CS1200 CMRR

Hello everyone,

for my application at hand* I need an extraordinary high CMRR (> 100 dB) throughout the full bandwidth of the CS1200 (200 MHz).

1) Why is the CMRR given only up to 50 MHz in the specifications? What happens above this frequency?

2) How is the CS1200-CMRR measured in the specifications? What is the exact setup (maybe a picture could be helpful)? Is it measured with or without probe? If it has been measured without probe, what happens if I measure the CMRR with a probe connected?

3) In the CS1200 user guide a "short x10 oscilloscope probe fitted with a common mode choke" is introduced. What is the best possible CMRR that can be reached with this special probe? What is the typical CMRR that can be reached with the standard probes?

Thank you very much for answering my questions. Is there already someone who employs the CS1200 to measure signals during an EMI immunity test?

* EMI immunity tests with fast transients (= burst)

Hello everyone, for my application at hand* I need an extraordinary high CMRR (> 100 dB) throughout the full bandwidth of the CS1200 (200 MHz). 1) Why is the CMRR given only up to 50 MHz in the specifications? What happens above this frequency? 2) How is the CS1200-CMRR measured in the specifications? What is the exact setup (maybe a picture could be helpful)? Is it measured with or without probe? If it has been measured without probe, what happens if I measure the CMRR with a probe connected? 3) In the CS1200 user guide a "short x10 oscilloscope probe fitted with a common mode choke" is introduced. What is the best possible CMRR that can be reached with this special probe? What is the typical CMRR that can be reached with the standard probes? Thank you very much for answering my questions. Is there already someone who employs the CS1200 to measure signals during an EMI immunity test? * EMI immunity tests with fast transients (= burst)

Hello Hans,
Sounds like an interesting application. Unfortunately I do not have a recent measurement of the CS1200 CMRR. It should be better than the CS548 CMRR which is shown below. We will do a measurement this week.

In answer:
1. We use our CS1070 power amplifier to boost the CS548 signal generator to 34V p-p. We then use this as the common mode source. The CS1070 only has a bandwidth of 52 MHz. We can run it up to 65MHz, which is the limit of the built in sig gen. We are also using our FRA Gain/phase to measure this, and for this we need a tracking generator. If we use a higher band width sig gen, the measured signal is in the noise floor. But we will give this a go with a reduced measurement bandwidth. So the answer was this was easiest for us to do.

2. The CS1200 CMRR is measured with a 50 ohm BNC cap placed on the CS1200 BNC. The CS1200 is mounted on a 50 mm thick plastic enclosure above an aluminium ground plane. The generator output is directly coupled to the BNC cap, via a BNC Tee, and the Generator common is conducted to the ground plane using 25 mm wide copper tape as short as we can make it spread over the BNC Tee shield. The other output from the BNC Tee is connected to another CS1200. An older picture is attached below, but this is not correct (the ground connection has too much inductance). We will take a better picture later in the week.

3. Any probe will reduce the CMRR.
a. The probe coaxial cable shield has series inductance. The series inductance forms an LC tank with the free space capacitance of the probe to the ground plane, and the capacitance of the UUT to the ground plane. The tank resonates with f = 1 / [2 pi sqrt (LC)]. The resonant signal can be large. One way to reduce it is to make the tank lossy with a common mode choke with a large resistive component in its impedance response around the resonant frequency. We are experimenting with short, double shielded cables and chokes to minimise this problem.
b. The probe attenuation attenuates the signal, but not the common mode. The result is that any attenuation reduces the CMRR by the attenuation value. For example a 10x probe reduces the CMRR by 10x or 20 dB. It is best to use a 1x probe as result - eg just a short piece of coax.
The standard 10x probe is about 30 dB worse in CMRR at 50 MHz.

I think most people are using the CS1200 inside high voltage enclosures. All the same the CS1200 is fully shielded. The battery connection is probably the most susceptible to the RF field, but we found no change in CMRR dependant on the battery being between the probe and the ground plane, or above it.

With the setup we had, the RMS signal amplitude is 34/(2 x sqrt(2)) = 12V RMS. The distance is 50 mm, so that is 12/0.05 = 240 V/m, which is a moderate level (well above the standard industrial 10 V/m). What field strength are you using?

