Scoping high side transistor of Two Transistor Forward.

[cupoftea]

When scoping the gate of the high-side transistor of an offline Two Transistor Forward (2TF),
I usually use a DIY (x1) coaxial probe, but with the 2TF powered through an isolation transformer,
so that the probe doesn't blow up the 2TF.

However, this means that the switching node is clamped to earth potential,
and then the entire converter's circuitry is banging up and down around that.
Hardly ideal for noise reduction.

The only alternative is to use a diff probe...but, with their hopelessly straggly
long leads, the noise on the scope is then massive.

Do you have a better way?

 

[danadakk]

There are high speed HV diff probes with very low L signal path
probe connections :

1 GHz and 500 MHz High Voltage Differential Probes

Key performance specifications1 GHz and 500 MHz probe bandwidth <1 pF differential input capacitance 1 MΩ differential input resistance ±42 v (DC + pk AC) differential input voltage >18 dB CMRR (at

www.tek.com

There are also DIY fast probes, just google "DIY Differential Probe"


Regards, Dana.

 

[betwixt]

The only alternative is to use a diff probe...but, with their hopelessly straggly
long leads, the noise on the scope is then massive.

I'm not sure why the 'straggly long leads' comment. Two normal scope probes are all you need and if you use x10 attenuation on them the loading and effect should be insignificant and not introduce noise.

Brian.

 

[danadakk]

On modern DSOs, the $500 kinds of scopes, the channel to channel isolation
is ~ 40 db or so. Pile on that probe to probe G accuracy, +/- 3%, and Z
differences that method has limitations.

Regards, Dana.

 

[FvM]

I understand that the question is referring to high common mode voltage of mains connected switcher, e.g. 400 V.
The mentioned Tek differential probes are only suited for 40 - 50 V CM.

Best non-optical isolated differential probe I have used so far is Keysight N2804A/N2805A. https://www.keysight.com/de/de/product/N2804A/high-voltage-differential-probe-300-mhz.html
Common mode rejection is still insufficient for SiC switchers according to my experience.

Classical 100 MHz HV differential probes like Testec SI9110 are often used for similar measurements, typical CMRR is at least 20 dB worse than N2804A, it's almost impossible to see exact gate drive waveform (overshoot, ringing) of high side switcher.

Best option is optical isolated scope, e.g. Cleverscope CS448/CS548 or fibre isolated probe, e.g. PMK Firefly.

 

[danadakk]

I understand that the question is referring to high common mode voltage of mains connected switcher, e.g. 400 V.
The mentioned Tek differential probes are only suited for 40 - 50 V CM.

Best non-optical isolated differential probe I have used so far is Keysight N2804A/N2805A. https://www.keysight.com/de/de/product/N2804A/high-voltage-differential-probe-300-mhz.html
Common mode rejection is still insufficient for SiC switchers according to my experience.

Classical 100 MHz HV differential probes like Testec SI9110 are often used for similar measurements, typical CMRR is at least 20 dB worse than N2804A, it's almost impossible to see exact gate drive waveform (overshoot, ringing) of high side switcher.

Best option is optical isolated scope, e.g. Cleverscope CS448/CS548 or fibre isolated probe, e.g. PMK Firefly.

Fascinating, these will do the CM range but CMRR still quite poor :

High-voltage Differential Probes

High-voltage Differential Probes

www.tek.com

Then there is this :

https://cdn.teledynelecroy.com/files/pdf/dl-iso-probe-datasheet.pdf

1 Ghz, but even this CMRR 30 db at 100 Mhz in the tank, 1000V CM.

None of these a human can buy due to cost.

 

[Easy peasy]

If you have a quality scope and good 100x probes - you can connect the probe tips to upper gate and source and use the subtract function to get the gate signal - scope grounds then go to zero volt on the psu.

alternatively if using a diff probe with long leads, put some CM ferrite clamps on the leads - but more importantly also solder some leaded 4k7 ( or 10K or 22k as suits ) resistors to the G & S and hook onto these with the diff probes - this often gives a better look at what is going on and gives you 4k7 CM isolation straight away.

