Hi, All,
As we know the serial connection of bridge measurement for resistors by voltage excitation, I think we can use parallel connection using current exciatation for many large resistors.
Any sense? Any good experience can be shared?
If you are interested in measuring a number of unknown resistors using a single current-source excitation, then you should connect the unknown resistors in series. Ohms Law (R=V/I) gives the resistance of each unknown resistor using the following method. The series connection makes the current in all unknown resistors the same, and equal to a value provided by the current excitation. By measuring the voltage across each of the unknown resistors (differential voltage measurements), and dividing each of those measured voltages by the common current, you can determine the resistance of each “unknown” resistor.
There are a couple details to consider with this approach. First, the accuracy of the current excitation (how close the actual current is to the desired value) will limit the accuracy of the resistance measurements. Some manufacturers use a ratiometric approach that improves the accuracy of resistance measurements. This approach “ratios out” possible errors in the current excitation with proportional errors in the differential voltage measurement. Please review the manufacturer’s accuracy specifications (including ratiometric accuracy) and consider how they apply to your specific measurements.
The second detail is the voltage compliance of the current source excitation. Voltage compliance is the maximum voltage a current source can produce in it’s effort to push the desired current through the string of unknown resistors. As you add additional unknown resistors in series, the required voltage increases. This required voltage is the product of the current excitation with the total resistance through which the excitation current is flowing. Please be careful not to exceed the current excitation’s voltage compliance.
To get to cabbageheart’s question (finally). I do not see an easy way to determine the resistances of a number of unknown resistors using a single current excitation if the unknown resistors are connected in parallel. The problem with this approach (as pointed out by SC) is that the excitation current splits in an unknown way through each of the resistors, and therefore, there is no common current in each unknown resistor.
Cabbageheart, would you be willing to explain why you desire to connect the unknown resistances in parallel when using a current excitation? Is it to enable only a single voltage measurement across the parallel connection of the unknown resistors? If so, I believe the extra effort of measuring the unknown current through each of the unknown resistors exceeds the benefit of the proposed approach. It’s uncommon to get something for nothing.
My apologies to the Forum for my lengthy post.
From my experience, we can measure 4 PT100s in serial from one current excitation and 2 PT1000s.
Let us think to measure a pt1000! If we use you a shunt 1000 Ohm, we can increase the excitation current because the total resistance is half! If we measure 5 PT1000s in one excitation, we can use a shunt resistor as 1000OHM, we still can measure them!
The biggest problem of serial measurement is that one of the cable broken will be a disaster. It is the limitation of the current excitation of CR3000 and CR5000. Maybe, you can ask us to use multiplexer.And then, voltage excitation can be used then!!
The reason why I think of parallel resistor measurement is for bigger resistor. As everybody know, a shunt parallel resistor will reduced totally resistance, and if we know the shunt resistor's resistance very well, we will know the exact current shared by the bigger measured resistor and we will get the value of it! It is the same as serial for voltage excitation!
I want to use a known resistance not an unknown resistor!
Dear Cabbageheart,
In your last post, you wrote:
“The biggest problem of serial measurement is that one of the cable broken will be a disaster. It is the limitation of the current excitation of CR3000 and CR5000.”
I agree regarding the liabilities of series-connected components (I assume your “serial” means the same as my “series”). One open connection failure renders measurements from the entire string unusable. I do not quite see how this is a limitation specific to the CR3000 or CR5000. Instead, it seems generic to all series-connected components.
In my previous post, I apparently responded to a different question than you asked. To avoid repeating this, I will refrain from responding before I more completely understand your question.
Please explain as precisely as possible what you are proposing. Is the current-source excitation connected to a known 1k-Ohm resistor that is connected in parallel with a single unknown resistor, or are there multiple unknown resistors in the network? If there are multiple unknown resistors, are they connected in parallel with one another or are they connected in series with one another? How many amps (or mA) does the current excitation provide? Is the current level programmable? What is the voltage compliance of the current excitation? What is (are) the resistance range(s) of the unknown resistor(s)?
I appreciate your understanding and patience as I try to understand your question and as I try to contribute to this forum.
Hi, Larry,
You are right.
Case1: the current-source excitation connected to a known 1k-Ohm resistor?
That is connected in parallel with a single unknown resistor. If this is a thermistor from 10K TO 100K range, it is impossible to use current excitation. But when you parallel a 1K, you can use 200 microamps or more current to excite them, and you will get the diff volt within the range of 200 mV. I don't want to use 100 ohm because I suspect the divided current of this measured resistance is too small to give a good resolution. Please note that CR3000 OR 5000's current excitation can be programmable.
