Talk:Non-return-to-zero

Proposed merge
I agree to the mege of the Non-return-to-zero, inverted data here on the condition of resolution of the disputes about the data there.--Mancini 17:46, 27 November 2006 (UTC)
 * Agree. It is useful to have such closely related topics on the smae page. The NRZ page isn't too long anyway and there is already some of the NRZI info on it.  -- Austin Murphy 14:23, 23 January 2007 (UTC)
 * Agree There is no point in keeping it alone. A redirect will do.  mfx  Q&A 11:27, 7 February 2007 (UTC)

I do not agree, they are two different codes, they should be referenced to each other under related topics, but not merged.

I also agree. -disgrutled EE student, 8 April 2007

Is this correct?

 * Taken from the article: "NRZ is used in the RS-232 serial protocol. Internal computer signals often use this code."

I don't believe this is correct. I was of the understanding that RS-232 used +12V to represent 0 and -12V to represent 1, and that was it. ...but at the same time, I don't want to post a correction because I could be mistaken.
 * According to my knowledge rs232 is not used for data encryption but only for the communication purposes beteen two devices it acts only as an controlled amplifier nothing else....
 * i know for sure NRZ is used in USB but in RS232 ????
 * NRZ (largely NRZ-L) is indeed used extensively in serial communications by network devices--Mancini 18:50, 22 November 2006 (UTC)
 * This appears to have been corrected, the "RS232 standard" is not a "protocol", it is just a definition of the voltage levels, typically used on old style modems, the control lines (as well as data) on the modems also use RS232 voltage levels. The "EIA RS232 standard" also defines short circuit currents and input impedances as well. Note that the "EIA RS232 standard" also requires a common wire, and all the signals are referred to that wire, this wire is usually physically connected to the chassis of the computer. Salbayeng (talk) 21:43, 1 January 2020 (UTC)

Response to queries
Just to respond to some of the points raise above:

--Gadget1700 17:09, 18 June 2006 (UTC)
 * RS-232/EIA-232 does indeed use ±12V to represent data with no null level, and so would quality as an NRZ scheme.
 * The page currently states that the "transmitter and receiver lack the timing synchronization". This is only true of NRZ if it is used with an asynchronous communication, and so is really a feature of the timing scheme, not the data representation scheme.
 * Regarding "DC blocking": I suspect that the author of this line may be referring to the fact that in a typical digital electronic system, the components generating the 0V and +V (representing 0 and 1) are CMOS transistors, which have a very small leakage current when not changing state, and so don't allow current to flow when the data isn't changing. If that's what the author meant, then this comment is superfluous. Such circuits still generate heat when they switch, and data loss is, in any case, highly unlikely.
 * Re "DC Blocking", this wording no longer appears in the article, and what is there (in Jan 2019) is sufficiently correct. The comment by gadget1700 above is not the explanation or reason for DC blocking.
 * "DC blocking" means that the signal can still be transmitted by inserting a capacitor in series with the data line, by definition the capacitor "blocks" the DC). This is important in communication links because the signal can then be transmitted through a transformer, which improves the reliability of the network connection, "in the old days" when lightning hit one building it would fry all the serial port connections of computers / terminals in adjacent buildings. Now with 100baseT networks there is an isolating transformer at both ends with 2kV isolation.
 * Old style telegraph, telephone, teletype data connections all required a DC current (or voltage) to be supplied to indicate that the device was available on the other end of the line. (The DC voltage also supplied power from the exchange to the telegrapth or phone). The "break" button on desktop PC keyboards would "break" the DC connection on the serial line, (normally used to get attention at the other end, same as in those old B&W movies where someone jiggled the telephone cradle to talk to the receptionist).
 * A related concept is "galvanic isolation" this essentially means that two pieces of equipment don't have any wires between them that can carry DC current. With larger pieces of equipment, there is an existing galvanic connection, due to earth bonding of the metal frames to the electrical supply, adding an additional galvanic connection (though a data connection) can create earth loops, and cause lightning induced transients to go the wrong way.Salbayeng (talk) 22:13, 1 January 2020 (UTC)

NRZI
I'm not sure is image Image:Nrz-i.gif correct; in the article Non-return-to-zero, inverted there's another image and it's different. First image shows that zero is represented by change of level, while in the article we can read "One" is represented by change in level, "Zero" is represented by no change in level.. Best, 212.33.90.36

I changed the NRZI definition, it was wrong. The image is correct. Mojodaddy 19:29, 18 December 2006 (UTC)


 * According to AMPR there's a differenc between NRZI and NRZ-I
 * BTW: I'm trying to find the correct terms for the german article about this - so tell there also if you find a exact source -- de:Benutzer:mik81 09:35, 1 March 2007 (UTC) -- 82.149.82.142 09:35, 1 March 2007 (UTC)


 * I thing I have an acceptable source. Look here: siemens
 * But convince yourself and search for "Nrz-i" nrzi inverted and get lots of hits
 * -- de:Benutzer:mik81 --82.149.82.142 09:52, 1 March 2007 (UTC)

NRZI definition
In NRZ-I encoding a '1' bit is represented by 0 volts or +V volts depending on the previous level. The encoding of '1' depends  on the current state of the line If the previous voltage was 0 volts then the '1' bit will be represented by +V volts, however if the previous voltage was +V volts then the '1' bit will be represented by 0 volts. A '0' bit is represented by whatever voltage level was used previously.