( note for idiots - do not solder these R's across the G & S )

 

[cupoftea]

Thanks, will try it. Also just wondering about doing a ~12v high side supply, and putting an opamp diff amp into the high side...and via say 20MEG input impedance resistors (pair).....feeding that back to the scope and setting scope.....i would use coax to bring the signal back to the scope. Kind of like a homebrew diff probe for a lot less cost.

A 10X 100MHz Differential Probe

To make...or not to make...that is the question. I needed a probe that can dig out a 1V differential signal riding on a 60Hz 125VAC mains voltage. Many of the commercial probes aren't suitable for my purposes and are pricey. Therefore, the answer is...make.

hackaday.io

Wondering about doing a stripboard/PDIP version of the above....prefer PCB but it will cost.

(ignore up to 08:30)

 

[Easy peasy]

As the Dilmah man says: " Do try it " and report back to group.

 

[cupoftea]

Thanks, all the diff probes on the web, even the DIY ones, seem massively over-spec'd for our needs, i mean, all we need to do is get a decent look at the high side drive of a 2 tran forward that has vin up to some 373V.
I will start with the bog standard opamp diff amp and work up from there.

 

[Easy peasy]

your bog standard diff amp will be a not very useful expense of time and effort.

 

[cupoftea]

Thanks, by bog standard, i mean along these lines.

Building a Differential Amplifier Probe

DIY 100MHz Differential Probe

www.paulvdiyblogs.net

...eg "topology 2" which is near-ish the top...it just uses three TL072's.

As you know, we dont need much accuracy, just to generally look at the upper gate drive rise time and ringing.

 

[cupoftea]

The attached (LTspice and PNG schem) shows a very cheap, "pretty pathetic" differential probe, which is never-the-less, perfectly
satisfactory. Satisfactory to tell what is going on at the high side gate drive of a two transistor
forward converter which has Vin = 373V. It gives a "divide by 36" ratio, but shows if the ringing
on the fet gate is above Vgs(max). Also, it shows the rise time of the upper fet gate-to-source.

Also, unlike using two scope probes and the subtract function, this one makes for better signal to noise
ratio because at pretty much all points, the signal path goes as tight "go and return" all the way back to the scope.

We could pay some £800 for a diff probe offtheshelf to do this, but the attached does what we need perfectly fine.
Would you agree?

In fact , the OPA891 opamp would be better for this application, and only costs $2.5
Also, it has rails to +/-18V so we wouldnt have to divide the gate drive voltage down so much...so we could make it a "divide by 12" diff probe , which is better.
OPA891....
OPA891

 

[Easy peasy]

respectfully, there is a lack of compensation for the " coax " and for the low pass filter formed by the 100k R's

 

[cupoftea]

Thanks, and i believe as such, it needs some small capacitance across the 100k's to compensate for the RC filter effect?

 

[Easy peasy]

are your 100k's rated for 400 volt ?

 

[cupoftea]

Thanks, yes they are.

 

[Easy peasy]

TL072 is only 13V/uS, this is not fast enough to see real gate information - there are much faster op-amps out there.

--- Updated Jun 27, 2024 ---

even the OPA891 @ 10V / 100nS may not be good enough to see everything ( gain of 1 @ 180 MHz, ), when you look at the graphs for 100mV signal out ( -3dB @ 220MHz ).

 

[cupoftea]

Thanks, and as we know, it is this very reason, which makes people drive top and bottom FETs of a two tran forward with a gate drive transformer...ie, so that they can scope the bottom FET gate, and infer that the top one is the same , it must be, because it is driven by the same componentry.

 

[Easy peasy]

it's not quite the same due to the currents flowing in the transformer capacitances - however very often it is close enough.