power distribution

Thank you, Sir! I never resent being kept on my toes! And yes indeed each of the four branches has its own dedicated fuse. Which reminds me... When choosing the optimal fuse amperage, am I selecting the closest available that is less than the actual load? I ask because 4A seems laughably small for the 50W HID, and/but 10A seems high for the 120W LED bar. OTOH, 10A seems just about right for the twin 55W Halogens. Am I using the right formula?

We have lift-off!

Do you want a chance to rewrite that before I answer? Ooops - too late! A fuse rated "less than the actual load?"... No - because then the fuse would constantly blow wouldn't it? (Ha ha, I finally get the chance to get sarcastic and dig the knife way in.... ) [ BTW - I rarely go to the trouble to include emoticons. To do so would require the utmost care and concern by me.... ] The "Rules": A fuse must be rated to handle the load - ie, at least equal to the load. Usually that is the max expected load - eg, at max volume, or max current (light output) at highest voltage - but you might decide lower is ok (eg, 5A for a 10A amp where you don't expect to go higher than ~1/4 max volume or half full output) - not that you can do that with lights. And there is a general fusing/protection/wiring design rule - namely that fuses (and wires & relays etc) should not normally run at higher than ~70% of its rating, though sometimes 90% may be designed (and sometimes 110% in reality - noting that a fuse or breaker may last indefinitely on a 110% loading (10% overload)).    Hence for a 50W HID, IMO... Assuming 50W HID means output hence ~60W input... Or using the usual ROT (rule of thumb) - divide the power by 10 to get current - hence 50W/10 = 5A. [ FYI - The 10 is simple, and it tends to factor in the "conversion" from output to input (at say 80% efficiency) or that 12V really means up to 14.4V etc and that rating may be based on old 13.8V car voltages if not newer 14.2 or 14.4V, or even 12.0V. ] So 5A. That fits in well with st'd fuse sizes (unlike 4A), but then if we apply the 70% rule it means a 7.5A or 10A fuse. But now the important thing - WHAT are we fusing? (Or rather, WHAT are we protecting?) It might be to protect the source - ie, battery or alternator - but that's not relevant here. (Let's assume some "master" upstream fuse or flink (fuselink) does that.) Usually it's to protect the load. And that's where people get con-fused. (Ah yes, a Master of Punnery.] The "load" is anything downstream from the fuse, and in our cases they are rarely for equipment protection - ie, HUs, amps, CPUs etc have their own fuses specifically designed to protect that equipment. Hence our protection/fusing is almost always for the distribution - ie, wires, connectors, relays. And the fuse rating (Amps) must not exceed the smallest rating downstream - ie, a 10A cable can have a 10A or 7.5A or 5A or 250mA fuse but NOT a 15A or 30A or 200A fuse. Note however that a fuse only protects its distribution until the next fuse. I'll call that segmentation... So, your design... You know you need (say) 5A for each 50W HID. Hence you need cable rated for (at least) 5A. You might use 5A cable and hence a 5A fuse. I'd probably use a 10A or larger cable but could then use a 5A or 7.5A or 10A fuse (anything up to the cable rating). Remember - that fuse is to prevent the wire from flaming if it shorts to GND. The fuse must blow before the cable has a chance to get too hot... And why use larger than necessary cables? To minimise the voltage drop. Or because it's what I happen to have on hand and isn't too expensive or big or heavy... Now, not that I know your design (hey man, emoticons are one thing, but to actually read what has previously been written in a thread... c'mon, get real!!), but... Two "5A HIDs" (ie, each with their 5A fuse & cable or 10A fuse & 15A cable etc) could be joined to a 10A distribution which might be a 10A fuse & 10A cable, or 15A or 20A fuse with a 20A cable etc. Likewise all loads can be joined. One danger you have is the splitting of the main feeder's core. Since they are not individually fused (upstream), if any were to short to GND or if an individual load exceeded that split's actual current carrying capacity... An outright short is probably not an issue since its current should be many times the upstream fuse rating, but it is any overload less than a direct short that is the danger. Provided a split's downstream fuse is rated reasonably below the split's capability, it may not be a big issue. IE 8G. Assume a 60A rating. 4 splits in theory about 15A each (not that that logic always follows!).   But you would not use a 15A downstream fuse... 10A maybe. The splits may be uneven, hence 13A or 12A etc. But also we do want the fuse to blow before the cable melts in an overload situation (ie, not a direct short). And since a fuse takes s time to blow at x overload and a cable takes t time to blow at the same x overload, we want to ensure the fuse time s is less than cable time t, hence in simple terms that the fuse is reasonably smaller than the presumed split rating. Simple eh? Of course most would say that such splits are unacceptable for the reasons I outlined and hence a DB (distribution block) or similar must be used - and professionals may have no choice on the matter - but I try to point out that "protection" can be physical instead of (or as well as) electrical, and there are different protection modes (overload, shorts, equipment protection). And I've seen enough installations that follow the rules that are IMO outright dangerous - eg, battery safety isolation switches on the +12V side, or useless and dangerous alternator-battery fuses as per big 3 upgrade on integra fuse box?. Even oft quoted rules like fusing within x mm/inches are not really rules at all - I prefer 'as close as practicable' to the battery (noting that practicable is the key word and subject to expert and legal opinion...). But I have written much on the above issues before - as well as on using self resetting circuit breakers for critical lights (and probably having them on separate distributions) - and I've repeated way too much here. As to your 10A for twin 55W HQs, you'll have figured that's probably a bit low (but see comments re my 10A CBs below). Though 55W HQs are 55W input, the 'div by 10' rule is till good as it factors in distribution NOT running at 100% of rating (tho the div-10 rule tends to assume EITHER an 80% efficiency ELSE a vehicle's voltage variation, whereas fuses etc might normally be consider a 70% loaded device). Hence 2 x 55W = 110W => 110/10 => 11A, hence a 15A fuse and cable. 10A may be ok, but why cut so close? Likewise the 120W LED => 12A => 15A distribution. Hey - does that work out nicely? All cabling from the feeder could be 15A (or higher) rated cable with 15A fuses. Of course that depends on how the 8G is rated and hence what each split can carry. (Different people use different cable current rating methods & tables. I never use them since I work from what I consider an acceptable overall voltage drop and that is always less than what the industry considers acceptable.) BTW - it doesn't matter if the total of the downstream (split) fuses exceed the feeder cable rating since the feeder's fuse (60A?) will protect that cable. Of course if that blows. you lose all. [ FYI - hence my car's main beam distribution via huge flinks intended never to blow (2 of; I can't recall if they are 50A or 60A or 100A; nor if I split as hi/low or left/right, tho probably the latter) which then feed relays each with self resetting circuit breakers. Though at the moment I'm using 3 relays for 6 filaments (4 lights), I intend to revert to my traditional dedicated relay per filament. The breakers are rated at 30A (for 2 100W inner highbeams) and 10A for the others that feed 65/55W H4s (normally they'd be 15A breakers assuming 100/90W or 100/55W H4s, but I have not had problems with mere 10A breakers (fuses) on what is obviously at least a 2x55W load (probably 55+65 = 120W or maybe even 65+65 = 130W). But I recall replacing the original 15A CBs with 10A CircuitBreakers to see what would happen... ] Anyhow, that's my design theory along with actual observations. BTW - my cabling well exceeds load requirements and fuse/breaker ratings. It met my design of no more than 0.5V less [across the bulbs compared to the source (ie battery or alternator output). Always nice writing a quickie before a sat'dy night gig. PS - I started writing this before your last reply. I wonder why I took so long to write?