I hope this provides a bit more clarity.

regards,
Bart

Hello Hans, Sounds like an interesting application. Unfortunately I do not have a recent measurement of the CS1200 CMRR. It should be better than the CS548 CMRR which is shown below. We will do a measurement this week. In answer: 1. We use our CS1070 power amplifier to boost the CS548 signal generator to 34V p-p. We then use this as the common mode source. The CS1070 only has a bandwidth of 52 MHz. We can run it up to 65MHz, which is the limit of the built in sig gen. We are also using our FRA Gain/phase to measure this, and for this we need a tracking generator. If we use a higher band width sig gen, the measured signal is in the noise floor. But we will give this a go with a reduced measurement bandwidth. So the answer was this was easiest for us to do. 2. The CS1200 CMRR is measured with a 50 ohm BNC cap placed on the CS1200 BNC. The CS1200 is mounted on a 50 mm thick plastic enclosure above an aluminium ground plane. The generator output is directly coupled to the BNC cap, via a BNC Tee, and the Generator common is conducted to the ground plane using 25 mm wide copper tape as short as we can make it spread over the BNC Tee shield. The other output from the BNC Tee is connected to another CS1200. An older picture is attached below, but this is not correct (the ground connection has too much inductance). We will take a better picture later in the week. 3. Any probe will reduce the CMRR. a. The probe coaxial cable shield has series inductance. The series inductance forms an LC tank with the free space capacitance of the probe to the ground plane, and the capacitance of the UUT to the ground plane. The tank resonates with f = 1 / [2 pi sqrt (LC)]. The resonant signal can be large. One way to reduce it is to make the tank lossy with a common mode choke with a large resistive component in its impedance response around the resonant frequency. We are experimenting with short, double shielded cables and chokes to minimise this problem. b. The probe attenuation attenuates the signal, but not the common mode. The result is that any attenuation reduces the CMRR by the attenuation value. For example a 10x probe reduces the CMRR by 10x or 20 dB. It is best to use a 1x probe as result - eg just a short piece of coax. The standard 10x probe is about 30 dB worse in CMRR at 50 MHz. I think most people are using the CS1200 inside high voltage enclosures. All the same the CS1200 is fully shielded. The battery connection is probably the most susceptible to the RF field, but we found no change in CMRR dependant on the battery being between the probe and the ground plane, or above it. With the setup we had, the RMS signal amplitude is 34/(2 x sqrt(2)) = 12V RMS. The distance is 50 mm, so that is 12/0.05 = 240 V/m, which is a moderate level (well above the standard industrial 10 V/m). What field strength are you using? I hope this provides a bit more clarity. regards, Bart

The earlier (incorrect setup):

The earlier (incorrect setup):

Hello Bart,

thank you very much for your detailed answer. During EMC tests with fast electrical transients (burst) we apply up to 6 kV (= 6000 V) between the reference table's metallic surface and the mains connection of our device. Up to now we can measure meaningful values only with a Langer EMV A200 (which has 500 kHz bandwidth only). This probe is optimized in size and hence is much less prone to common mode currents. Maybe we should give the CS1200 a chance. Would you suggest to contact our local distributor (www.meilhaus.de) to get a trial kit?

kind regards,
Hans

Hello Bart, thank you very much for your detailed answer. During EMC tests with fast electrical transients (burst) we apply up to 6 kV (= 6000 V) between the reference table's metallic surface and the mains connection of our device. Up to now we can measure meaningful values only with a Langer EMV A200 (which has 500 kHz bandwidth only). This probe is optimized in size and hence is much less prone to common mode currents. Maybe we should give the CS1200 a chance. Would you suggest to contact our local distributor (www.meilhaus.de) to get a trial kit? kind regards, Hans

Hello Hans,
I think that is a plan. We can sell the CS548 without any internal channels, and just a CS1200 if you want, which reduces the cost. Please email us at support @ cleverscope.com, and let us know your address and email details, and we will organize this with Meilhaus. They do not have a CS1200 at present, and our European office will have to ship it to them.

kind regards, Bart

Hello Hans, I think that is a plan. We can sell the CS548 without any internal channels, and just a CS1200 if you want, which reduces the cost. Please email us at support @ cleverscope.com, and let us know your address and email details, and we will organize this with Meilhaus. They do not have a CS1200 at present, and our European office will have to ship it to them. kind regards, Bart