Author Topic: Time Splitter  (Read 12719 times)

0 Members and 1 Guest are viewing this topic.

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Time Splitter
« on: December 31, 2016, 03:16:49 am »
Hi Guys :)
I finished this project a year or so ago,
and since it took a year to make, I wanted to be sure I'd actually finished the whole thing before doing the final video demo :D

https://youtu.be/c93-IrQbUck



At least four seperate parts of this project were Hackaday features!
Strictly speaking, the demo is part 2,
but I think this one wraps up how I spent a year of hobby time quite well!
I hope you enjoy the video as much as I did not enjoy producing it ;)
Maybe that's what I was trying to avoid.
Cheers, Brek.
« Last Edit: January 03, 2017, 12:46:38 pm by @rt »
 
The following users thanked this post: SeanB, cdev, cyr, tautech, hexreader, Cliff Matthews, MK14

Offline joeqsmith

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Time Splitter
« Reply #1 on: December 31, 2016, 04:19:57 am »
It's pretty rare I will watch an entire YT tech video end to end.  Very nice work! 
 
The following users thanked this post: @rt

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Re: Time Splitter
« Reply #2 on: December 31, 2016, 06:15:50 am »
Thanks, I try my best :)

 

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Time Splitter
« Reply #3 on: December 31, 2016, 08:33:21 am »
I'm not going to spend 18 minutes of my remaining life to find out if there is something interesting in there.

Do you have a 30s elevator pitch?
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Re: Time Splitter
« Reply #4 on: December 31, 2016, 08:34:52 am »
You are in a forum sub-section about blogs.

The first 30 seconds of the video I suppose!


I'm not going to spend 18 minutes of my remaining life to find out if there is something interesting in there.

Do you have a 30s elevator pitch?
« Last Edit: December 31, 2016, 08:37:49 am by @rt »
 

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Time Splitter
« Reply #5 on: December 31, 2016, 08:58:29 am »
You are in a forum sub-section about blogs.

The first 30 seconds of the video I suppose!


I'm not going to spend 18 minutes of my remaining life to find out if there is something interesting in there.

Do you have a 30s elevator pitch?

(Reverting to the normal bottom posting. read to difficult top-posting is Because)

1) Shrug; so what.

2) Based on most yootoob videos, I doubt it. In 30s there's a chance I might find if the general topic is interesting.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Re: Time Splitter
« Reply #6 on: December 31, 2016, 09:16:32 am »
I don’t know what to tell you. I’m not selling anything, just sharing my work.
There’s no add support on my YouTube account.
 

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Time Splitter
« Reply #7 on: December 31, 2016, 10:14:12 am »
I don’t know what to tell you. I’m not selling anything, just sharing my work.

Sure. I'm just noting how you might be able to get more people interested in your work. Maybe it will help them, but not if there are impediments to their finding that out.

Quote
There’s no add support on my YouTube account.

I have no idea what you mean by that.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Re: Time Splitter
« Reply #8 on: December 31, 2016, 11:18:21 am »
I’m not sure I have understood you, or even that I really do now.
Are you asking your initial question in a manner on behalf of everyone?

What I meant was I have nothing financial to gain by you watching it or not (unlike add supported YouTube accounts),
although I’d prefer to think there are like minded people out there interested in the same thing.
So there is no need for me to push it, and I fail to understand why someone would try to get me to pitch it to them
any further than the thumbnail suggesting dead bug prototyping, and the description already given underneath that.

This being a sub-forum about blogs other than EEVBlog, if you come in here, it should already be because you are interested
in finding or watching other electronics related YouTube videos, or whatever other type of written or pictorial article.

So with that in mind, if my description and thumbnail does not appeal to any particular person,
I wouldn't expect them to be interested in the video itself, and would anticipate they’d simply move on to the next thread.





 

Offline hexreader

  • Frequent Contributor
  • **
  • Posts: 260
  • Country: england
Re: Time Splitter
« Reply #9 on: December 31, 2016, 11:58:08 am »
I found both the video and the project interesting. Nice mix of art, nostalgia and electronics  :)
 
The following users thanked this post: @rt

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 6145
  • Country: ro
Re: Time Splitter
« Reply #10 on: December 31, 2016, 12:00:35 pm »
I like the steampunk look, very nicely done, congratulation!
 
