Voltage loss over many circuits

"eromlignod"

** You need to have multiple DC power feeds to all those modules - not just a single bus.

Break them up into groups of say 10 or so and run a cable back to the PSU for each group.

..... Phil

Wow, this is great! I'd try and lay things out so that each circuit has about the same amount of resistance between it and the power supply. Then you can add a sense line back to the PS to adjust the voltage. (but only one and it might not be necessary.)

George H.

A

not

Can you mix more than one 'sense' feedback line?

George H.

"George Herold"

** Piss off fool.

It's only 15 amps. How are the connections being made between the supply and the boards? How are they grounded?

The best wey to distribute power like this is to use a higher voltage,

24 maybe, and use a "point of load" switching regulator on each board. But that may not be feasible at this point and, at 15 amps, isn't really necessary.

#10 wire is about a milliohm per foot. Four feet at 15 amps will drop

60 millivolts, not enough to bother logic. So your problem isn't purely wire resistance.

John

John

A quick hack maybe to try to equalize the voltage using resistors. Attempt to add a small resistance to the closer loads to drop the voltage to about the same as the furthest ones then up the voltage from the supply slightly if you can. This may be next to impossible to do effectively if the topology if the circuit does not facilitate any easy estimations of the resistances.

Alternatively you can divide the circuits into several groups and go from there... as phil has suggested.

Or if you got a lot of copper to spare, use the same length wires, one for each circuit. The is effectively the above suggestion but with groups of size 1 but doesn't require any calculations or anything since every circuit board will have the same drop due to the same length wire(and guage of course).

Really it's going to depend on your topology. If it a bus with very short stubs then the drop should be relatively linear and can be computed by finding the relative distances. The resistance to add would be quite simple to do(although it depends on the final power supply voltage which is unknown but can be estimated).

Of course the optimal way would have been to add regulators/converters on the boards from the getgo.

What I have done in those cases is mount copper rails(~3mm by 20 mm) like they use in power stations. then you only need 2 or 3 short wires the each connector.

As others have pointed out, point-of-load regulators solve the problem nicely but your system is of such small scale that shouldn't be necessary. Using thicker wire for a trunk line (10 gauge copper) and letting the last-few-inches stay at a more convenient size will work well. You can also configure the trunk as a loop (put ferrite beads on if the ground-loop pickup or radiation is a problem); this doesn't use much extra wire, but halves the copper resistance.

Is this for your self-tuning piano?

ticle.PDF According to the article at that link, a piano could pull as much as

500 or 600 watts. That's a lot of power and you will get a voltage drop in inverse proportion to power supply voltage, as I'm sure you know. Are you running the heater current through the same wires that power the logic chips? Maybe you should try to find a way (power-wise) to separate the brawn from the brain, as it were. You could run your wire-heaters at a higher voltage than the logic and use local voltage regulation for the logic. Or use remote voltage sensing -- a dedicated wire between the power supply and the load that doesn't carry any power.

if resistive loss is the wire is the cause then increasing the amount of copper is the solution.

If you have the system wired so that the second unit is powered by short wires from the input termials of the first and the third is likewise powered from the second your 90th unit has 89 short wires (that probably total much more than 48" ) and also 90 mechanical joins that probably have higher resistance than than short wires themselves between it and the powerssuply.

Adding power feeds that go direct to the supply every 10 or so units (like Phil suggested) will alleviate most of that.

As others have suggested swapping to a bus-bar system where each unit connects directly to metal rods would also work, but at 15A smells like overkill to me.

--- news://freenews.netfront.net/ - complaints: snipped-for-privacy@netfront.net ---

a

Am wondering if the problem is not with the 5 volt 'feed' side??????? Maybe it's with the 'return' side? maybe the wire gauge or metallic path of the return needs to be looked at?

100 units at 150 ma is only 15 amps (total) and normal gauge wire Such as 10AWG or 12AWG; unless something stupid like 26 AWG has been used to a carry the whole 15 amps, has very low resistance per foot etc.! Something weird about this question. Assuming the 5 volt supply, is, say the positive? How is the negative side wired; through a frame 'ground' or equipment case or something with a significant resistance path? This perhaps somewhat similar to a motor vehicle when there is a 'bad' ground to the vehicle frame due to corrosion etc. Suggestion anyway!