why isn't anyone talking about this method instead? What am I missing?
Volt dropping fuses using a reference chart to find a draw is a superior method for many reasons; first you do not need an amp clamp, second, you can find the circuit without disconnecting and reconnecting anything thus disturbing sleep and having to inadvertently start over and have a pile of fuses that need to go back correctly. You can also go directly to whatever you'd like to in the circuit once you've found which fuse is at fault with the draw still active and start disconnecting components in that circuit directly 1 at a time.
Do you have a picture of this 0.1Ohm resistor you used?
Where did you source it from and do you have a part number?
What's the maximum continuous current handling capability?
I've been trying to find a current shunt online, I found one from Keysight Technologies;
However it is only 1mV per A, with a resistance of 0.001Ohm so that gives 10x higher voltage drop than a battery sensor current shunt (0.0001Ohm). The 4425 input sensitivity is just not good enough if we want a resolution in the 10s of milliamps, as per a multimeter in the mV range. However this product can also handle 15A continuous current
I really like the product design of this device, just that it doesn't quite fulfill our needs on the resistance side. Perhaps this could be a future probe idea from Pico
Using the battery sensor current shunt idea I think can be ruled out, as the voltage drop is too low for the scope to handle, we would need to build some kind of amplifier circuit in order for this measurement to be viable. I believe the circuit design for these devices is utilizing exactly that method. Frustratingly I have the equipment to measure the output of these sensors on the LIN bus, but during quiescent mode the sensor stops transmitting data..
So if you have any suggestions for suitable current shunts it would be great if you can share.