11

Background

I am a junior engineer who works with testing and designing inverters for motors. I have been looking at a new oscilloscope since the one I had gave up. There are a lot of Tektronix TPS2024b being used since it has dedicated isolated channels on it. I was not sure how isolated channels worked to begin with but I now understand their usefulness. It is very handy to have a scope being able to do the same measurements as a multimeter without making mistakes, but also very nice to be able to avoid long paths to ground for more precise measurements.

So I decided an isolated scope that can measure high and low voltages well would be a a good choice for me, especially since differential probes sometimes can cost quite a bit, and adding that to the cost of the oscilloscope it quickly takes off.
So I started looking and it seems it is quite unusual, very unusual in fact. So I thought it would be a good idea to ask here to see if you can tell me what I am missing. I have noticed that the bandwidth seems lower on the isolated scopes, but it is hard to draw any conclusions with so few alternatives to compare.

Question(s)

  1. Why are there so few bench oscilloscopes with dedicated isolated channels?
  2. Is it a design issue, a cost issue, a business issue or is it just better with differential probes?

Other

Oscilloscopes I have found:

  • Tektronix TPS2000 series (~15 year old design)
  • Cleverscope CS448
  • Handyscope HS6 DIFF

There are also some hand-held isolated oscilloscopes that I could add to the list but I have not written them down since I was not looking for a hand-held scope. USB scopes seem to me very good and like a no-brainer, if it is possible to get one.

Null
  • 7,448
  • 17
  • 36
  • 48
MisterB
  • 113
  • 1
  • 7
  • 2
    Use two standard scope inputs; one for each of the differential signals then use the scope's subtract function to get differential. I've never used a differential probe in my life. – Andy aka Dec 06 '19 at 11:22
  • 2
    High dv/dt over the isolation barrier has an effect and may falsify your measurement. – Jeroen3 Dec 06 '19 at 11:48
  • @Andyaka I do know it can be done that way, but I am trying to understand specifics in why things are as they are before I commit to buying a new scope. The math function on the TPS2000 series works, but that takes up two channels aswell as the math channel for one measurement. I would assume it on top of that turns out to be less precise aswell, especially at higher voltages. My question is aimed at trying to understand why almost none of the big players design and sell isolated scopes, and I am pretty sure I know less about scope design then them, – MisterB Dec 06 '19 at 13:02
  • Here's a useful commentary http://www.techni-tool.com/site/ARTICLE_LIBRARY/Tektronix%20-%20IsolatedInputOscopes.pdf I'm sure it's a cost issue, the TPS2024B is about twice the price of the non-isolated but otherwise comparable TDS2024C. I've been working on motor controls for years, and I have one isolated probe that plugs into my TDS-5034. For monitoring multiple channels, I float the EUT with a transformer, and make the ground reference one of the rails feeding the output bridge. – Phil G Dec 06 '19 at 15:04
  • Yeah I have read that article. It is the same as on Tektronix page; https://www.tek.com/document/fact-sheet/fundamentals-floating-measurements-and-isolated-input-oscilloscopes. Like you say, it probably does have to do with initial unit price and also with not wanting to close of selling of extra probes. – MisterB Dec 06 '19 at 15:33

2 Answers2

15

Why are there so few bench osilloscopes with dedicated isolated channels? Is it a design issue, a cost issue, a business issue or is it just better with differential probes?

It is a cost (price) and demand issue. Most users don't need isolated inputs. Most users have ground-referenced signals so a scope with non-isolated inputs is OK.

Cost is an issue as isolated inputs require more complex circuits compared to the standard solution. This makes the oscilloscope more expensive.

Also note that isolated inputs and differential probes are different things! You can use an isolated input as if it is a differential (floating) input but it's not really the same as the connection on the oscilloscope is a BNC connector which is a single ended input. The two electrical connections on a BNC plug aren't precisely the same, one is also the shield, the other (inner connection) is shielded. For DC and low frequency that usually doesn't matter much but for higher frequencies ore low noise measurements it can matter. Then a true differential probe is the better option.

