CMOSprinkles' Instant D Flip Flop!

[CMOSprinkles]

Hello again, some of you may remember the 1 tick DFF I was working on using the Dropper SR Latch. As you may recall, it did not work properly, "for some unknown reason".



Turns out, the problem was indirect power. A very helpful member of Minecraftforum pointed out to me that Droppers can become indirectly powered the same way that Pistons can. As you can see in the picture above, the bottom Dropper is one block below the block that the repeater is powering. This causes that Dropper to stay powered even when the repeater has stopped powering the block; thus the first redstone pulse did not move the item, but only updated the Dropper to remind it that it was off. The second pulse could then move the item. This new design not only completely averts that issue, but it also makes it 1 tick faster; it can now instantly save data and output in 1 tick!

Here is an overview of 4 bits stacked 2 wide:


Here is a view of the outputs, 2 wide:


Here is a side view:


Now, here is the Shared Clock version, much neater looking:


The clock input goes in one neat little line, no BUD's here:



This works because instead of indirectly powering the bottom Dropper, it indirectly powers the top Dropper, which is then updated instantly by the powered rail. Each bit is 7L*[(x*2)+1]W*6H (x is the number of bits) stackable in any direction if it has its own clock, and 7L*[(x*2)+1]W*5H if they have a shared clock. It can save data instantly, and a RAM will be coming soon that will read in 1 tick. There you have it; an instant DFF! That means an instant write/1 tick read RAM! Here is a world download and a MCEdit schematic, let me know what you think!

Individual Clock Version:
World Download: http://www.mediafire.com/?8boxucnfly6v6uf

MCEdit Schematic: http://www.mediafire.com/?6itkrp7jm4ab4nb

Shared Clock Version:
World Download: http://www.mediafire.com/?vovghxui9u0pf3g

MCEdit Schematic: http://www.mediafire.com/?bexn9r5kq94cqp8

[Thor23]

I'm not understanding how those pistons are indirectly powered. They appear to be extended with no source of power around. Reminds me of a glitch I myself recently encountered, where some pistons that are powered fail to retract when being unpowered. Now that I see this, I think it may have to do with them being powered from directly above. BUD glitch strikes again?

[CMOSprinkles]

Lol, indirect powering causes a piston/dropper to become a BUD. So yes, it is a BUD glitch.

[CMOSprinkles]

Hm, considering that this is 3 ticks faster than most DFF's out there, I thought there might be a little more interest. Has this already been done?

[Chibill]

I may use this for registers in MCX's CPU.

[Thor23]

Lol, indirect powering causes a piston/dropper to become a BUD. So yes, it is a BUD glitch.

No, I mean I don't SEE the BUD. As far as I can tell from the pictures you've provided, it should be working fine. So I have no idea why it would be BUDing.

[CMOSprinkles]

Do you mean that you cannot see what is causing those pistons to BUD? It is a glitch that I have encountered before, it is based on signal strength and placing the wire above the piston I believe, notice that the first piston isn't affected by the glitch. It is indeed strange.

[Billman555555]

Hmm.. What are the Dimesions,
* Billman555555 thinks he might use this in SAC

---

This has been done before by properenglish but with a different save techneique.
Great Job CMO.

[CMOSprinkles]

7x7x2 stackable per bit, plus 2 more to the width for the write control line. I'd love to see Proper's version of this, if it exists then there should be a thread or a video. Could you link that here? Also, thank you Chibill, you are more than welcome to use this in any of your builds if you just mention my name to give credit.

---

Gah! Y U DO DIS BILL. I just took a look, and Proper does indeed have an SR latch that is even more truly instant than the dropper SR latch... Both designs save the data instantly, but there is one difference; Proper's also outputs instantly, with no comparator delay. True, this version of an 'instant DFF' is most likely more compact than any you could make using Proper's SR latch and is probably functionally identical in a computer circuit since the data does save instantly. However, it is still true that a DFF using Proper's SR latch would output data faster. Thank you for pointing that out Bill. Though, I will most likely still being using this since it is seemingly more compact and functionally identical.

---

Also, I just saw his RAM video too, but that was much less optimized than this is. His RAM was 2 ticks to write and 2 ticks to read. I believe that he wanted to avoid using pistons, so he didn't actually manage to create an instant DFF. Although what he had was a pretty compact version of this, it was slower.

[Billman555555]

IK THT I AM HPLES T HLPN PPL. But yea maybey if you can change the control line you can get it smaller IDk.

---

Anywho i will do a youtube video comparison as soon as i have time.

[Chibill]

Can you update the schematic and world download to the fixed on?

[CMOSprinkles]

Of course, I'll have to do it when I get home today though, so in about 7 hours. Also, another very helpful member of Minecraftforum gave me an idea on how to compact the circuit to be only 6 tall and much more easily accessible. I'll upload the updated version tonight, look forward to it!

[CMOSprinkles]

Sorry, my computer is only letting me work for 5 minutes at a time, it might not be up tonight, but I can safely tell you that the dimensions are looking much better. Now the clock input doesn't have to be shared, each flip flop can have its own clock in two wide space. The new dimensions are 7*[(x*2)+1]*6, where x is the number of bits you have stacked. It still saves data instantly. Also, I'm going to use Proper's instant RS latch to make another instant DFF that also outputs instantly, so hooray! I'm not sure what the dimensions of that one will be, most likely bigger than this, but I won't be able to find out until my computer lets me build it.

[Chibill]

Insta is Good I can deal with size. That's the words of himehowareu the lead designer and builder of MCX I am the second in command of HCI where we visualize minecraft! XD.

[CMOSprinkles]

Alright, here we go! I have updated the main post with the new designs! And great news; the shared clock version (the one that would be used for RAM) is now 5 tall stackable! It is all completely instant, I quite enjoyed watching how fast it changed with the shared clock version. Cheers!

[CMOSprinkles]

Bump.

[Chibill]

You have been added to our thanks list that's in a book on the server for parts for MCX!

[CMOSprinkles]

Well thank you! Also, I'm working on a RAM right now that will read/write instantly on all edges, I'll have that up as soon as my computer wants to turn on again!

[Iceglade]

That would certainly be interesting to see. Can't wait

[CMOSprinkles]

Thanks! It utilizes a two wide stackable instant inverter to create an instant AND gate as the writing system. The inverter does need some staggering, but I have already put together a prototype in a circuit builder, so I really just need to get on actual minecraft to build it. I will note that there will be one delay in the system; the comparator that pulls the signal from the dropper will always make it take one tick to change the output, so that is unfortunate. However, data will be saved instantly (on the same tick) when the clock goes high and that data will be able to be accessed and blocked instantly.