diy differential probe?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I need a differential probe for my scope to be able to measure both high and low side gate waveforms ín a UcD simultaniously, so I thought I might ask in this forum.
Does anyone know of a design that helps me accomplish this? The bottom end signal rides on 0-5V compared to ground, and the high side rides on 0-100V compared to ground, with the offset being a square wave of 550 kHz with sharp edges. I guess I need to insert whatever I build in both channels to get equal delay since I am interested in the interaction between the waveforms.
 
How many channels do you have on your 'scope?

A four channel scope with matched probes can usually be used to make "quasi floating" differential measurements.  Common mode rejection can be enhanced by wrapping the cables of two identical probes in-hand several times (half dozen or so) through a large ferrite toroid.  Of course, the scope must be set to difference (A-B) mode.  Verify the measurement quality and integrity by first connecting both probes to the same point on the source of the floating MOSFET (grounding clips are not used).  If all is well, no signal will appear (= perfect common mode rejection).
 
You can build two pulse transformers and connect the primary of each one between the gate and source of each output device. A small value series coupling capacitor in each primary is also required as gate drive waveforms contain DC and audio frequencies that otherwise would saturate the cores.

When using such a setup, it's preferable to feed the amplifier only with low frequencies, as any quick change in duty cycle is probably going to saturate the cores even with the capacitors.

Also, in self oscillating circuits the transformers have to be sized to accomodate the flux produced by the lower expected operating frequency (either that, or avoid any high amplitude signal that would cause the frequency to fall considerably).

BTW: If I was you, I would take the output H/L signal from one of the legs of the comparator and then I would use a quad XOR 74HC86 to derive complimentary gate drive pulses with fixed dead times and equal propagation delays.
 
I have constructed pulse transformers from unknown scrapped ferrite transformers from old PC PSUs using a large number of bifilar turns. The larger the number of turns the less the influence on the gate drive waveform as compared to without the transformer, I guess my drivers dont like parallell inductance. Now to the strange part - even though I use an exact match in the number of primary and secondary turns, I get about 10% higher amplitude on the secondary side but with no visible distortion. Could this be because of the ferrite material I use, or does this have any other explanation? I tend to believe the unknown ferrite causes by having non-constant permeability over the operating frequency range.
 
Ex-Moderator
Joined 2005
analogspiceman said:
How many channels do you have on your 'scope?

A four channel scope with matched probes can usually be used to make "quasi floating" differential measurements.  Common mode rejection can be enhanced by wrapping the cables of two identical probes in-hand several times (half dozen or so) through a large ferrite toroid.  Of course, the scope must be set to difference (A-B) mode.  Verify the measurement quality and integrity by first connecting both probes to the same point on the source of the floating MOSFET (grounding clips are not used).  If all is well, no signal will appear (= perfect common mode rejection).


Won't that work with just a two channel scope operating in differential mode (probe grounds at the power supply ground)? I've measured bridged outputs of an amp using that technique before and it worked.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.