The following users thanked this post: @rt

Offline joeqsmith

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Time Splitter
« Reply #11 on: December 31, 2016, 06:13:00 pm »
Quote
What I meant was I have nothing financial to gain by you watching it or not (unlike add supported YouTube accounts),
although I’d prefer to think there are like minded people out there interested in the same thing.

I appreciate you taking the time to put the video together just for the pure enjoyment of it.  Your comment about being able to do a fair amount of it with a micro, and the documentation (or lack of it) fit right into it.   I like seeing what other people build even if the end goal was just to have fun.   When you posted your core memory, I read through your entire blog and really enjoyed seeing it tied into the rest of the project.   

I've certainly done the same thing, using wire wrap and very old technology just because.  It's a hobby and I enjoy it.  Again, thanks for taking the time to put the video together.
 
The following users thanked this post: @rt

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: Time Splitter
« Reply #12 on: December 31, 2016, 06:38:44 pm »
What I meant was I have nothing financial to gain by you watching it or not (unlike add supported YouTube accounts),

Ah; I see. You mean "advert", not "plus". It didn't cross my mind that you might have been "selling" something :) (In English the contraction of advert is "ad".)

Anyway, it is your choice whether or not to entice people into believing they want to look at what you have done.
« Last Edit: December 31, 2016, 07:07:40 pm by tggzzz »
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline MK14

  • Super Contributor
  • ***
  • Posts: 4527
  • Country: gb
Re: Time Splitter
« Reply #13 on: January 12, 2017, 04:30:24 pm »
Thanks!

I enjoyed the videos.

I'm very impressed with the build quality of your project. It partially reminds me of HP build quality, and has a look of great solidness to your project(s), on the videos.

I'm amazed that you have (presumably) hand built your very own 64 bit core memory for the project. That is just amazing!

You spend of lot of time to make sure that your various displays, have interesting "animations", as part of them. Such as the rolling digits. That is an interesting and neat feature.

I can understand why it is difficult to classify your project, here.

A core memory, LED/LCD display, super accurate (Rubidium) time piece, with its own GPS unit, for calibration (I think, not 100% sure). Assembled with a giant 1960's Nixie tube calculator kind of build quality. With (presumably) lots of logic chips and other electronics.

I was wondering where you managed to get the tiny cores from (my best guess's would probably be by taking an old/broken/scrap one apart or they are just ferrite bead cores {I don't think those would work, as they DON'T retain the magnetic force. But I'm not 100% sure} you can just still buy for general electronics use). Presumably your other videos, would answer that question ?
 

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Re: Time Splitter
« Reply #14 on: January 13, 2017, 09:15:42 am »
Hi MK14, thanks for your comment :) ... and also thanks to the others that posted!

The core RAM, yes I did make it from scratch, and didn’t have to break down any vintage items. The memory cores I used are still available:
http://www.ebay.com.au/itm/BOX-of-45k-VINTAGE-RAW-FERRITE-CORES-5221-3-2113-35-FOR-CORE-MEMORY-BOARD-/310724678963?hash=item4858a25533:g:NckAAOSw5VFWJiYR

This is indeed true square loop ferrite. It’s probably mentioned in a previous video,
but there’s also some misinformation I should probably go back and correct.
For the video about the core memory flip-flop circuit, I was convinced that the larger cores I used (about 4mm) was square loop as well,
but have since found it isn’t. It’s more “S” shaped, but does have enough remnant magnetisation to work in that circuit.
The flip-flop doesn’t rely on inducing current in a sense wire when flipping the core's direction, but instead,
only has to present greater impedance for a wire turn in the wrong direction.
So strictly speaking, you could still use a lot of these flip flops to make a large memory, it would just take twice as many cores.

I see I’m already ranting, you can maybe tell the core memory is the only reason I still have an attachment to this project :D

There is still a square ferrite available today from a company called Ferroxcube. The material is 3R1, but these are large.
The smallest core, is I believe about 9mm (or maybe larger).
I have made a friend in the US, who has already plotted their curve, and that’s what I’ll be using 8 of next to have another go at
a core logic shift register.

This is his plot he showed me today! This plot is actually still less than 12 hours old as I’m typing this.
 
The following users thanked this post: MK14

Offline MK14

  • Super Contributor
  • ***
  • Posts: 4527
  • Country: gb
Re: Time Splitter
« Reply #15 on: January 13, 2017, 09:52:24 am »
Hi MK14, thanks for your comment :) ... and also thanks to the others that posted!

