In theory, isolation transformers (IST) sound like “the” essential safety device for an ECE workbench. The question is, are you trying to protect your equipment, yourself or both?

The two other essential safety devices for your ECE workbench are a decent variac and Differential Probes.

A Note from HL: I wanted to go ahead and share this post. It has been sitting in my drafts for way too long. My senior year of Computer Engineering has been quite a doozy, especially with COVID causing a late start (result: accelerated pace). I apologize about having to delay the SS2590 review until May/June 2021 (College). By doing the ES9038 HL Stage I and Stage II mods, I ate the time I did have, but it’ll be worth it.

Differences Devices for Different Protection

1: Protecting yourself and DUT: Isolation Transformer + Variac
2: Protecting your equipment: Differential Probe (I recommend Probe Master’s 4232)
3: Protecting Test Equipment and Yourself: Isolated Differential Probes (Note: these usually run $2000 and higher) or an IST + Differential Probes

Important Caveats

Care must be taken here, as the IST has become a hot topic for debate, partly due to incorrect usage scenarios. It is also due to the varying isolation and safety of the different types/classes. Medical-Class IST seem to be the best bet, as they can be easily modified (which is what we’ll be using).

There has also been an onslaught of DIY isolation transformers, where people are using types of transformers never intended for isolation. This can lead to dangerous results and if you aren’t qualified, please don’t attempt to make one. Even if you are qualified, you can buy a brand new M-Class IST, from a brand such as Triad, for $150 (DigiKey).

Other Important Notes of Interest: Grounding
Neutral vs. Ground:

The neutral wire or “grounded conductor” is a normally current-carrying conductor, similar in many ways to a phase wire in that it will carry the same amount of current in single phase system.  It is grounded (connected to the earth) at the center-tap point of transformer for a number of very complex reasons.  The neutral conductor is isolated from accidental contact because it is a very dangerous normally current carrying conductor.

The ground wire is a normally non-current carrying conductor, designed to carry the electrical energy should a fault occur.  This normally non-current carrying conductor is also bonded to every exposed metallic part in your structure, to ensure that there are no hazardous differences in potential on the objects you touch.  You should be able to touch exposed ground wires with no concern, although all exposed wires should always be treated as live and dangerous.,energy%20should%20a%20fault%20occur.

Let me point out something, the average engineer, hobbyist, technician, scientist, IT specialist, etc., does not need to probe the raw mains (with a grounded oscilloscope). Those that do, almost never actually probe the mains directly. If for some reason, this need arises, the best bet is to use a variac to adjust the AC output + an IST (after the variac). The variac alone will alter the signal and the IST surely won’t help, but it will make it safe and provide you a good idea of the real-time behavior from the outlet.

Please, do NOT probe the raw mains with a grounded oscilloscope and no variac. It’s incredibly dangerous and the waveforms you’ll see are not worth the risk (your life)..

Here are some common reasons to probe with a scope across a DUT’s power inlet:

Designing boards for power delivery
Electron Tube Based Equipment Design/Repair
Power Supply (Switching, Linear, etc) Design/Repair
Transformer Design
Building, repairing, upgrading and diagnosing the mains

If you are not qualified and do not have a specific reason you are doing this, just don’t do it. It’s cool to look at the grid itself, I get that. However, it’s EXTREMELY dangerous (without proper precautions) and people have been killed.

You can check an outlet with a multimeter, that is fine, but a scope is another animal. This is due to the fact the oscilloscope’s negative terminal/probe (alligator clip) is ground referenced. Due to this ground reference on the BNC, you can easily cause a short-circuit, if you connect the probe ground to a non-Grounded node. (Note: This is a totally separate issue from the danger of probing the raw mains) This leads to possible lethal shock to the user and fire/explosion hazards (on the probe ground clip).

Normally, this alligator clip is what will blow up, if you tried to give it 15A out of the mains. There are probes made for doing that, but I am not going to go into that, due to the dangers involved. This article is mainly to touch on these topics and give you the references below to dive in deeper.

You can get around the dangers to the scope (itself), by using differential probes.

This article is for those who understand the principles of electricity and the way that device grounds interact with each other. As well as, how to hook the equipment up, once the IST is safe to use (i.e. don’t hook a scope or variac into the IST output, only the Device Under Test goes in there.

K. Hallman

Fundamental Concepts of IST

Fundamentally, an IST is a 1:1 ratio transformer, although this is not always true. In fact, you can find multi-tap IST designs, with step-up and/or step-down configurations on the same bobbin/core, while maintaining full isolation. These step-up/step-down ratios are calculated by the number of winding, between the primary and the secondary.

