Hi Folks...

I'm hoping someone can clear my confusion about floating voltages and what appears to be stray/parasitic side-effects. Included are very simple diagrams/images that represent the environment and data.

It's a simple 10:1 transformer, full bridge rectifier and storage capacitor. As you can see, everything on the right side of the transformer is floating which is my desired configuration. Various points in the circuit are labeled. Point "E" is a connection to Earth ground from a grounding point on the wall outlet.

Please see the waveforms when measuring from point E (ground) to various other points in the circuit. I strongly suspect these waveforms are due to parasitic effects from the transformer. If different transformers are used, the waveforms remain similar but with different amplitudes. I've looked at various circuit models for transformers and understand the concepts of leakage capacitance, leakage inductance and mutual capacitances between the windings. I'm having a hard time getting my head wrapped-around which of those side-effects are causing the observed waveforms. Can anyone describe the source and path that leads to this? The second part to this question... how to eliminate those side-effects while still maintaining a floating voltage?

BTW: There's not a lot of power in that signal. It's easily shunted to ground by placing a 1uF cap across points C-E or D-E.

Your thoughts / help is greatly appreciated.

Regards

Ray

 

Welcome to AAC!

What you are observing is 60Hz AC pickup. This is normal. This is not stray/parasitic side-effects.
Simply touching the oscilloscope probe tip with your finger will give the same picture and has noting to do with your circuit.
(Unplug your circuit and the 60Hz signal will still appear on the oscilloscope.)

What you need to do is measure the signal between Points C and D.
There are two ways to do this. You can use two probes and measure the differential signal.
You can also connect the probe to one point (C or D) and the probe GND clip to the other point.

 

Yes, because of the very high impedance of the 'scope input, you are seeing the voltage from stray capacitive and magnetic coupling (some of which you can see by just having a piece of wire connected to the probe in the air).
If you put a low resistance (e.g 1kΩ) from the probe to the 'scope ground, what you see should be greatly reduced in amplitude.
If not then you may have a problem.

 

Just connect the scope ground to point D.

 

Hi MrChips, crutschow and Dodgydave...

Much thanks for the replies and welcomes to this site. ... Yes, in my dabbling in electronics, I've seen the AC waveform once or twice before.

No doubt, the folks outside North America experience a 50Hz version of the same thing. Indeed, when measuring with a 1x vs 10x vs differential amp probe, the waveform changes a good amount due to varying circuit loading.

In the case of this particular circuit, the waveform is very pronounced and changes considerably when toroidal vs EI transformers are used. As a mental exercise, I'm trying to figure-out what elements of a theoretical model of the transformer, combined with the rest of the circuit (and/or power supply in the 'scope) are mainly at-play. This my "deep-dive" so I can stop writing-off that waveform as parasitic, stray, coupled noise from some mysterious source. Just looking for some clues on how to unravel this mystery. In addition, I'd like to find-out if this problem is preventable vs using a cap or high value resistor and dumping it to ground. Putting caps to ground has its own set of problems. I've tried various snubber circuits within the isolated part of the network -to no avail.

Thanks... Ray

 

Did you read my post #2?

The 60Hz AC appearing on the oscilloscope screen has nothing to do with your circuit, toroidal or EI transformer core.
Disconnect the transformer from the AC mains and the signal is still there.

If you want to make the signal go away you have to go to the middle of an open field far away from any power lines or transmission towers. By the way, you also have to use a battery powered portable oscilloscope.

What you see is the fact that we are bombarded with 60Hz AC radiation in our modern industrialized society.

If you connect the GND of your probe to point-D, the 60Hz AC on the oscilloscope will go away, as others have already stated.
There is no need for any snubber circuit.

 

Did you read my post #2?

The 60Hz AC appearing on the oscilloscope screen has nothing to do with your circuit, toroidal or EI transformer core.
Disconnect the transformer from the AC mains and the signal is still there.

If you want to make the signal go away you have to go to the middle of an open field far away from any power lines or transmission towers. By the way, you also have to use a battery powered portable oscilloscope.

What you see is the fact that we are bombarded with 60Hz AC radiation in our modern industrialized society.

If you connect the GND of your probe to point-D, the 60Hz AC on the oscilloscope will go away, as others have already stated.
There is no need for any snubber circuit.

Hi... Yes, I read your post. Thank you!

When the transformer is energized, the voltages (as you can see from the images) are around 12 to 15V RMS AC. When unplugged, the voltages drop way down around 30-40 mV (RMS). That's a difference of about 50dB. Using different transformers, I've measured differences ranging roughly from 20 to 40dB -which is enormous. Honestly, I don't know the precise differences because probe loading varies the values considerably.

I conducted the same experiment by measuring some various switching power supplies. In one case, potential from ground to one of the conductors was 156 volts. It dropped to 50 mV when unplugged. You might be thinking there's something wrong with my house ground... -Nope. I've tested this at home and in different locations. I also borrowed a handheld (battery powered) scope and got similar results.

