FRA of current measurement with CS1070

I'm looking to run frequency response analysis (gain/phase) on a current measurement signal chain where the current sensor is a hall effect type. The current sensor gain is fixed at 13.3mV/A and can't be modified externally, unlike a shunt + amplifier current measurement. Consequently, it seems like I need to use a signal source that can output an appreciable amount of current to perform the FRA.

I have a CS320A-FRA. Would the CS1070 be a good fit for this application? With the 1Ω output, of course. ±1Apk seems just large enough to get passable SNR throughout rest of the signal chain (hall effect current sensor output, post filters, etc.). I'd probably use the CS1070 current sense output for channel A and a 1x probe for channel B unless you have other suggestions.

The only other alternative that I can imagine is a wide bandwidth isolating transformer with a step down turns ratio, but I don't know if anyone makes something like that commercially.

Side note: Your forum software does not handle special characters (Ohm symbol and/or plus minus symbol). "Conversion from collation utf8mb3_general_ci into latin1_swedish_ci impossible for parameter".

I'm looking to run frequency response analysis (gain/phase) on a current measurement signal chain where the current sensor is a hall effect type. The current sensor gain is fixed at 13.3mV/A and can't be modified externally, unlike a shunt + amplifier current measurement. Consequently, it seems like I need to use a signal source that can output an appreciable amount of current to perform the FRA. I have a CS320A-FRA. Would the CS1070 be a good fit for this application? With the 1Ω output, of course. ±1Apk seems just large enough to get passable SNR throughout rest of the signal chain (hall effect current sensor output, post filters, etc.). I'd probably use the CS1070 current sense output for channel A and a 1x probe for channel B unless you have other suggestions. The only other alternative that I can imagine is a wide bandwidth isolating transformer with a step down turns ratio, but I don't know if anyone makes something like that commercially. Side note: Your forum software does not handle special characters (Ohm symbol and/or plus minus symbol). "Conversion from collation utf8mb3_general_ci into latin1_swedish_ci impossible for parameter".

Hi Sgleysti,
I will leave Bart to respond to your technical question but just to let you know, I have taken on board your comment about the lack of symbol support in the forum and fixed it - as well as adding a toolbar to give you instant access to them - Ωµ°±√π≤≥.

Hope that helps for next time!

Regards,
Roger

Hi Sgleysti, I will leave Bart to respond to your technical question but just to let you know, I have taken on board your comment about the lack of symbol support in the forum and fixed it - as well as adding a toolbar to give you instant access to them - Ωµ°±√π≤≥. Hope that helps for next time! Regards, Roger

Hi Sgleysti,
Thanks for your query. You can achieve what you want, but you will have to pay attention to maintaining SNR. Can you please tell me the frequency range you want to do the gain/phase over.

Here are some steps:
1. Do a self cal on the CS320A-FRA to get good offsets. Click the CS1070 enable switch up to disable it.
2. Connect the Isense output on the CS1070 to the Chan A input, preferably with Coax. The I sense uses a 100mΩ resistor to sense current, so 100mV/A, or like a 10x probe. Set the Chan A probe setting to 10x, and the Chan A units to (A) by overtyping in the Channel Control above the Scope Graph. In this way you can read Amps directly of the graph.
3. If the Hall Effect Sensor is floating, attach one side to the Out 1R signal, and the other side to the Out 1R common shield. It’s best to use Coax and pigtails. Do not use a 1x probe, there is a 300Ω series resistance in it.
If the Hall Effect sensor is ground referred, connect the ground side to the I sense Output common. Take note of the schematic on the CS1070 lid – Out1R is not ground referred.
4. Connect your sensor output (13.3mV/A) to Chan B. You can either use coax or a 1x probe. The probe will limit bandwidth to about 6 MHz, but you need the 1x setting to maintain SNR.
If you want to read the sensor output in A, open Settings/Analog Names and units, and set the Pt. 2 pair for channel be to 13.3m 1, which will display in A. Change the Chan B units to (A).
5. Assuming both channels are displaying as A, adjust the range on both to be ±1 A. You may need to vertically expand your scope display so it has 10 or 20 divisions.
6. Use View/FRA to see the Frequency Response Analysis panel, and click the top right button to FRA System ON. Set the current frequency to somewhere in your pass band – eg 10 kHz. Set the Sig Gen Amp to 0.1V. Set the Start/Stop frequency range as you want, and set the other values.
7. Start capturing, with the scope graph setup to capture a cycle or two (100us width or more). Turn on the CS1070 by clicking the enable switch down.
8. Looking at the View/Signal Information adjust the offset to get about 0V DC on the A channel.
9. Now adjust the Sig gen up in 0.1V increments (position the cursor to the right of the 1/10ths digit, and use the up down keys on the keyboard) to get around 0.9A. You may need to go finer. If you are not getting fine enough control, you may need to put a 10 or 20 dB coax attenuator between the sig gen and the CS1070 input.
10. Once you have the Sig Gen amplitude and Offset set as you need, click the Start Sweep button. The app will adjust the time base and filters suitably.

I used a 9 Ω load on Out 1R, and got the attached graphs. This wound up with an SNR of 38dB, allowing a 20 dB reduction in gain over the pass band before you run into too much noise.

Please let me know if you have questions.

