LVTTL Logic Ground.

Our weekly LIVE video chat. Every Wednesday at 8pm ET!

Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.
Locked
User avatar
abalducci
 
Posts: 14
Joined: Wed Aug 20, 2014 10:17 am

LVTTL Logic Ground.

Post by abalducci »

Dear Adafruit Engineers,

I am currently working on a project that involves LVTTL signalling to an external device. Of course, reasonably there are the signalling lines out, and then of course a 'return to ground'. This thus may otherwise then seem like perhaps a rather 'naive' question then, but after watching some of Paul Rako's videos at Atmel Edge, I do at least get a sense that 'ground' is not always as simple as one might make it seem, therefore my question.

So, lets say I had 10 paired LVTTL (3.3 v) signalling lines, or 20 total (10 logic, 10 ground), at the end of the interconnects I can simply shuttle all these to the same 'power ground?'. Further, let's say for a moment I had multiple power sources on this board (i.e. LVTTL, TTL, and then a 24 V for controlling a motor). Even though of course the supply lines are all separate, can they then still share the same 'common ground' ? Or do the grounds have to be separate (I'm not sure what that would look like-- assume I have a single [higher] DC voltage input source and then am scaling down via regulators).

I know this is a somewhat 'basic' seeming question, but I want to make sure I understand so as not to let the smoke out of the bottle.

Also, just a 'feedback comment' and less Ask An Engineer-- But for one I would like to first say 'I love Adafruit'-- Though, to be honest (living in MA), I've been a bit less than crazy that certain areas like mine, especially, have been moved away from USPS, to only UPS say-- Why ? Well for one thing, NYC to MA is already basically 'overnight' shipping, even through USPS. Second, the cost, or on this project or that, sometimes I find myself needing just a few parts at a time-- but to pay maybe $5 in parts, and then pay $12 in shipping... It doesn't quite make as much sense. Perhaps I am the only customer that feels this way, but I find myself waiting longer to place orders in this case, at least until I have a complete BOM to justify it. Just wanted to friendly relate, as way of feedback and unrelated to the question.

Best,
- A

User avatar
Franklin97355
 
Posts: 23912
Joined: Mon Apr 21, 2008 2:33 pm

Re: LVTTL Logic Ground.

Post by Franklin97355 »

I find myself needing just a few parts at a time-- but to pay maybe $5 in parts, and then pay $12 in shipping..
Adafruit now sells through DigiKey and they provide USPS shipping, perhaps that would be an option that would work for you.

User avatar
adafruit_support_mike
 
Posts: 67454
Joined: Thu Feb 11, 2010 2:51 pm

Re: LVTTL Logic Ground.

Post by adafruit_support_mike »

abalducci wrote:I do at least get a sense that 'ground' is not always as simple as one might make it seem
That's extremely true.

To scale the effects up so they're easier to understand, imagine you have to build all your connections between components with 100 ohm resistors. Suddenly these two layouts (which look identical if you assume all wires have zero resistance) become different circuits:
gnd.jpg
gnd.jpg (18.57 KiB) Viewed 626 times
In both cases, the bar which indicates GND represents the physical negative terminal of your power supply.
abalducci wrote:Even though of course the supply lines are all separate, can they then still share the same 'common ground' ?
They can -- and should -- all meet up eventually. Voltage always has to be measured between two points, and there always have to be two conductive paths between the points you measure. The measurement is a comparison of how hard it is to push current along the two paths. GND is the common reference point for all voltage measurements in a circuit, so if you don't have explicit/known current paths between two points, the current will follow whatever path it can.
measure.jpg
measure.jpg (18.92 KiB) Viewed 626 times
The upper drawing shows a common trap: If you use a meter to measure the voltage between GND and P1/P2, you'll get valid readings because the meter itself provides the second path between either point and GND. The measurement you get for P2 will be a mess though, because the current flows through a bunch of unknown resistances between P2 and GND.

Thing is, if you take the meter away, it's impossible to know P1's voltage at P2 or GND because there's only one conductive path between them. That kind of mistake can drive you nuts because everything behaves rationally when you measure it, but doesn't behave at all on its own.

The lower drawing shows a well-behaved circuit. You'll get decent measurements between GND and P1/P2, and P2 can know the voltage at P1.

The problem of connecting GNDs comes when you think about the current that flows in the circuit.

Digital circuits use practically no current when holding state, but use a lot of current to change state quickly. When that current flows through the resistance of the GND connections, you get voltage spikes along that path.

When current from two different parts of a circuit flows through the same resistance, the spikes generated by one part become visible in the other part. That's called 'coupled noise', and it can cause big problems for signals and thresholds in parts of a circuit that are supposed to be isolated from each other.

To keep that from happening, you use debounce capacitors to create local voltage rails that are more stable:
isolate.jpg
isolate.jpg (11.35 KiB) Viewed 626 times
abalducci wrote:I know this is a somewhat 'basic' seeming question
No, it's a mid-level issue. It gets swept under the rug to keep descriptions and schematics simple, and is only discussed once you get to a level where the effects of coupled noise become important enough that you have to deal with them.
abalducci wrote:Just wanted to friendly relate, as way of feedback and unrelated to the question.
You're not the only person to make that comment, and we understand your point of view. To understand our point of view, you have to think in terms of shipping around 20,000 packages a month.

It isn't a question of UPS being cheaper for us, it's a question of how well they do the job:
http://blog.adafruit.com/2014/03/14/usp ... rbusiness/

The on-time delivery and tracking performance we see from UPS is in the 99.8% region. The USPS can't match that, and the performance we track has been getting progressively worse.

We recently worked a deal to have our stuff distributed through Digikey, so if you want USPS shipping you can order from them. They'll even order things they don't hold in stock directly from us and re-ship it to you.

User avatar
abalducci
 
Posts: 14
Joined: Wed Aug 20, 2014 10:17 am

Re: LVTTL Logic Ground.

Post by abalducci »

Dear Mike,

Thank you greatly for your quite detailed 'reply' post. I feel you have elucidated the matter rather well and I am still examining it. For the sake of anyone else who runs into a similar question/issue again, I will at least 'restate' insofar as why I had the inquiry in the first place:

In typical introductions, the student is presented with the common 'water' (voltage = pressure, current = flow rate, resistance/load = constraint) analogy... At yet here we seem to have this rather 'mystical', and as you say, taken for granted, outlet called ground-- which at first one might logically think of as the 'drain' on the circuit tub-- Only, as I feel you have reaffirmed/elucidated, where as a typical 'drain' can only absorb 'so much' water at once, 'ground' seems to have this almost 'strange' and 'mystical' property, like a black hole, of being able to 'change size' so to speak, at least in the way to absorb whatever 'pressure' or 'flow' of electrons are being thrown at it.

For other's do correct me if I am wrong here in this observation, but it does seem to be a rather 'peculiar property', as far as the usual analogy at least goes.

And yes, it is okay about shipping, if in the meantime I can hope at least to give back to the general conversation/discussion. Also, if my reply here is not 'totally off', it could be a potentially interesting topic for Ada's future children's related video series-- both 'why is it needed', and 'why is it strange'-- voltage, current, and resistance could also be explained at the same time.

Best,
- A

Locked
Please be positive and constructive with your questions and comments.

Return to “Ask an Engineer! VIDEO CHAT (closed)”