The core RAM, yes I did make it from scratch, and didn’t have to break down any vintage items. The memory cores I used are still available:
http://www.ebay.com.au/itm/BOX-of-45k-VINTAGE-RAW-FERRITE-CORES-5221-3-2113-35-FOR-CORE-MEMORY-BOARD-/310724678963?hash=item4858a25533:g:NckAAOSw5VFWJiYR

This is indeed true square loop ferrite. It’s probably mentioned in a previous video,
but there’s also some misinformation I should probably go back and correct.
For the video about the core memory flip-flop circuit, I was convinced that the larger cores I used (about 4mm) was square loop as well,
but have since found it isn’t. It’s more “S” shaped, but does have enough remnant magnetisation to work in that circuit.
The flip-flop doesn’t rely on inducing current in a sense wire when flipping the core's direction, but instead,
only has to present greater impedance for a wire turn in the wrong direction.
So strictly speaking, you could still use a lot of these flip flops to make a large memory, it would just take twice as many cores.

I see I’m already ranting, you can maybe tell the core memory is the only reason I still have an attachment to this project :D

There is still a square ferrite available today from a company called Ferroxcube. The material is 3R1, but these are large.
The smallest core, is I believe about 9mm (or maybe larger).
I have made a friend in the US, who has already plotted their curve, and that’s what I’ll be using 8 of next to have another go at
a core logic shift register.

This is his plot he showed me today! This plot is actually still less than 12 hours old as I’m typing this.

That scope plot, looks like a very nice BH curve (and hysteresis circuit symbol), as one would expect. Amazingly clean and well shaped (as expected by theory).

A shift register, using core memory techniques sounds fun!

I might have already seen your core memory techniques, logic (flip flops). But if not I plan to watch it.

I fairly recently watched, a whole pile of core memory youtube videos, which is why I'm not certain if I have watched it yet or not.

I have either skimmed through it (watching bits of it) and plan to watch it later, or already seen it. Hence I can't remember if I watched it all yet!

I'm interested/fascinated by how core memory has to be erased, in order to be read. Then it has to be re-written. I've watched numerous videos about it. But I'm still a bit confused by it. But I understand that the fixed magnetism polarity, CAN'T be sensed by coils, since it is NOT a changing magnetic field. So it is using the changes in saturation, while writing to the core again, to detect if it was at a logic '1' or '0' (magnetically), BEFORE the write (which actually is for a read). Then it needs to put the information back in again.
Modern day Dram memory is a bit similar. Because it needs constant refreshing, otherwise the data will start to fade, some milliseconds after the last refresh. Obviously core memory does not need refreshing until it next needs to be read. But there is a sort of basic similarity of sorts.

It is also fascinating, that they use to make (or intend to make, as I think it was relatively unsuccessful as regards sales), computers of the time (probably 1950's), using core memory elements as logic gates and similar.

The main videos I have watched so far (as regards core memory) are these long (approx 1 hour each), black and white, ancient military videos about it. I find them quaint and fun to watch.

Because (e.g.) in the 1950's, transistors were extremely expensive. Using core memory techniques, to allow somewhat few transistors, to make powerful, computer logic systems. Makes a lot of sense, and tiny cores are fairly cheap, compact, and practical to make into computer logic boards of the time.
« Last Edit: January 13, 2017, 10:04:14 am by MK14 »
 

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Re: Time Splitter
« Reply #16 on: January 13, 2017, 10:19:26 am »
I’d bet I have watched all of them several times over the last couple of years :D

The destructive read of the memory shouldn’t be difficult to understand if you concentrate on one bit (one single core).

If you pass current through the core in the direction that will set the core to binary 1,
and your core was already set to binary 1, nothing will happen, and you know three things.

A: The direction you passed current through the core (because it was your circuit that did it)
B: The core was previously set to binary 1 (because nothing happened),
C: The core is currently set to binary 1 (the circuit passed current in a particular direction).


If you pass current through the core in the direction that will reset the core to binary 0,
and your core was previously set to binary 1, you are changing it’s digital state.
When the coercive current you pass through the core is great enough, it will flip magnetic direction at great speed,
inducing a small current in the sense wire for a short time while
the magnetic direction is switching. For this moment the core is being a transformer.
The current driving line is the primary, and the sense wire is a secondary.
When your sense circuit reads the sense pulse as your current driver sent the current pulse,
you know three things.

