Hi there,

I'm currently reaching the end stages of my breadboard 8-bit CPU project. The project uses about 28 breadboards, it's very similar to Ben Eater's breadboard computer ( https://eater.net), he inspired me to actually crack on with taking my designs and building something. The problem I'm having now is that when I plug in the system to a 5V 2A, or even a 5V 5A PSU, the system is dead. A few lights turn on but it's obvious not enough current is reaching half of the ICs. My question is, what is the best practice when connecting power for a system like this, how would I distribute the power properly?

Thanks very much.

 

Use thicker wires on your power and ground distribution lines.
Don't daisy chain your power and ground lines. Use a star connection from the power supply.

 

Use thicker wires on your power and ground distribution lines.
Don't daisy chain your power and ground lines. Use a star connection from the power supply.

Thanks! I'm using the standard copper wire. What gauge would you recommend? Again thanks so much!

 

Use thicker wires on your power and ground distribution lines.
Don't daisy chain your power and ground lines. Use a star connection from the power supply.

Also, what is a star connection in this context? I understand what it is in context of 3 phase power.

 

I think what he means is that you should not run the power and ground from the supply to board A, board A to B, B to C...

Run power from the supply to one rail on each board, then chain only to the rail on the far side of the board. That allows each board to draw the current it needs without having each of the previous boards in the chain impact the current available.

Edited to clarify: Star in this case is with the power supply at the center of the star and one board on each ray. As a final thought, you can also redundantly chain the ground rails between boards. Never hurts if isolation is not needed or desired.

 

I think what he means is that you should not run the power and ground from the supply to board A, board A to B, B to C...

Run power from the supply to one rail on each board, then chain only to the rail on the far side of the board. That allows each board to draw the current it needs without having each of the previous boards in the chain impact the current available.

Edited to clarify: Star in this case is with the power supply at the center of the star and one board on each ray. As a final thought, you can also redundantly chain the ground rails between boards. Never hurts if isolation is not needed or desired.

Oh okay, do you think I could get away with my standard gauge copper wire? I don't really want to have to buy more wire if I don't have to. Thanks very much btw. I was very scared that this would stop me completing the project, you've genuinely helped me so much.

 

do you think I could get away with my standard gauge copper wire?

Well...standard gauge divided by standard amps = standard result.
Got it?
Neither do I.

How are we supposed to guess what, "standard gauge" is? Or how much current each board uses?
No matter what gauge you use, you will need a reservoir capacitor and a bypass capacitor on each board.
That's something like, 100uf in parallel with a nanofarad to filter local voltage on each board. Your mileage may vary.

 

That's something like, 100uf in parallel with a nanofarad to filter local voltage on each board.

I think you mean 10nF to 100nF, no?

 

Well...standard gauge divided by standard amps = standard result.
Got it?
Neither do I.

How are we supposed to guess what, "standard gauge" is? Or how much current each board uses?
No matter what gauge you use, you will need a reservoir capacitor and a bypass capacitor on each board.
That's something like, 100uf in parallel with a nanofarad to filter local voltage on each board. Your mileage may vary.

Oh sorry, I just assumed that there was a standard size of copper wire for using with breadboards. I use the 0.6mm stuff single core stuff. I don't understand, why would you need caps for every board?

 

why would you need caps for every board?

Despite my horrible memory about what size the caps should be, the theory is still in my head.
Digital chips do fast switching. The inductance of a few inches of wire can cause slight dips in the power voltage. Local capacitors supply the quick switching current, thus rendering the impedance of the wires almost imperceptible from the point of view of the chips.

https://forum.allaboutcircuits.com/threads/decoupling-or-bypass-capacitors-why.45583/

 

Despite my horrible memory about what size the caps should be, the theory is still in my head.
Digital chips do fast switching. The inductance of a few inches of wire can cause slight dips in the power voltage. Local capacitors supply the quick switching current, thus rendering the impedance of the wires almost imperceptible from the point of view of the chips.

https://forum.allaboutcircuits.com/threads/decoupling-or-bypass-capacitors-why.45583/

Ohhhhhhhhh, fair enough. I understand. Cool, so a reservoir cap and a bypass cap. 10nF to 100nF. Sounds awesome. Thank you very very much, I learned something new and found a solution!!!

 

The explanation I gave you is rather simplistic. That's why I linked to a rather comprehensive discussion about the theory and application. In your case, several switches flop at every clock cycle. In some machines, thousands of switches flop at once. That multiplies, "slight dips" into results that can completely stop a digital process. I suspect you just found that out the hard way.

A daisy chain power rail and no local capacitors, and half the boards crashed?
This is not a surprise to the well experienced designer.

 

Standard hookup wire is 22AWG or 24AWG.

Star connection from the power supply means, divide up your boards into separate sections. Run two wires from the power suppy to each section, one wire for +5V and another wire for GND. Use heavier wire such as 18AWG.

Use one reservoir capacitor for each section, about 100 to 470μF.
Use a few 100nF sprinkled across each section.

For specifics, we need to see your layout and need to know the current demand for each section.

 

Just to supplement the excellent information already provided, this is the deal with breadboards and wire gauges:

Typical solderless breadboards will accept wire gauges from 21 to 26 AWG. That's 0.016 to 0.028 inches or 0.4 to 0.7 millimeters in diameter.

Your wire is 0.6 mm in diameter. That actually falls between 23 AWG and 22 AWG. It's best to assume the more conservative 23 AWG specs. Wired in the open air, as opposed to through a conduit, 23 AWG will take a maximum of 7 A. The more its ability to dissipate heat is limited by being enclosed, the closer it gets to its 5 A maximum rating.

Assuming a 5 amp maximum current drawn by any given board and longest length of 3 feet, your cable would produce a voltage drop of roughly 0.3 V. That's pretty nominal, but if the tolerances of your components all conspire against you, it's something else to consider. Thicker wire (lower AWG) means less resistance in the wire itself. In a solderless breadboard, you can't use much thicker wire, but I suppose you could run two wires in parallel if it came to salvaging a project.

Just my two cents. I could be wrong. Actual mileage may vary. This bag is not a toy.

caltain

 

Depends on the distance between your power supply and your boards.

18AWG stranded wire will most certainly not plug into your bread board strips. That is why they put banana jacks on experimental bread boards.

 

Just for reference purposes, 0.6 mm (0.0236") is slightly smaller than 22 Ga American.
And caltain just got there before me.

 

[QUOTE="And caltain just got there before me.[/QUOTE]

I wasn't slowed down by needing to impart Truly Useful Information...

 

[QUOTE="And caltain just got there before me.

I wasn't slowed down by needing to impart Truly Useful Information...[/QUOTE]
LOL! You probably weren't slowed down by having to walk to the electronics room to read the wire gauge chart.

 

LOL! You probably weren't slowed down by having to walk to the electronics room to read the wire gauge chart.

Dude! There's an app for that!!

Actually, I use three, depending on my need and my preference of how each approaches the issue at hand.

I am not involved with either product in any way. I just find them extremely convenient and useful, hence the rest of the post...

I use Electronic Toolbox and RF Toolbox by Marcus Roskosch, both available on the iOS App Store. These are almost the same, but the RF Toolbox clearly has additional tools for RF and high frequency matters. (www.electronic-toolbox.com)

I also use the Electronics Engineering Toolkit Pro by Thomas Gruber, also available on the iOS App Store. There is also a Pro version specifically for the iPad which has a larger database of components and a few other features not in the universal Pro version. (www.ee-toolkit.com)