Isolation of the RIGOL DS1054Z from the mains with three transformers

[Edison]

Because of physics works 

Well, physics may work, but the head doesn't - a floating oscilloscope is nonsense - there is a filter on the power three-wire socket that is connected to the ground, and in that case the supply voltage will reach the frame of the oscilloscope. The oscilloscope may be more susceptible to interference from the environment. The device under test is connected via an isolating transformer. If you have a necessary need, buy a battery oscilloscope or an oscilloscope that has galvanically isolated inputs by the manufacturer (for example Tektronix 🤔 but for that price you have about twenty of your own)
Well, if the head works as it should, then when a galvanically isolated measurement is needed with a classic oscilloscope, a normal galvanically isolated differential probe is used.

[bdunham7]

a normal galvanically isolated probe is used.

Can you provide an example of one?  They typically either have significant performance limitations or are very expensive.

[Edison]

Differential optical probes measure from DC to GHz, but you're right, they are expensive, the cheap Chinese ones work wonders, so I solved it with a 2x 100Mhz battery oscilloscope, which is enough for my purposes

[Edison]

I've four RIGOL PVP3150 Passive Probes so it's easy to make one differential probe out of two by connecting the crocodile clips together 

     In this type of measurement, the ground clamps are not used at all, thanks to the activated mathematical function, the signal from both channels is calculated, one probe with the tip, for example, on GND and the other on the test pin

The differential probe is galvanically separated - two probes against each other are not - it uses mutual isolation of the active part of the oscilloscope input

[Edison]

wating for the 4 5 6 transformers ideas  pffffffffffff

Don't be surprised, the more transformers, the more Adidas 🤣

[Edison]

Professionalism means attention to detail. 

If you do not understand from what has been written here by everyone that the isolation transformer should not be connected to the power supply of a classic oscilloscope, take your oscilloscope, open the window and throw it out of the window and the transformers behind it - that will be the most professional thing you can do

[coromonadalix]

loll  i'm waitning for the op answer to that   loll   this is a joke   

[bdunham7]

The differential probe is galvanically separated - two probes against each other are not - it uses mutual isolation of the active part of the oscilloscope input

Differential probes are not truly galvanically isolated, they have an input impedance spec to ground of xM + ypF and a probe-to-probe spec of 2xM + ypF/2.  The circuit doesn't look much different then two probes on a scope.  That was sort of my point--true galvanic isolation comes at steep price.

[Edison]

sorry I left out the word optical

edit: my heart almost stopped when I first saw the price of a galvanically separated probe up to 1GHz 😂

[Andy Chee]

30 years ago there was a Silicon Chip project for galvanically isolated differential probes based on the IL300 linear optocoupler.  It only had a measly bandwidth of 100kHz, so it's almost useless in today's high speed switching environment.

Perhaps the design could be re-worked to use a modern device like Broadcom HCPL-4562? (still limited to 17MHz bandwidth) Turns out that device has worse THD due to lack of feedback photodiode.  Neverthless, Broadcom has an improved IL300 drop-in replacement, HCNR200 with 1.5MHz bandwidth, which is still slow but might be a tad more useful.

[bdunham7]

edit: my heart almost stopped when I first saw the price of a galvanically separated probe up to 1GHz 😂

True, but the specs on those Iso-Vu units are jaw-dropping as well.  160dB CMRR and 60kV common mode voltage!  I wonder if that 60kV is good all the way to 1GHz?

[bdunham7]

Perhaps the design could be re-worked to use a modern device like Broadcom HCPL-4562? (still limited to 17MHz bandwidth) Turns out that device has worse THD due to lack of feedback photodiode.  Neverthless, Broadcom has an improved IL300 drop-in replacement, HCNR200 with 1.5MHz bandwidth, which is still slow but might be a tad more useful.

Siglent had an isolation module, the two-channel ISFE, that used regular probes and had a 1MHz BW.  Perhaps they used this part or something like it.  Unfortunately they discontinued it for unknown reasons.

[magic]

I solved it with a 2x 100Mhz battery oscilloscope, which is enough for my purposes

This sounds like an ordinary oscilloscope with GND connected to all probes?

It still has most of the problems:
- common mode of both channels must be the same
- loads the "negative" input with capacitance to ambient (even more so when you are touching the device)
- can be dangerous if common mode is high voltage with respect to earth

It eliminates one disadvantage of "ungrounding" a mains scope:
- no mains noise injection through capacitance of the PSU

[Edison]

It eliminates one disadvantage of "ungrounding" a mains scope:
- no mains noise injection through capacitance of the PSU

Yes, the inputs have a common GND, because I only use one probe, so it doesn't matter, but as I wrote - it's enough for my needs.

Connecting the oscilloscope through the isolation transformer just solves the removal of the connection to the ground, but badly - strangely floating oscilloscope, capacitive coupling of the transformer, etc.

[magic]

I have only ever lifted ground to test power supply primaries, where capacitive current from the scope's PSU is irrelevant since it's absorbed by the low impedance (and high voltage) negative rail.

Safety nuts hate me

[Zero999]

you can power a Rigol DS1xxxZ scope off approximately 50V DC placed across the live and neutral terminals.  That is enough for the internal flyback stage on the PSU to function.  It'll pull around 12W in this configuration.

Good to know 
On the label on the back, the minimum frequency range is 44Hz to 400Hz max, so I didn't expected that this oscilloscope could be powered by DC voltage. 

Of course it'll work off DC. The input to the power supply is just a bridge rectifier. The minimum frequency will be specified because of the size of the smoothing capacitor.

It will work at lower DC voltages, because the smoothing capacitor doesn't have to do any work. I found mine works down to 36V. I considered making a battery pack for it, but lost interest when I bought an oscilloscope with a battery. Others decided to make their own battery pack with tool batteries. Refer to the following thread.
https://www.eevblog.com/forum/testgear/battery-pack-for-rigol-ds1054z/msg717702/#msg717702

[eneuro]

Professionalism means attention to detail. 

If you do not understand from what has been written here by everyone that the isolation transformer should not be connected to the power supply of a classic oscilloscope, take your oscilloscope, open the window and throw it out of the window and the transformers behind it - that will be the most professional thing you can do

PCBs designed to determine transformer polarization have just arrived, so after soldering optoisolators, a few resistors and diodes, it will be possible to connect all transformers to isolate the oscilloscope and the tested 230VAC system as shown in earlier simulations.
Throw your computer out the window because your typing on the keyboard makes no sense Pepiczku 

[S. Petrukhin]

I think it is better to isolate the circuit, rather than the oscilloscope.