This means that only a '1' bit can 'invert' the voltage, a '0' bit has no effect on the voltage, it remains the same as the previous bit whatever that voltage was. Rait 13:27, 2 March 2007 (UTC)

Alternative definitions
There do seem to be a number of different definitions of these data encoding schemes depending upon the manufacturer of the communications device. Back in the 1986 the Intel 82530 Technical User Manual had NRZ with a 1 as a high level and a 0 as a Low level, NRZI with a 1 as no change in level and 0 as a change in level. More recently the Motorola 68360 User Manual defines NRZ with a 1 as a high level and a 0 as a Low level, "NRZI Mark" with a 1 as no change in level and 0 as a change in level and "NRZI Space" with a 1 as a change in level and a 0 as no change in level. These do not entirely match what is on the main page. It is also worth noting that the hardware between the comms chip and the physical comms line may invert the signals - thus apparently inverting the output. Thus an encoded "high" will be represented by - 12volts and an encoded "low" by + 12volts on RS232 for example Commsguy (talk) 13:12, 16 September 2008 (UTC)


 * In the article there is no clear definition of NRZ-Mark; worse, it does not give a clear answer to the difference between NRZ-Mark and NRZI. From the images the two seems to be the same. There's something mentioned about the clock, but it's not clear for me (even though I did learn this stuff some time ago). Hoemaco (talk) 15:04, 12 October 2011 (UTC)


 * I get the feeling that there is no difference 116.90.140.41 (talk) 22:54, 20 September 2016 (UTC)

unipolar-bipolar
unipolar says:

"One is represented by one physical level (such as a DC bias on the transmission line).

Zero is represented by another level (usually a negative voltage)."

bipolar says:

"One is represented by one physical level (usually a positive voltage).

Zero is represented by another level (usually a negative voltage)."

which are really the same. Also from the image and the mention of On-Off Keying I'd guess that the unipolar should be when one level is zero volts.

Also, it says

"Disadvantages of an on-off keying are the waste of power due to the transmitted DC level"

which I'd contest. The DC level is a problem due to the spectrum and possible high-pass media, but not because of the power. Actually OOK is the best for power saving, as there's no power transmitted for zeros, ie. roughly half the time.

Hoemaco (talk) 08:23, 10 March 2012 (UTC)

Nrz-s.gif misleading
The Nrz-s.gif has a bit pattern labeled on top as "101100011010". The misleading aspect is the first bit which should be unknown or a 'X'. The gif does not show the previous low state. Alternatively, the gif could be edited to omitted that first '1', leaving just a fraction of the first bit's end of low state. — Preceding unsigned comment added by 70.99.114.194 (talk) 22:56, 9 May 2013 (UTC)

WHAT is, or is not, returning to zero?
I think this article would be less confusing if - somewhere - it was explained what property or value is - or isn't - "returning to zero."

We're told that "NRZ can refer to any of the following serialiser line codes.." but even the linked reference doesn't say whether the "line codes" are a string of bits or a sequence of physical values! I suppose we can rule out the former, as bits can be zero and (I guess) have no place in discussion about NRZ schemes, so presumably "NRZ" refers to the physically encoded sequence that corresponds to a sequence of bits. On the other hand, it's claimed that NRZ(L) is a sequence of "raw binary bits without any encoding" so (again) it's not clear what isn't returning to zero!

Without knowing anything about this topic, I'll guess that a binary sequence like {1, 1, 0, 0, 1, 0, 0} would be mapped to a voltage function of (say) V(t)

= +12v, for t in [0, 1) seconds          = +12v,  for t in [1, 2) seconds = -12v, for t in [2, 3) seconds          = -12v,  for t in [3, 4) seconds = +12v, for t in [4, 5) seconds          = -12v,  for t in [5, 6) seconds = -12v  for t in [6, 7) seconds assuming a transmission rate of 1 bit per second.

Is this right? The voltage doesn't "return to zero" so does this qualify as an NRZ voltage stream? (Of course, I haven't specified how the start and end of the bit sequence is encoded, but you could (say) set V(t) = 1 for t 7.)

I know a fair bit about computer science and electronics, but this article is such a mess that I can't even work out what it's try to say, let alone the details!

StandardPerson (talk) 02:49, 30 March 2017 (UTC)

Mention variants in first paragraph
I think the first paragraph should have a simple sentence that mentions that there are many variants. Somerandomuser (talk) 15:43, 29 September 2018 (UTC)

Randomized non-return-to-zero
didn't actually contain any information about Randomized non-return-to-zero. Another editor who tagged the section recently and I are not able to readuly find any sources on this topic. I am inclined to delete the section outright but we have a lot of redirects pointing to this section. No pages on Wikipedia are actually using these redirects so perhaps they should all be deleted with the section. ~Kvng (talk) 00:05, 4 September 2021 (UTC)