Case2, we want to measure 5 or more PT1000 by one current excitation port, but, it can't work! we still not sure why the Resistance instruction can't measure more when we connect serially more together. (Resistance (r_1,1,mV200,1,Ix1,4,200,True ,True ,0,_50Hz,1.0,0)), then we use ExcitI instruction, and when we use an a known 1k-Ohm resistor to parallel, we can get the measurement!
Dear Cabbageheart,
For your Case 1:
A current excitation source connected to a parallel combination of a known 1k-ohm resistor and an unknown 10k- to 100k-ohm resistor.
It is unclear why you are limiting yourself to a +-200 mV input range. Why not use the CR3000’s +-5000 mV input range? By doing so, you could eliminate the parallel 1k-ohm resistor. The voltage compliance (defined in my first post of this thread) of a CR3000 current excitation is +-5 V. To avoid going beyond this compliance limit when driving a 100k-ohm resistance, the current excitation must be set at or below 50 uA, and the voltage across the 100k-ohm resistor will be 5 V. Therefore, I suggest using the +-5000 mV input range on the CR3000. By the way, the voltage will be 0.5 V when the resistance drops to 10k ohms.
Please bear in mind that the offset accuracy of the current source on a CR3000 is 0.5 uA. This inaccuracy will give about a one percent error in the measurement of the unknown resistor.
Your proposal using a 1k-ohm resistor in parallel with the unknown 10k- to 100k-ohm resistor will also work. You can measure the voltage across the parallel combination of resistors, compute the current through the fixed 1k-ohm resistor, and then subtract this current from the total excitation current. The resulting current is the current through the unknown resistor. Because you know the voltage across, and current through, the unknown resistor, you can determine its resistance (R=V/I). You can program these calculations in the CR3000.
There are also liabilities to this approach. The 1k-ohm resistor’s inaccuracy, especially over temperature, will affect the accuracy of your measurements. If you stick with this approach, I still recommend that you increase the input range on the voltage measurement above 200 mV, and optimize your parallel, fixed resistor value and current excitation value.
Case 2:
Five PT1000s, each connected in series with one another, and then connected to a current excitation.
I believe it is possible to measure this network without the 1k-ohm resistor in parallel that you suggest. Perhaps your issue is too low of a voltage input range on the Resistance instruction. I offer the following without the 1k-ohm parallel resistor.
A PT1000 has about 1000 ohms of resistance at zero degrees C, and a temperature coefficient of about +3.85 ohms per degrees C. At 64.9 degrees C, each PT1000 will provide about 1.25k ohms of resistance and the five PT1000s in series will provide about 6.25k ohms of resistance. To stay within the +-5 V compliance of the CR3000 current source, the maximum excitation current is 800 uA. The voltage across each PT1000 will be about 1.0 V at 64.9 degrees C. Therefore, I recommend the +-1000 mV input range on the datalogger. By the way, the 0.5 uA current source accuracy considered in Case 1 represents only a 0.06 percent error in this case.
Here is your CR3000 instruction modified to accomplish the steps described above:
Resistance (r_1, 1, mV1000, 1, Ix1, 4, 800, True, True , 0, _50Hz, 1.0, 0)
You can use similar logic for six or more PT1000s in series. Simply adjust the current source value and the input ranges accordingly.
For case 1, I'm only give you an example. In many cases, the range of thermistor are not so easy to control, when it is too big, maybe the parallel is a selection because we would like to prepare one proper precision resistor. By the way, I would like to know your consideration about introduced heat because of larger amps?
We had tested the resistance instruction, we found only 2 pt1000 can be works and 4 PT100 can be done. I will re tested this here.
Perhaps the self-heating question best belongs in a new thread. It seems to be a separate issue from the parallel-resistor question.
A summary of my thoughts expressed in this thread:
I believe the proposed parallel resistor essentially accomplishes two things, both of which the CR3000 can achieve without additional hardware. First, the parallel resistor keeps the current excitation source within its voltage compliance limit, and second, it reduces the measured voltages across the variable (or unknown) resistor(s). I believe CR3000 users can achieve these same objectives for many applications (including the ones mentioned by Cabbageheart) without the parallel resistor. The key is to program the current excitation to a lower value and to program the voltage input range a higher value. Ohms law (V=IR) provides the relationship required to determine the appropriate current-output and voltage-input levels. I look forward to hearing if hardware testing shows otherwise.