Yes, I know that touching a probe will produce a similar result -usually about 2-5V RMS. There's stray AC propagating thru the air and we're just antennas. I get it. What I don't understand is, what specifically about transformers causes 20 to 40dB amplification when they are energized (some kind of resonance perhaps?). Finally, I've taken good quality lab power supplies (with isolated output) and performed the same experiment. The readings from ground were very low -around 50mV or less as I recall. It was a borrowed lab supply and I could not take it apart to see if there was any obvious magic inside.

I noticed this a few decades ago and just brushed it off -but now I'm curious and wish to understand the details.

Thanks...

Ray

 

You are overthinking this.
This is not a problem with which you need be obsessed.
Connect the probe GND clip to point-D and the probe tip to point-C.

 

Think Ohm's Law.

I = V / R
V = I x R

The scope probe impedance is 10MΩ
If the probe picks up 1μA stray current, it registers as 10V.
10μA becomes 100V.

 

You are overthinking this.
This is not a problem with which you need be obsessed.
Connect the probe GND clip to point-D and the probe tip to point-C.

Yeah, I wish it was that easy to just forget about it. Here's the back-story... I regularly use some buck converters to which I make some minor improvements; furthermore, I designed and fabbed a board to filter the last little bit of switching noise from that buck converter. The board has a CM choke and a differential filter to tame the noise -other things on that board too. Depending on what lab supply I use to power the whole thing, the final pole-to-pole noise coming out is around 1.2mV RMS. Now, I wish to build an integrated PS (instead of a lab supply). Using simple AC rectifiers like the one shown above, increases the final noise to about 5mv RMS. Toroid transformers produce the best results so far.

Sorry to say but, at this stage of the game, I'm hell-bent on knowing what's going on. Yes, I could put a linear regulator in there and be done with it but, I'd like to avoid that if possible.

Ray

 

Did you do as suggested?
Connect the probe GND clip to point-D and the probe tip to point-C.

 

Did you do as suggested?
Connect the probe GND clip to point-D and the probe tip to point-C.

Oh yeah, that's very clean.

 

Oh yeah, that's very clean.

Well, what a surprise!

 

Most of what you are seeing is caused by the capacitive coupling from the primary to the secondary windings of the transformer. Some power transformares have a grounded copper shield between the two windings to reduce this effect. It really is of no consequence unless you are designing very sensitive circuits or medical equipment. It is good practice to connect one of the power rails of a circuit to ground to avoid a static charge build-up, which could damage sensitive components if it is discharged to ground. It also gives a common reference point for making measurements.

 

If you are building high impedance or high gain circuits and are concerned about this, wire the transformer primary with neutral connected to the end of the winding closest to the secondary. That will give minimum signal to be transferred to the secondary by capacitive coupling. If you don't know what the primary connections are, measure the voltage from the secondary to ground with the power connected each way. Make sure the transformer frame is grounded too.

 

If you are building high impedance or high gain circuits and are concerned about this, wire the transformer primary with neutral connected to the end of the winding closest to the secondary. That will give minimum signal to be transferred to the secondary by capacitive coupling. If you don't know what the primary connections are, measure the voltage from the secondary to ground with the power connected each way. Make sure the transformer frame is grounded too.

Much thanks. The transformer data sheet shows the Dot notation and color code. I'll take some measurements and try different ways just to be sure. Grounding the frame is not possible on this one; its a toroid.

Thanks again... Ray

 

Ray you have brought up a very interesting topic. There are some great responses as well. I think we are still dancing around the specific cause and I think that’s why you are having a hard time feeling settled with the answers. I think this is a fantastic subject for a video! I will post a video this week

 

One thing to keep in mind is that as soon as you connect your scope probe's ground to any point in the secondary of the circuit, that circuit is no longer floating -- it is now ground referenced, via the scope, to the earth conductor of it's power cord (unless the scope is battery powered, of course). If you NEED the circuit to be floating, then do NOT make this connection. More to the point, if you do make this connection, be very sure that there is no point elsewhere in the circuit that is somehow, intentionally or otherwise, referenced to earth ground. If there is, then at best you have created a ground loop situation that can wreak havoc with noise pickup, and at worst you have created a short across part of your circuit that could potentially (no pun intended) have disastrous results. If you need the circuit to remain floating, then you need to use a battery powered scope or a scope with differential inputs (neither common nor cheap) or a suitable differential amplifier between your circuit and your scope.

 

Ray you have brought up a very interesting topic. There are some great responses as well. I think we are still dancing around the specific cause and I think that’s why you are having a hard time feeling settled with the answers. I think this is a fantastic subject for a video! I will post a video this week

Welcome to AAC, Kiss!

Just to let you know, AAC is a privately owned enterprise. There are two sides to AAC which began as a free not-for-profit open platform for electronics hobbyists and professionals alike. The owners have ventured into the for-profit commercial advertisement side and are not too keen to allow other parties to promote there own enterprise.

 

Ray you have brought up a very interesting topic. There are some great responses as well. I think we are still dancing around the specific cause and I think that’s why you are having a hard time feeling settled with the answers. I think this is a fantastic subject for a video! I will post a video this week

Since site- and self-promotion aren't allowed here, be sure to make this video as a blog entry or, perhaps better yet, contact EETech about making an article/video on the commercial side of the house (you can then get paid for it, too!). If you just post a link to a YouTube video (or similar), we will have to delete the link.