Jump to content

Talk:Ring counter

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Misc comments

[edit]

I'd love to know why such counters are/were used.--64.126.73.194 (talk) 19:15, 15 January 2013 (UTC)[reply]

I added a bit of an applications section, but I'm only speaking from my area of expertise. There may be other applications of which I'm unaware. 128.97.68.15 (talk) 02:37, 6 April 2013 (UTC)[reply]
That application section must have been removed or absorbed at some point. I stubbed in another. Dicklyon (talk) 03:22, 10 June 2018 (UTC)[reply]

The image purporting to show a Johnson counter is incorrect. It should show the /Q of the last shift register feeding back to the D of the first shift register. — Preceding unsigned comment added by 209.81.2.75 (talk) 17:41, 3 September 2015 (UTC)[reply]

The diagram is correct; I clarified the comment pointing out the inversion indicated by the bubble. Dicklyon (talk) 03:22, 10 June 2018 (UTC)[reply]

oeis.org/A000031 is the number of output sequences from a simple n-stage cycling shift register, should this be added to the article? Praoubl (talk) 20:51, 14 March 2019 (UTC)[reply]

It is usual to make FF's with Q and Qbar output, one above the other. Looking at the diagram, one looks at the height more than the little circle. Gah4 (talk) 08:28, 23 May 2023 (UTC)[reply]

Hamming distance

[edit]

The Hamming distance for a Johnson counter is 1. The hamming distance for an Overbeck counter is 2. For example, Johnson(4-bit): Decimal: 0, Binary Code: 0000 Decimal: 1, Binary Code: 1000 Single substitution at bit location 0.(0->1) Overbeck(4-bit) Decimal: 0, Binary Code: 1000 Decimal: 1, Binary Code: 0100 Double substitution bit location 0(1->0), bit location 1(0->1) See NASA GSFC management directive 500-PG-8700.2.7 page 10 section 4 "finite state machines" as reference. 128.97.68.15 (talk) 02:10, 6 April 2013 (UTC)[reply]

Counters don't have Hamming distances. See discussion below. I removed that bit and added it back more sensibly, with citations. Dicklyon (talk) 03:23, 10 June 2018 (UTC)[reply]

Hardware implementations where registers are more expensive than combinational logic (e.g. FPGA)

[edit]

Is it in the time-domain (i.e., registers take more clock cycles and FPGAs have more registers than other configurations)? — Preceding unsigned comment added by 79.16.162.227 (talk) 09:51, 18 July 2014 (UTC)[reply]

Question not understood. Dicklyon (talk) 03:23, 10 June 2018 (UTC)[reply]

Overbeck?

[edit]

An anon added Overbeck counters in 2016.[User:RTC added Overbeck counter in 2006]. When I look for sources, they're all newer than this. Is this a recent attribution of a classic ring counter design to a classic dude, or what? Where does this come from? Does it belong? Dicklyon (talk) 21:43, 5 June 2018 (UTC)[reply]

Do you ever read references? Or have you just decided again that they're wrong and that You Know Better?
Here's the very first reference in this article: https://www.google.com/patents/US2427533 It's a 1943 patent, in the name of Overbeck, for a ring counter. It has been in this article for over two years. It may not be the first ring counter or the first ring counter patent, but it is clearly a source for Overbeck's name in relation to them.
I don't know what, "When I look for sources, they're all newer than this. " means, other than a failure of WP:COMPETENCE. Andy Dingley (talk) 23:02, 5 June 2018 (UTC)[reply]
I read the concern as being that if all the sources post-date the Wikipedia article then they might have got the information from us. SpinningSpark 23:20, 5 June 2018 (UTC)[reply]
Which is obviously not the case when there's a 1943 patent, and when that's already so prominent in the article. Andy Dingley (talk) 23:32, 5 June 2018 (UTC)[reply]
Yes, but Overbeck's patent does not establish that ring counters are called Overbeck counters. But in any case, the issue is moot since earlier sources have now been identified. SpinningSpark 23:53, 5 June 2018 (UTC)[reply]
Sorry, I had the date wrong. Need sources before 2006. So not moot. All I find before are "Overbeck ring". Dicklyon (talk) 05:56, 6 June 2018 (UTC)[reply]
(after EC) Here's one from 2010 [1], but that's still fairly recent. One of the cites in the article is to a 1947 Overbeck patent. This is using tetrodes, but not in pairs as conventional bistable latches. Instead it makes use of negative resistance from the secondary emission effect. The solid state equivalent of this kind of latch is the tunnel diode. This is a somewhat left field application of tetrodes. I would be surprised if someone had not produced a counter with vacuum tubes in a more conventional manner prior to this. Overbeck's patent can hardly be taken as an RS establsihing his priority. Of the other cites, ref#3 does not make any claim for Overbeck's priority and besides, names several other researchers. Ref#2 is dead. None of the general references mention Overbeck at all. SpinningSpark 23:11, 5 June 2018 (UTC)[reply]
I think it's pretty cool that Overbeck had an efficient way to make a cyclic pulse generator, using a lot less than a flip-flop per stage. And in discussion at Wikipedia_talk:WikiProject Electronics#Anyone heard of an Overbeck counter? I note that I've found that "Overbeck ring" is mostly what it has been called. When I made such a thing with neon bulbs (analogous to tunnel diodes) back in the 70s, I had no idea it had a name. Anyway, I do see a very few books since 2010 attributing the old standard one-hot ring counter to Overbeck, and this is what's new and surprising to me. Why did his name resurface, and is it worth putting forward as if this is what people call it. Andy, you didn't say if you've actually every heard it called that; I'd be interested to know. Dicklyon (talk) 02:24, 6 June 2018 (UTC)[reply]

