I need to make some differential measurements on a device, but both nodes need to be floating from earth.

I found out the hard way that the negative clip of my oscilloscope voltage probes are connected to the mains earth conductor, but I don't have access to a differential probe.

I was wondering if there is an easy solution to this, and I figured this: inside the scope I have seen that the mains are only connected to a power, old style transformer, while the earth lead is screwed to the metal chassis. Since of course the transformer output is isolated from its input, I think that if I just remove the earth screw from the chassis and isolate the cable I can now connect my device to the scope, avoiding the pesky ground current that destroyed my first board.

The negative terminal I would hook to the probe clip is about 60V over earth, the maximum current it can source is just above 30mA. The input transformer can surely withstand a 60V offset, right?

I think that this is safe because the shell of my scope is all made of plastic, the only metallic parts are the BNC connectors and a couple of blade connectors that I use to calibrate my probes, but they should all be safe.

Is it ok to remove the earth connector from my scope to make such a measurement? I have seen this suggestion all over the web, so I think it must be safe, but I want to ask a board of professionals just to be extra sure.


4 Answers 4


The best answer is to isolate the equipment you are trying to measure. When it is isolated, you are free to ground any one point, which will be done thru the scope probe ground lead. That keeps the scope, its chassis, and anything it may be connected to or touching from becoming live.

If isolating the circuit you are trying to measure is not possible, then a portable battery-operated scope is the next best option. These are intended in part for this sort of thing, and are usually well insulated, unlike scopes intended to sit on a bench.

If you are really really careful and can ensure everyone else around you or that might enter the room will be really careful, then you can put the isolation transformer on the scope. Before you do that, very carefully make sure the scope isn't touching anything conductive, and there are no connections to is to elsewhere, like a USB connection to a PC or something. This way you get what you suggest without modifying the scope.

I just can't recommend what you suggest. You may think you found the only connection to ground, but the scope wasn't designed to be modified like that, so you don't really know without testing. A Hi-pot test would tell you, but may also blow up the scope in the process.

Do it right. Use a isolation transformer, preferably on the equipment under test, not on the scope.

  • \$\begingroup\$ +1. I just don't understand why people choose to risk life/limb when isolation transformers are common and basic differential probes are 5% (or less) of the cost of replacing a blown scope. I mean, is $300 too much to keep a $6000 scope (and all of the potential users of said scope) safe? \$\endgroup\$ Aug 24, 2016 at 18:21
  • \$\begingroup\$ @AdamLawrence: 300 is too much for people that saved up years their 300 for a scope \$\endgroup\$
    – PlasmaHH
    Aug 24, 2016 at 19:27
  • \$\begingroup\$ Some equipment in the industrial world has instructions saying you need to supply isolating devices when connecting other equipment (laptops or network hubs etc) even if they are using ethernet and RS232. This is a related issue. \$\endgroup\$
    – Spoon
    Aug 24, 2016 at 19:53
  • 1
    \$\begingroup\$ @PlasmaHH I would think that such a burden would encourage someone to take even better care of their scope vs. deliberately defeating the earthing. \$\endgroup\$ Aug 25, 2016 at 13:44

The more common way I've seen "poor-man's" differential probes implemented is by using 2 channels, then using the math A-B mode. The two ground clips are attached to proper ground.

  • \$\begingroup\$ This answer doesn't really address the original question, but it's really good advice. I would always use this method vs. floating the scope if there's no diff probe available. Results will vary depending on matching of the two probes, but it can work fairly well. Did it yesterday because my diff probes couldn't handle high enough voltage. \$\endgroup\$
    – John D
    Aug 24, 2016 at 18:03

It's possible, and we used to have a 'special' cord sets with the ground pins broken off in the old days, but it's not particularly safe, especially if the voltages are relatively high. It cannot be recommended, mostly for safety reasons.

In this millennium, I have a Tek scope with four (mutually and from ground) isolated channels for this kind of work. It also has a Li-ion battery pack for when more AC isolation is required. Differential probes can work, but their common mode rejection ratio isn't all that great so if you are trying to look at a 10V waveform on a noisy 300V bus the results may not be all that representative of the truth.

As well as the internal power transformer you may also need to worry about the isolation (or lack of isolation) of communication ports such as Ethernet, IEE488, USB or RS-232.

I have an isolation transformer with an internal electrostatic shield (two shields would have been even better) that is also useful in these situations.

Keep in mind that there will be (sometimes large) capacitance across whatever isolation device you insert to isolate the measuring or the device under test and that may load the signal you are trying to measure and distort the results.


The line filters to ground would become open and a charge accumulation from line voltages on the primary coil may result.

It would be better to use an isolation transformer if the internal design is unknown between neutral and ground.

Normally using two matched and calibrated 10:1 probes can produce reasonable results using no ground leads and tip removed with adjacent test pins for V+, current sense V and earth ground. Then verify on the same point in A-B mode for flat line.

or better yet , make a diff amp board to output the voltage and put near DUT test points. Then attenuate inputs within CM range and amplify with sufficient gain.


Not the answer you're looking for? Browse other questions tagged or ask your own question.