Reduce Ground Disturbances when using HF2TA Transimpedance Amplifier in a Measurement Setup

Ground loops can potentially pose a serious problem in many measurement setups. Due to the multiple connections to and from the HF2TA current amplifier and other external components, several ground loop related issues can pop up in the signal being measured. These issues can be high-frequency oscillations (>10MHz), 50/60 Hz power line crosstalk, and other pickup from strong electromagnetic fields. In this blog, some of the common ground loop issues and the possible remedies will be presented.

Test setup diagram

The diagram below shows an example of potential ground noise issues when using a HF2TA transimpedance amplifier with a HF2LI or HF2IS. The HF2 high frequency output generator is used to drive a device-under-test and the resulting current is amplified by the HF2TA before being fed back into the HF2 again. The measurement can use either single or both channels of the HF2. There may also be external instrumentation connected to the device-under-test (DUT). The potential grounding issues will be discussed later.

ground_test_setup_diag

Checking for ground noise

In order to check a setup like the one above for the presence of ground noise feedthrough, one can take advantage of the Oscilloscope tool that comes with the ziControl GUI. Here, the sampling rate is set to minimum of 6.4 kSa/s to try to capture any low frequency noise. RMS averaging can be used to clearly see a noise pattern when a pattern is not easily distinguishable in the time domain waveform. In this example, the HF2 output generator is turned off i.e. no modulation signal. However, one can see that there is clearly a 50 Hz power line coupling, including its harmonics, measured at the input. The gain of the HF2TA is set at 10k A/V in this case. The Oscilloscope tool is a convenient way to check quickly if ground noise issues exist already before the start of an experiment.

oscilloscope_noise

Sources of ground noise

Ground noise can sometimes couple into the signal path capacitively but more likely through amplifiers where the signal input and output voltages are referenced to ground. Therefore, any disturbances on the reference ground will also show up as noise in the signal. Here are some examples of possible sources for ground noise:

EMF (electromagnetic field) disturbances

As can be seen in the previous setup diagram, connecting different cables and ground shields together form multiple ground loops. According to Lenz’s Law, an induced current will be generated in a closed, conducting loop when the loop experiences an external emf. This also applies to the ground loops and, since a ground loop is never perfectly conductive (i.e. it is always resistive and inductive), a voltage noise will result from the induced current. Having multiple ground loops makes the prediction of where the noise will travel even more difficult since each ground loop may have slightly different resistance and inductance, thus experiencing different noise coupling.

different reference ground

Very often , when connecting to other systems or instruments, the reference ground of two different instruments may not be the same. Or, for example, one reference ground may be noisier than the other. In these cases noise coupling might result due to this uneven potential.

imperfect grounding

Most instruments have a common (earth) ground. This earth ground can sometimes be very noisy due to flaws in the building grounding.

Why the use of HF2TA accentuates the problem of ground noise

Whether the noise coupling comes from emf or an unwanted “dirty” ground injection, the use of HF2TA seems to worsen the effect of ground noise problems. This is due to the fact that any noise injected into the signal path will be amplified by the HF2TA which then can show up clearly in the input signal noise spectrum. Therefore, the HF2TA is often not the source of the problem but an indicator of a potential ground noise issue.

How to minimize ground noise coupling in a setup

For the problems mentioned above, an obvious answer would be to connect everything to a clean ground reference. However, clean grounds are hard to come by, especially in a lab environment where many different instruments (including high-power, high frequency generators) are either running on the same power supply or in close proximity. There are however some guidelines that should be considered which can significantly minimize the noise coupling through ground:

Better HF2TA/DUT ground connections to HF2

Use a low-ohmic ground/earth reference as close as possible to the DUT, where the sensitive current signal is measured. Remember that connection to this ground must be low-ohmic and the ground itself must be “clean”, otherwise it is unlikely to bring any improvement.  This precaution is, however, not effective if the ground noise is coming from some ground leakage/coupling to the signal path elsewhere.

Minimize number of ground loops or break ground loops

This measure is especially effective against EMF-type noise coupling. Rearrange the connection cables (BNC, SMA, Zctrl) and minimize the loops which are enclosed by the cables. As mentioned previously, the Oscilloscope tool in ziControl can be used to monitor any improvement brought by the re-wiring of the cables.  The improvement can be seen especially with high frequency pickup from strong electro-magnetic fields. Although putting the whole setup in a Faraday cage against EMF interferences may also be an option, it might not be always practical.
Sometimes, breaking ground loops may be an option although it is not always recommended. For the HF2TA, the ground loop can be broken by switching the shield bias from ground to an external bias, as shown in the screenshot below. Again, the shield bias should be put to a clean ground in order for this technique to be effective. Be aware that the separation of shield ground and the analogue ground may cause a small DC offset. Using coaxial isolators may also be an option.
HF2TA_input_shield
USB de-coupler might help

In the past some customers were concerned about the ground loop connection between the reference ground and the earth ground of the HF2LI through the USB cable. Although internal tests at Zurich Instruments have shown that ground noise injection due to the USB connection is negligable, a specific make of USB-decoupler is nevertheless recommended for those who just want to be on the safe side. In fact, an USB-decoupler makes more sense in cases where the earth ground node of the HF2LI has been galvanically isolated from the  earth ground of the power supply. Here it is highly recommended to remove this ground loop connection through the USB cable in order to preserve the galvanic isolation.

Conclusion

Ground noise coupling is often hard to debug. Moreover, there is no one-stop solution to get rid of it. Understanding the causes, accompanied by careful wiring practices, can help to minimize this issue. Users with more hints on reducing ground noise are welcome to share their findings in the comments section.