Bart

Hi Sgleysti, Thanks for your query. You can achieve what you want, but you will have to pay attention to maintaining SNR. Can you please tell me the frequency range you want to do the gain/phase over. Here are some steps: 1. Do a self cal on the CS320A-FRA to get good offsets. Click the CS1070 enable switch up to disable it. 2. Connect the Isense output on the CS1070 to the Chan A input, preferably with Coax. The I sense uses a 100mΩ resistor to sense current, so 100mV/A, or like a 10x probe. Set the Chan A probe setting to 10x, and the Chan A units to (A) by overtyping in the Channel Control above the Scope Graph. In this way you can read Amps directly of the graph. 3. If the Hall Effect Sensor is floating, attach one side to the Out 1R signal, and the other side to the Out 1R common shield. It’s best to use Coax and pigtails. Do not use a 1x probe, there is a 300Ω series resistance in it. If the Hall Effect sensor is ground referred, connect the ground side to the I sense Output common. Take note of the schematic on the CS1070 lid – Out1R is not ground referred. 4. Connect your sensor output (13.3mV/A) to Chan B. You can either use coax or a 1x probe. The probe will limit bandwidth to about 6 MHz, but you need the 1x setting to maintain SNR. If you want to read the sensor output in A, open Settings/Analog Names and units, and set the Pt. 2 pair for channel be to 13.3m 1, which will display in A. Change the Chan B units to (A). 5. Assuming both channels are displaying as A, adjust the range on both to be ±1 A. You may need to vertically expand your scope display so it has 10 or 20 divisions. 6. Use View/FRA to see the Frequency Response Analysis panel, and click the top right button to FRA System ON. Set the current frequency to somewhere in your pass band – eg 10 kHz. Set the Sig Gen Amp to 0.1V. Set the Start/Stop frequency range as you want, and set the other values. 7. Start capturing, with the scope graph setup to capture a cycle or two (100us width or more). Turn on the CS1070 by clicking the enable switch down. 8. Looking at the View/Signal Information adjust the offset to get about 0V DC on the A channel. 9. Now adjust the Sig gen up in 0.1V increments (position the cursor to the right of the 1/10ths digit, and use the up down keys on the keyboard) to get around 0.9A. You may need to go finer. If you are not getting fine enough control, you may need to put a 10 or 20 dB coax attenuator between the sig gen and the CS1070 input. 10. Once you have the Sig Gen amplitude and Offset set as you need, click the Start Sweep button. The app will adjust the time base and filters suitably. I used a 9 Ω load on Out 1R, and got the attached graphs. This wound up with an SNR of 38dB, allowing a 20 dB reduction in gain over the pass band before you run into too much noise. Please let me know if you have questions. Bart

Hi Bart,

Thanks. This is really helpful.

Understood about only using 1x probes to measure and not to source signals. My concern with using coax for measurements is ringing if not properly terminated. It looks like the CS320A-FRA doesn't have built in 50Ω terminations, but I could use an inline device.

The most important measurement is phase response at 1kHz. Runnerup is attenuation near the sampling Nyquist frequency, which tops out at 300kHz. The target is ~40dB, and most of this is from a post filter. My best bet is probably to measure the hall sensor response in this range to ensure it's mostly flat, and then measure the post filter separately.

The hall sensors are isolated types. I'm guessing the CS1070 I sense output common is connected to earth ground just like ground leads of the CS320 probes. I'm of a mind to make board modifications to ensure no connection between the input and output sides of the hall sensors through Y capacitors or similar, just in case. Such things should be much higher impedance than the hall sensor pass element, but 300kHz is a long way from DC.

Hi Bart, Thanks. This is really helpful. Understood about only using 1x probes to measure and not to source signals. My concern with using coax for measurements is ringing if not properly terminated. It looks like the CS320A-FRA doesn't have built in 50Ω terminations, but I could use an inline device. The most important measurement is phase response at 1kHz. Runnerup is attenuation near the sampling Nyquist frequency, which tops out at 300kHz. The target is ~40dB, and most of this is from a post filter. My best bet is probably to measure the hall sensor response in this range to ensure it's mostly flat, and then measure the post filter separately. The hall sensors are isolated types. I'm guessing the CS1070 I sense output common is connected to earth ground just like ground leads of the CS320 probes. I'm of a mind to make board modifications to ensure no connection between the input and output sides of the hall sensors through Y capacitors or similar, just in case. Such things should be much higher impedance than the hall sensor pass element, but 300kHz is a long way from DC.

Hi Sgleysti,
You are quite right. If using Coax in the frequency range <1MHz, you don't need a termination, but you do need to keep the coax short, so any ringing is well above your passband. The main deal is shielding.

I agree, your main goal is to measure the Hall sensor and filter, not the Y caps. Maybe removing them is a good plan. You could also check how they change the response.

Attach is a photo of the top label of the CS1070, which shows how the connections work.

It sounds doable to me!

Bart

Hi Sgleysti, You are quite right. If using Coax in the frequency range <1MHz, you don't need a termination, but you do need to keep the coax short, so any ringing is well above your passband. The main deal is shielding. I agree, your main goal is to measure the Hall sensor and filter, not the Y caps. Maybe removing them is a good plan. You could also check how they change the response. Attach is a photo of the top label of the CS1070, which shows how the connections work. It sounds doable to me! Bart