OK, I was wrong on the dates. It came into Wikipedia in 2006 in this edit. So the 2010 and 2014 sources are very likely WP:CIRC. If that editor is still volunteering at the Computer History Museum around the corner, I'll drop in and discuss it with him. The have a RAMAC there, and he's an old IBM guy, it looks like, and there's Overbeck ring in the 1959 RAMAC manual, so I bet that's the connection. Dicklyon (talk) 02:35, 6 June 2018 (UTC)[reply]

And maybe it should be called a Mumma ring counter. Dicklyon (talk) 03:21, 6 June 2018 (UTC)[reply]

  • Andy and Dick: you could both afford to cool it off by one notch. I know it's hard to de-personalise, but ... try. WP needs the talents of both of you. Tony (talk) 05:47, 6 June 2018 (UTC)[reply]

Proposal – unless we can find a source using "Overbeck counter" before 2006 when it was added to Wikipedia, we should change to "ocassionally known as an Overbeck ring" or something to that effect, in recognition of the 1959, 1960, and 1966 uses of that term. There's not a lot else we can do about the circularity problem of a handful of modern sources using the term we promulgated here. Dicklyon (talk) 05:54, 6 June 2018 (UTC)[reply]

I made an edit. Please comment or tweak it. I cleaned a lot of cruft out of the lead while at it. The stuff about the Johnson counter was really sketchy, sourced to de.wp and largely about a decimal specialization. If some of it is found reliable, but it not in the lead please. Dicklyon (talk) 06:15, 6 June 2018 (UTC)[reply]

I moved the following conversation from Wikipedia talk:WikiProject Electronics to keep the discussion together. SpinningSpark 08:58, 6 June 2018 (UTC)[reply]