A: The direction you passed current through the core (because it was your circuit that did it)
B: The core was previously set to binary 1 (because a sense pulse was received),
C: The core is currently set to binary 0 (the circuit passed current in a particular direction)

In resetting the value of the core from 1 to 0, you destroyed the memory content of the core which was 1.
But you also know the old and new value, so can write it back if you wish.
 

Offline MK14

  • Super Contributor
  • ***
  • Posts: 4527
  • Country: gb
Re: Time Splitter
« Reply #17 on: January 13, 2017, 10:50:49 am »
Thanks for the very detailed explanation. That makes a lot of sense.

So that also explains (hints at), why you (and maybe old 1950's computers), would prefer it to be a single bit at time (or at least a small number of bits). To minimize all the circuitry needed to accomplish that.

Your analogy or way of thinking about it, as a (tiny) transformer, is a very good one, and one I don't think I've come across before.

Thinking of it as a tiny transformer, so that if the logic 1 magnetic force, was already present, the transformer (primary) will either see little or no magnetic change. Since it was already at a logic 1. Hence the transformers secondary (sense) wires, will see little or no signal, briefly produced in it.

But if it was already at a logic 0, the transformer WILL see a significant change in magnetic level, and hence the secondary will show a brief, electrical change.

Yes, that makes a lot of sense. Sorry to repeat what you just said. But I was just showing that your attempts at explaining it to me ARE working.  :)

Now that I better understand how core memory works. Thanks to your explanation in the last post. I can watch the videos again, and hopefully better understand them.

The video(s) which end up 101% totally confusing me (no need to try to explain it to me, I'm so, so confused by it, that I have given up for now). Is a particular military core memory techniques for performing logic functions. Probably flip-flops or even shift register. It is where a tiny cores coil is connected up to the one on the right and then maybe another on the right. Then the signal propagates through.

I'm use to standard logic functions (such as TTL). But unfortunately these logic core (memory) devices/techniques operate in a significantly different way. So it is much harder to understand. I was expecting them to be like normal logic gates and stuff. But instead it uses a different approach. Hence my complete confusion.

Here (I think) is the video which ends up badly confusing me, sorry!
I think I have seen parts of that or similar video(s), embedded into your videos. So I guess you have seen them!
It is where the core directly connects to the core on the right. With data transfer between them, because they are connected together, electrically.

https://www.youtube.com/watch?v=X0WnddW5gZI&feature=youtu.be&t=1056

Time position about 17 minutes 36 seconds: (direct link to that time position directly above this text. For some reason forum software won't let me embed that time position)



I think that the old skills from the 1950's and 1960's are getting somewhat forgotten. So that not many people really know how to do core memory and core memory techniques to make computer logic systems. These days.
« Last Edit: January 13, 2017, 11:03:16 am by MK14 »
 

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Re: Time Splitter
« Reply #18 on: January 13, 2017, 11:15:56 am »
I really hope you keep up with my channel, because I know that this year I’ll be successful with the core logic shift register detailed in those two videos!
That project is what the 3R1 cores mentioned above are for. A shift register made of “Single Diode Transfer Loops” detailed in that video.
The second parts links them together at the end and explains the magnetic shift register, which I have so far failed at once (but I know those cores are unsuitable now).

That first failed attempt is also pretty detailed on my channel. I do understand the application, but didn’t properly implement it.
I decided to leave the failed experiment up, because I just know I’m going to follow through ;)

I want to strongly relate it to those two videos, and am hoping the future project shows up with similar searches.
I’m happy to have found the modern 3R1 cores because the project could be reproduced by others.

Yes, your explanation of the memory sounds right. If you apply current in the direction the core is already set,
The current passed through it does not have to permeate it, and there might as well be no core.
The only additional complication for core memory is the matrix grid used to supply half of the switching current
through a number of cores, so as to be able to select a single core in the matrix.





 

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Re: Time Splitter
« Reply #19 on: January 13, 2017, 11:20:33 am »
For the video you linked at about the time you mentioned, the current passed into the first core,
and between the first and second core, are only pulses. Not continuous current.
You can think of the first core setting the second core as a single action that happens once, and very quickly.

The winding around the first and second cores is a secondary on the left coil, and a primary for the second core in the right coil.
 
The following users thanked this post: MK14

Offline MK14

  • Super Contributor
  • ***
  • Posts: 4527
  • Country: gb
Re: Time Splitter
« Reply #20 on: January 13, 2017, 12:09:25 pm »
For the video you linked at about the time you mentioned, the current passed into the first core,
and between the first and second core, are only pulses. Not continuous current.
You can think of the first core setting the second core as a single action that happens once, and very quickly.