The main difference in an IST and any other transformer, is that all 3 primary connections (from the mains), are meant to be totally isolated from the secondary. The issue we run into is that many ISTs have the common ground of the secondary, connected directly to a chassis ground (which connects directly to the both transformer winding). One solution changes the secondary from directly connected, to an RC safety network (more below)

One of the rules, with a (true) IST, there can only be one plug on the output. This is mainly for safety reasons, so that users don’t accidentally bypass the very isolation that is provided (by mistakenly connecting the grounds of a DUT and a piece of Test Equipment (which connects to ground with a probe)). This is why I will be removing all the IEC outlet and replacing it with a single Journeyman-Pro 15A chassis mount outlet (below).

You might ask, “Why buy a medical transformer for this, when Triad and TrippLite are selling them on DigiKey and Ebay brand new, for $100-$150?” The first reason that jumped out to me was the leakage rating of medical class transformers, they are less than 100 µA (in a worst case scenario/load). My transformer has a little sticker with “40.3” written on it, so that was probably my max leakage reading at the factory (40.3µA). Typical leakage is likely more in the range of 10µA

Medical-Class also can come with double wrapped windings, further reducing losses and improving product-life. With the double enamel coatings, in the case that one should come loose, the other will still be there to hold it. Enamel coming loose, is often what induces vibration and, the famous, 60Hz hum.

Back to transformer cores

Medical-Class comes with potted cores (or the whole transformer is potted). Potted means it has been filled/covered with some sort of epoxy, which is essentially a sort of fiberglass. The higher end brands often coat/paint them, once potted with epoxy.

  • You then have grounded inter-winding screen or electrostatic-shield. These are commonly seen on HV-Transformers and the higher-end ISTs.
  • Some of designs come with physical Faraday shields, in addition to their potting and electrostatic shield.
  • I heard that some companies ground the core of the transformer, but have yet to confirm this is true in an IST, I know it’s true for a HV-Transformer.

Note: the core/screen and Faraday cage, should never be disconnected, as these are not equivalent to chassis ground (in the eyes of the outlet and DUT).

Note: The below quote is in regard to High-Voltage transformers, but it still applies here; 15 amps is nothing to sneeze at.

Because of the construction of the transformer core/winding assembly, there can be high potentials induced into the core, causing partial discharges which will damage the transformer. The core ground diverts this voltage safely to ground. The core ground also provides a low-resistance path to ground if there is a short circuit between the winding and the core, allowing protective relaying to detect it.

There are a few other features, that are often, only found in the medical series isolation transformers. In my Toroid Corp M-Series, I have a thermal fuse block. This is rated for 10A and set to a specific thermal load temp, at which point the connection is severed. This is to protect against over-voltage conditions, such as a short circuit (which overloads and overheats the transformer windings). There is a 5-pole EMI line-filter and in-rush current protections included in the IEC inlet.

As you can see, there are many reasons why myself and others, pick (used) Medical Class transformers, over what can be found at retailers new.

Option 1:

Is to cut the ground connection between the transformer’s secondary and the chassis. This will effectively “float” the ground, but now the chassis loses all electrical protections (required by some countries). However, if you look on YouTube for “Modding Medical Isolation Transformer for Lab“, this is exactly what you will find. Some say this is totally fine to do, as long as you are only putting the DUT (device-under-test) on the isolation transformer, not your measurement equipment (like an oscilloscope).

What you are losing, if you “float”/remove the secondary ground:

By connecting a green ground wire from the metal frame of the refrigerator, if the chassis inadvertently becomes charged for any reason, the unwanted electricity will travel through the wire back to your electrical panel, and tripping the circuit-breaker stopping the flow of electricity. Additionally, that wire must be connected to something that is in turn connected to the earth or ground outside.,energy%20should%20a%20fault%20occur.

One user on YouTube added in banana jacks, then ran the transformer ground to one jack and a jumper to the chassis ground to the other. This was so he could quickly reconnect the chassis ground if he wanted. I’d say that is a good idea, if you decide to go with option #1. You could always stick a fuse between those two points as well, providing some safety (go with fast blow..)

Audio Isolation 1:1 Transformers (obviously incorrect)

You can find 1:1 isolation transformers on EBay for about 25 cents. However, these are signal isolation transformers (and usually terrible to use, in almost all HiFi circumstances). However, a real 1:1 isolation transformer will weigh at least 10-20 lbs. The one I purchased is a transformer made by the “Toroid Corporation of MD” in 2006. In terms of condition, the unit is nearly spotless and it looks like the plugs were barely ever used. If they did use them, they didn’t repeatedly plug and unplug them. The inside is clean enough to eat off of, nearly any dust even in there.

The problem with this price range, nearly every kind of isolation transformer you can find for under $200-$300 is going to tie the ground to the chassis (medical isolation transformers). Or it just won’t include 3 prong plugs at all (many Tripp Lite products are like this). They likely did this due to laws in many countries, regarding floating grounds. We are going to look at that exact topic.

The main reason not to float the chassis ground is the fact the chassis can become “hot” and you can actually get hit by an in-rush current, just by touching it. The reason for this, is because you can (and do) have high transient spikes on the grid (as well as your isolation transformer) and they have to go somewhere. Usually they are dissipated to ground, from all major points in the circuit.

With a floating ground, you have essentially made the chassis one side/plate of a capacitor, which is slowly accumulating charge. You are the other (low-side) plate. If the chassis plate builds up enough charge, it’s going to be a bad day, for our “plate”.

K. Hallman

Now, in most medical class transformers, you will see a 2nd ground. This is an additional feature to improve performance (EMI rejection), where the core of the transformer is directly tied to the chassis (ground). As stated before, you do not want to remove the grounding feature of any additional performance or safety features of the transformer. This includes the inter-winding or electro-static shielding.

A true, fully isolated, isolation transformer should only have a single plug connected to the transformer secondary (at any one time). This plug ground, should not have any direct connection (i.e. 0 ohm) to the general chassis ground (in AC or DC states), for a 3-prong device’s isolation.

Instead, you’ll want to use an RC safety network. Providing good safety, especially with an Insulation Monitoring Relay watching the network for resistance shifts or network failure (see below). This gives us two independent ways that the circuit should disconnect, which is what we want (for safety). Then we have our backups, which are fuse and thermal related. Worth noting, the fuse and thermal safety capacitors likely won’t be enough to save you from a critical fault, as they have a longer failure delay than our RC network + IMR. If you are smart about it, you can have that IMR pick-up a large current spike and cut off, because current will shift the resistance (R = V/I). You don’t have to wait for the network to blow. The only thing that may be a problem is large in-rush currents, if so get a soft-start circuit to add in.

For any other questions, feel free to ask or see the references below. Above all else, be safe. If you aren’t confident on anything, this isn’t the project you want to do. I haven’t finished mine yet (partly due to time, partly due to research + caution) and I am a Computer Engineer.

K Hallman
Fuck it, Ohm's law compass : PoliticalCompassMemes
Ohm’s Law Formula Wheel

Anything other than this, is not a “true” isolation transformer, it’s more like what you’d see on medical equipment, where all equipment is tied together through a very low impedance grounding plate.

Instead of directly floating the ground, I found a 2nd option, that sounds slightly more intelligent.

Insulation Monitoring Relay (CM-IWS.1)

Option 2:

Use at least a 1 MΩ resistor, capable of dissipating any inrush of high current. To be safe I’ll probably use a 2W or 5W. AC Power calculations for 1 MΩ resistors, vary by location, but it should be below a half watt; I’d still go 1-2W minimum. Then, add a low valued Y safety-capacitor (<10nF) between outlet ground and chassis/ground.

To protect against a major fault/failure in the RC network, between the plug and chassis, I recommend an Isolation Monitoring Relay (IMR). This relay can be dialed to a specific minimum resistance and if the relay detects any resistance less than that, it will trip the breaker. This is BY FAR, the best way to ensure the safety of Option #2. I found an ABB CM-IWS.1 (lightly used) for $50 on EBay (The seller accepted a $40 offer, so you might could get it even lower).

Info on Y-Safety Capacitors:

These two components are to be wired in a parallel configuration, on the secondary ground. The capacitor will help add some protection, with some added RF filtering. I recommend using a fusible safety resistor, as well as a Y2 rated safety capacitor. By combining these two measures, you should be much better off than directly connecting or floating, ground.

Toggle Switch for Ground “Float”/IMR-Mode

This should be a good middle ground, between not isolating the ground at all (which is not a good option, for use on an electronics bench) and floating it. I’m going to have to do some testing and some more research to be sure of my setup, but I wanted to share what I had so far.

You can limit the dangers, when doing testing of your new isolation transformer, by putting a variac before the isolation transformer. This allows you to test your isolation at much safer voltages. Plus, many times you won’t need the full 120-125V, so it’s common to pair these up.


Join the discussion and let HL know what you think!

This site uses Akismet to reduce spam. Learn how your comment data is processed.