Bimpelrekkie
  • 80,139
  • 2
  • 93
  • 183
  • Ok, so if I understand what you are saying correctly it means that when I use a "differential probe" the circuitry is, as much as possible, the same for both signal wires, whereas when I use a scope with isloated inputs it is up to the design of the probe, but that generally the probes ground wire does not take that into account. – MisterB Dec 06 '19 at 12:55
  • The scope I have been most interested in is the Cleverscope cs448 and I think the senior engineer mentions what you are talking about 25min into this video https://www.youtube.com/watch?v=I7ppDNLlEL4. I would appretiate if you listened to what he says about it. I interpret it as him saying they have changed the probe design to account for that, is that right? – MisterB Dec 06 '19 at 12:58
  • Sorry I think I edited to many times, the time for the video is rather 23:50, sorry about that. – MisterB Dec 06 '19 at 13:07
  • I have seen (part of) that video before. I am not sure that the guy says what you claim he does. I think he's not doing anything differential. However you can still use two channels and look at the difference between their signals. That boils down to a proper differential measurement without using a differential probe. But that's nothing new, all scopes can do the same. – Bimpelrekkie Dec 06 '19 at 18:44
  • That seems like a strange thing to say. Just a "regular" measurement with a "regular" oscilloscope is differential but with regards to ground. In the case of having a floating/isloated scope the measurement is with regards to where ever one would connect the ground wire to. Or am I wrong? To me it seems like there is some ambiguity to all this. Tektronix for instance has both differential probes as well as isolated probes, both to be used in floating measurments. The engineer showing of the isolated probes (IsoVu) even brag about how they show a "true differential measurement". – MisterB Dec 09 '19 at 08:04
  • 1
    You're trying to couple isolated and differential measurements but these are **different things**. Differential just means that I measure a voltage using two probes and both probes use the same ground reference. Isolated means that the ground of the probe is isolated so not connected to ground. A floating measurement just means that the ground reference does not need to be the common ground, it can be anything. So for a floating measurement you need an isolated probe. That's still not differential. For differential you need **two** probes. – Bimpelrekkie Dec 09 '19 at 13:49
  • Ok, thanks for trying to explain but I dont think you are correct, or alteast not being able to explain it properly. I don't think I'll be able to convice you so I'm just going to post some links: https://www.tiepie.com/en/articles/differential-measurements , https://www.se.com/ww/en/faqs/FA244497/ .A differential measurment is often defined as a floating measurement, i.e. not referenced to ground but rather the difference between two points. Then if that is trough an op-amp(diff probe), via math function(two probes+common gnd) or a floating ground(one probe) should not matter. – MisterB Dec 09 '19 at 15:31
  • I think the information under "Differential input oscilloscopes" in the link below answers my own question that I stated in my second comment above: https://www.picotech.com/library/oscilloscopes/differential-and-floating-voltage-measurements , thanks for listening. – MisterB Dec 09 '19 at 15:38
  • 1
    *not referenced to ground but rather the difference between two points* That is the **commonmode voltage**. Often that voltage isn't available in a circuit or it is high impedance so you can't use it as a ground reference (for an isolated probe). *A differential measurment is often defined as a floating measurement* No, it is **sometimes** a floating measurement. You're still confused that one (differential) requires the other (isolated) which is not the case. – Bimpelrekkie Dec 09 '19 at 17:35
  • Do you know of anywhere where I can read about this in more detail? Because I still don't get it – MisterB Dec 09 '19 at 20:18
  • Thank you for your patience with me @Bimpelrekkie, I will try to read up more on how it works to be able to understand the differences. I talked to Bart Schroder, the owner of Cleverscope, and he might give some thoughts on the matter here related their product. I might turn these questions into a better formulated one and make a new question, seeing as this no longer answeres the original question I asked. – MisterB Dec 10 '19 at 09:38
3

Source: I work for a scope company with scopes

Essentially, the answer to "can it be done" is almost always "yes" unless it's a physical impossibility. However, the number of people that actually need this and make it a purchase requirement isn't really that high. Often there are workarounds that folks find.

There's also some discussion in this thread about differential probes vs. channel isolation. If you want to learn more about differential probes, there are a couple killer videos on our YouTube channel about them:

Single ended vs. Differential probes: https://www.youtube.com/watch?v=a28lVMsWFEY

Differential Probe Guide: https://www.youtube.com/watch?v=OZDijMDHmtI

One way to think about it is that an isolated measurement is a differential measurement, but a differential measurement isn't necessarily an isolate measurement. You can actually use a differential probe as a single-ended probe, which is covered in that first video.

Daniel Bogdanoff - Keysight
  • 730
  • 3
  • 6