The winding around the first and second cores is a secondary on the left coil, and a primary for the second core in the right coil.

Thanks again!
That makes sense.
Also, in the meantime I've also rewatched that video

I'm convinced that (if I wanted to, or need to in the future), I CAN fully understand those videos, and core memory/logic concepts/techniques, as and if necessary.

I was being thrown (into NOT understanding it) because I was expecting it to be like "Normal" TTL logic like concepts.

But in reality (as you have just kindly explained above), it is NOT stationary logic 0 and logic 1 levels. As you would have in a TTL (and similar) logic device/circuit. Which are (typically) waiting for the next clock transition. Then the logic states will rapidly change again.
But a flow of what are essentially pulses.

Hence that is why I was having difficulty understanding it (the logic side of the core memory used as logic circuits).

Where I am a bit disappointed in that video I LINKED TO is that he suddenly introduces the concept of multiple cores (coils), connected together. WITHOUT (apparently) explaining why there are multiple cores and what the multiple cores are going to be used for.

I did not realize that understanding how core memory logic circuits work (especially as regards multiply connected core memory elements, to perform computing logic functions), would be so difficult to understand. When one is use to the "static" TTL logic way of doing things.

I guess most logic things these days are clocked (synchronous). Which although core memory/logic are also potentially clocked/synchronous as well. They are potentially relying on various pulses, flying through the system. In order to make shift registers, and other computer logic functions.

Or to put it into different words. What I mean is that if a modern day TTL/FPGA system, was to greatly slow down or even stop its clocks. There would be greatly slowed down (or even frozen) logic levels to observe. Potentially making it easier to understand and experiment with.

Whereas these core memory/logic systems, appear to always need these relatively high speed/fast (to us humans), pulse transitions. In order to function.

It is a bit like some very, very old logic (discrete germanium transistor), which had speed up capacitors, across various resistors (a sort of RTL logic, Resistor Transistor Logic). So that there was a sort of high speed part to the signals. Even if it is clocked at very low speed.

tl;dr
I reasonably fully understand how basic core memory works, as regards simple read/write operations. Thanks!
I somewhat/partially now understand how core memory/Logic. Such as (transfer) shift registers work as well.

If I ever end up making stuff like that and/or repairing stuff like that. I will then spend the time to much more fully understand the intricacies of how it is done. But I have enough of a grasp now, to have a partial understanding of how the more complicated aspects of core memory/Logic is doing its stuff.

Anyway thanks again for all the explanations!

As time goes on (after your explanations and rewatching the video). My understanding is getting, better and better. I bet that if I was to actually try and make some core memory and/or core logic types of devices. I would have many more technical issues/problems. Which I just don't realize about at the moment.

Finally, thanks for explaining where you got the ferrite cores from. It is amazing the stuff you can get on ebay.
« Last Edit: January 13, 2017, 12:13:34 pm by MK14 »
 

Offline cdev

  • Super Contributor
  • ***
  • !
  • Posts: 7350
  • Country: 00
Re: Time Splitter
« Reply #21 on: January 13, 2017, 03:10:35 pm »
Really gorgeous from a steampunk design standpoint! Have you ever considered doing some kind of designed space like a bar or restaurant ?


Perhaps as the inside of some device/devices?


Really. Say in San Francisco.


 I KNOW you could make a lot of money doing that!
"What the large print giveth, the small print taketh away."
 

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Re: Time Splitter
« Reply #22 on: January 14, 2017, 01:48:23 am »
I never thought of interior decoration, but closer to the time I did this,
someone suggested looking into making movie props.
I looked into it enough to find out everyone wants to do it! :D
 

Offline cdev

  • Super Contributor
  • ***
  • !
  • Posts: 7350
  • Country: 00
Re: Time Splitter
« Reply #23 on: November 19, 2017, 01:25:36 am »
Just watched these videos again and just loved this project. So inspirational.
"What the large print giveth, the small print taketh away."
 
The following users thanked this post: @rt

Offline @rtTopic starter

  • Super Contributor
  • ***
  • Posts: 1051
Re: Time Splitter
« Reply #24 on: November 21, 2017, 04:57:34 am »
Thanks :) It seems so long ago. Even the video was made a year after the project :D
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf