Getting less noisy scope traces on our Boost PFC?

[cupoftea]

Hi,
We wish to probe the signals/pins on our L4981B PFC controller while its driving a 1kW Boost PFC. We have a Micsig DP10007 diff probe, but, due to its long dangly probes, the signals are enevitably hopelessly swamped in common mode noise.

We don’t have an isolation transformer, and cant afford one.

I know about the serious dangers of floating our scope…but I believe the only way we can get good clean signal traces on the scope is to use a home-brew coaxial probe and solder it to the probe point on the PCB, and also literally cut off the earth lead to the scope, and then probe the circuit like this. We are aware of the potential highly serious danger to health of doing this, but we must do it.

Do you agree that the scope signal traces, when using the coax probe, will be just as “clean” as if we had used an isolation transformer, and not removed the scopes earth wire?

 

[Warpspeed]

You could try winding your cro probe lead through a large high permeability ferrite toroid.
Something with a fairly high Al (inductance per turn). Lower frequency ferrite material preferred.

That can help considerably to reduce common mode noise, as well as adding some series impedance to the ground connection through the probe shield.

A bit of experimentation will be required, but a very worthwhile improvement is possible.

 

[cupoftea]

Thanks, the ferrite torroid sounds good...doing that in conjunction with the above cutting of the scope's earth lead....with the homw-brew coax probe being taken through the torroid.
In fact.....i'm thinking of having the earth wire connected to the scope...but via a 1nF series y capacitor..........so that the earth is "there in effect"...and this could be the trick to clean up the signals?

i suspect a scope with earth wire cut away means noisy scope traces?

 

[Warpspeed]

It would be much safer to give the mains cord to the cro, the same treatment as the probe.
You might like to try the wound steel toroid from a 50//60Hz mains transformer with the original windings removed.

Similar concept, but it works down to a much lower frequency. Maybe one full layer of three core plastic appliance cord. You will probably need to make that up, as the length of mains cord may be significant, and neither plug is likely to fit through the hole for winding.

It puts a bit of high frequency common mode impedance into both the mains ground and the 120/240v supply without compromising the safety ground. As above, a bit of experimentation may be called for.

The trick seems to be to get a high inductance per turn. That means aim for a large magnetic cross section and shortest magnetic path length. When choosing a victim toroid, fat with a small hole is the way to go if you can find one. One way to cheat in achieving this, is to stack smaller toroids.

 

[cupoftea]

It would be much safer to give the mains cord to the cro, the same treatment as the probe.

Thanks, but we must cut off the earth lead in the mains cable to the scope......... since as you know, otherwise we cant attach the gnd of the home-brew coax probe to the primary side.

If anyone can give us a heads up as to what the noisiness of this method pf probing would be, then i would be much appreciative. The new PCBs will arrive in 6 days....will take me some 4 days to kit and build them, then there will be inevitable debugging where we will need to see signals clearly, and so we are definitely going to cut off the earth wire and probe like that, with the nice home-brew coax probe....we dont have a choice as we cant afford an 1kva isolation transformer.

Unfortunatley the entire first board had to be scrapped as the UCC280780A's were ESD damaged, and when the control board was removed to replace them, the PTHs coudlnt be cleared of solder, so control board couldnt go back properly.....plus the board was difficult to probe, so the nxt board is built for test....but these events (all 3 purchased UCC28070A's dead-on-arrival from ESD) have totally wiped out our budget.

Does anyone know if non-earthed-scopes give more noisy scope traces? Just so we can get a kind of heads-up.

 

[danadakk]

Isolation safety :

https://download.tek.com/document/3AW_19134_2_MR_Letter.pdf

https://www.eeweb.com/wp-content/up...power-supply-noise-measurement-1290016135.pdf

https://assets.testequity.com/te1/Documents/pdf/handheld-floating-measurement.pdf

You can isolate a scope with an isolation transformer, or use a car battery, a cheap auto car inverter
to provide the scope power. Those inverters can be noisy, sometimes a line filter appropriate.

DIY diff probes :

High Voltage Differential Probe for Oscilloscopes - PCBWay Collaboration​

High Voltage Differential Probe for Oscilloscopes - PCBWay Collaboration: In the last ten years, the power electronics industry has been growing at an incredibly fast pace. However, despite this trend, articles and projects in the maker community regarding this topic are very scarce. This could...

www.instructables.com

DIY Cheap HV differential oscilloscope probe - Page 1​

DIY Cheap HV differential oscilloscope probe - Page 1
www.eevblog.com


Regards, Dana.

 

[Warpspeed]

The old fashioned way to make a differential measurement was to use two cro channels, and select "difference" mode, A minus B. With a modern high performance oscilloscope and carefully compensated probes, it may be almost be usable.

 

[danadakk]

As Warpspeed mentioned :

Regards, Dana.