I don't know the origin. But here is a patent from 1954 that mentions an Overbeck ring circuit, seen in Fig 6. --Mark viking (talk) 22:07, 5 June 2018 (UTC)[reply]
They are mentioned at [ http://ds.opdenbrouw.nl/digse1/ring_counters.pdf ]. Look at the "Straight Ring Counter" section. --Guy Macon (talk) 22:28, 5 June 2018 (UTC)[reply]
OK, it appears that "Overbeck ring" is a better search term. I found this 2014 book. And here in 1959. Still, quite rare. Where has this term been buried, and why has it arisen recently? Is such a rare attribution worth treating as if it's what an ordinary ring counter is called, when practically nobody actually calls it that? Or maybe say "occasionally called..."? Dicklyon (talk) 02:02, 6 June 2018 (UTC)[reply]
My sense from the search is that occasionally called would be a due-weight adjective. --Mark viking (talk) 02:30, 6 June 2018 (UTC)[reply]
Even that might be too much. I see mentions in 1959, 1960, and 1966, and then nothing until after User:RTC added it to Wikipedia in 2006. See the linked discussion; sorry we got forked to here. Dicklyon (talk) 03:53, 6 June 2018 (UTC)[reply]
Your "research" skills really are as bad as your subject knowledge. Do you know who Wilcox P. Overbeck was, and why he was interested in counting? He was at the Chicago Met Lab during the Manhattan Project and was the electronic engineer responsible for the neutron counting on CP-1. (I hope at least some of these links are blue.) His counter circuit wasn't intended as a "counter" in the usual sense, so much as a rate scaler or divider for fast pulse trains, from a G-M tube, thus making a Geiger counter or ratemeter.
1950s-1970s textbooks of applied nuclear engineering use the term. I've only seen it applied to ring counters based on valve devices, usually thyratrons and particularly dekatrons (i.e. fast pulses). Once solid state came along, it seems to have fallen from use. Nor have I seen it myself in other fields, except for one 1960s reference in ECM - again, fast pulses in a world of valve equipment.
Oh, and he'd started out pre-war with Vannevar Bush on a computing project, the "Rapid Arithmetic Machine" (or as Dicklyon will want to call it, the "Ickle Kiddy Computer" or something). He worked on high speed valve counter circuits with that too. One of his innovations was the bi-quinary divider, a divide by ten circuit using sequential division by two and division by five - thus using seven thyratrons, not ten - no idea if that was from the computer or the ratemeter period. He also had some involvement with early phase control circuits for power electronics, using a (very) obscure valve called a permatron (a gas-filled triode(?) controlled by an externally applied magnetic field from a solenoid).
I have no idea what the "P" stood for. In Dicklyon logic (see Manchester Small Scale Experimental Machine), that means he didn't exist. Andy Dingley (talk) 12:20, 6 June 2018 (UTC)[reply]
The P was for Pratt. Thanks for the interesting back story on him. Let's do a bio. Also on the "Ickle Kiddy Computer" as you call it, though sources call it Rapid Arithmetic Machine and Rapid Arithmetic machine and rapid arithmetic machine (or perhaps Rapid Machine or Rapid Analytical Machine, if those refer to the same). But back to the question: if an Overbeck ring is a way to make a one-hot counter/prescaler with few vacuum devices, as attested in sources, is there any source in support of what digital designers call ring counters being called Overbeck counters? I'm not seeing it. Dicklyon (talk) 14:53, 6 June 2018 (UTC)[reply]
I have no idea on the priority of ring counter ideas. Uniselectors were counting long before this, I don't know if more binary (or multi-element 1-ary) devices were being used too, before Overbeck. I'd be happy to see the term "Overbeck" applied to only those circuits which involve a single counting valve (per stage), such as a dekatron. Maybe it was applied to the multi-valve circuits too, or even non-valve circuits, but I can't source that. Andy Dingley (talk) 15:06, 6 June 2018 (UTC)[reply]
Oh, and the Rapid Analytical Machines were totally unrelated, in purpose, technology, timescale and country of origin. Please don't confuse those too. Andy Dingley (talk) 15:07, 6 June 2018 (UTC)[reply]
I found at least one book associating Vannevar Bush with the US Navy's Rapid Analytic Machine (RAM) project, but that may be a red herring, so I'll be careful about that, thanks. On priority for one-hot counters, I don't think we need to take that up; they clearly existed before Overbeck. So I take it you're more or less OK with my edits so far. Dicklyon (talk) 15:16, 6 June 2018 (UTC)[reply]
Don't believe everything you read (in at least one recent book) about Rapid Analytical Machines as a US term, and rather anachronistic. U-571 was fiction too. Read Cryptologia. Andy Dingley (talk) 15:35, 6 June 2018 (UTC)[reply]
Agreed! So much of what you read in modern sources is copied from Wikipedia, and every now and then some of that is just ahistorical. Like Sir Malcolm Thornhill making the first Cardboard box. Dicklyon (talk) 16:29, 6 June 2018 (UTC)[reply]
We should do an article on Robert Royce Johnson, too. Dicklyon (talk) 03:47, 7 June 2018 (UTC)[reply]

Andy Dingley, please stop being so snarky. I don't know what your problem is, but it doesn't belong here. --Guy Macon (talk) 00:57, 7 June 2018 (UTC)[reply]

+1 SpinningSpark 15:21, 7 June 2018 (UTC)[reply]
WP:COMPETENCE - and see the thread below for "hamming distance" Andy Dingley (talk) 16:48, 7 June 2018 (UTC)[reply]
Yes, I made a mistake there in editing an old version and taking over 2 hours to notice and fix it. Mea culpa. Still, +1. Dicklyon (talk) 23:18, 7 June 2018 (UTC)[reply]

I got in email touch with User:RTC, who had added "Overbeck counter" in 2006. He gave me permission to quote from his reply:

I am not sure where I got that term, probably from discussions with an IBM CE on the 1401 Restoration (they had a lot of terminology for various circuits I had never encountered before, some may have been internal IBM terminology of the time not used by the rest of the industry).

Both the IBM 1401 and IBM 1620 computers use that type of ring counter, aka “one-hot counter" (as described in the Wikipedia article) to implement their Control Unit state machines. The IBM 1620 Model I Control Unit has a very very complex “branched ring counter” with roughly 200 states (200 flip-flops) in it. The IBM 1620 uses the other type of ring counter, aka “Johnson counter” to reduce its 1Mz crystal oscillator to its 50KHz (20 microsecond) memory cycle clock with 20 states using only 10 flip-flops.

I know none of the IBM 1620 documents I have mention Overbeck (but of course I don’t have “everything”) don’t know about IBM 1401 documents as I don’t have any of them. I am reasonably sure I got the term verbally and most likely from one or more of the IBM CEs on the 1401 Restoration.

— User:RTC

Some of the guys he worked with on the 1401 restoration at the Computer History Museum also worked on the RAMAC restoration, and the RAMAC docs do have "Overbeck ring", so there's still a likely (verbal) path there. Anyway, there's still no support in sources for "Overbeck counter" before it was added to Wikipedia. Dicklyon (talk) 20:34, 8 June 2018 (UTC)[reply]

Agree 100%. Interesting, though. --Guy Macon (talk) 22:10, 8 June 2018 (UTC)[reply]

Hamming distance removal

[edit]

Why don't we want information on Hamming distances in this article? I can't say I really care all that much one way or the other, but it would be nice to have an explanation for its removal. SpinningSpark 16:43, 7 June 2018 (UTC)[reply]

The statement in the lead was imprecise and out of context. If you find sources that comment on Hamming distance relative to ring counter patterns, please do add it to the article. Dicklyon (talk) 16:46, 7 June 2018 (UTC)[reply]
"hamming distance" !
When stuff like that keeps going on, all it does is demonstrate the power of ignorance, dogma and persistence. 8-( Andy Dingley (talk) 16:47, 7 June 2018 (UTC)[reply]
Well, sources certainly do discuss Hamming distance, for instance [2] and there are numerous scholarly papers, but as the person who deleted the information, you might want to take responsibility for putting it back rather than ask others to do it. SpinningSpark 17:12, 7 June 2018 (UTC)[reply]
  • I didn't delete it, Dicklyon did.
It's yet another of his dogmatic "all capitalisation must go" crusades. Wiki policy against reality. It's Hamming distance, not hamming distance. It's eponymous.
Yes, Hamming distance ought to be in here. We might even look at things like cyclical binary and de Bruijn sequences. Andy Dingley (talk) 17:22, 7 June 2018 (UTC)[reply]
That edit you linked was me accidentally editing an old version where "hamming" was lowercase; when I saw the diff, I immediately self-reverted and said what I did wrong. I apologize for my imperfect editing. Dicklyon (talk) 22:20, 7 June 2018 (UTC)[reply]
S, the article said The Hamming distance of an Overbeck counter is 2 and the Hamming distance of a Johnson counter is 1. But this was nonsense, since counters don't have Hamming distances. There are two Hamming distances of potential interest: that between adjacent states (keeping it down to 1 can be a good property, like in a Gray code); and the minimum distance (keeping it big can be good for error detection, like in a Hamming code). None of this was explained, contrasted, or sourced, and yes I agree it should be. Sounds like Andy has an opinion, too, though I can't discern what it is. It's not really clear that this article is a good place to go into such things. Other than that one point on the Johnson encoding, which is still mentioned in the article (without mentioning Hamming), I don't find any source that mentions "ring counter" and "Hamming distance" close together (if you see it among those articles you linked, please point out what it says). Dicklyon (talk) 22:00, 7 June 2018 (UTC)[reply]
OK, I added a bit of sourced stuff on Hamming distances and encodings that is almost about ring counters, if not quite. Dicklyon (talk) 00:17, 8 June 2018 (UTC)[reply]

Reworked

[edit]

I reworked the article a bit, with some additions, sources, reorganization. Feedback is hereby solicited. Or just go ahead and change it to make it better. I cited one of my own works; feel free to remove or replace if that's not good enough. Dicklyon (talk) 03:43, 10 June 2018 (UTC)[reply]

I still find nothing on the history of the Johnson counter. His patent that we cite has elaborated shift-register counters of various sorts, but not the basic Johnson counter, as far as I can discern. Anybody know more? Dicklyon (talk) 16:35, 10 June 2018 (UTC)[reply]

CD4017 and Hamming distance

[edit]

Might mention the CD4017. It uses a Johnson counter, and a decoder with 2-input AND gates for each output. That means no glitch on any count transition, since only one bit transitions, or, in the usual explanation, Hamming distance of one between successive counts. Gah4 (talk) 08:37, 23 May 2023 (UTC)[reply]