Zounds! Thanks very much - I have to digest that. In fact, what I ended up doing is re-joining the 4-way split into one. (I had intended to do it like sketch #1, but it ended up as #2) I still haven't decided whether to install a main feeder fuse. It seems like overkill to me, but I suppose there's always a tiny chance that the main feeder could GRND somewhere between the battery & the split. I guess. < scuttles off muttering >

You *absolutely* want a main feeder fuse. You're protecting the battery (and thus the entire vehicle) with it. (unless this is within, say, a foot of the battery already). I have 3/0 welding cable going to the back of my truck for winch and jumpstart use. Since winching can induce loads up to 300A, I am running a 400A fuse at the battery- it's not there to protect the winch motor, it's there to protect the cable and vehicle in event of a short of the cable to chassis. While there's certainly some science in oldspark's approach, I disagree with a few points. Here's what I do: *Always* protect the distribution within a foot of the battery- for the same reason I fused my welding lead as above. In this case, the load doesn't matter. Take the wire gauge and the length (of entire circuit, ground and positive) and look it up on an ampacity table. Fuse should be a fast-blow sort, rated near the top of that range (a little over is perfectly fine, as there is a very large safety factor inherent when you're using fine-stranded cable, which you definitely should be). This protects the battery, cable, and your car, only. Fusing of loads is handled at the loads. To figure out the amps required for your load- take the watt rating, divide by 12 (nominal voltage in our systems, with rare exception). I don't know where he's getting the "divide by 10" bit. It's close, but not correct- Ohm's Law is what we're dealing with here. When you get a value that is between commonly available sizes- step up. In my case, I am running two 55W driving lamps. 110W total, obviously. Divide by 12, you end up with 9.2A. Well there's a 5, 7.5, and 10A fuse commonly available- I'm running the 10A and am perfectly well-protected. Additionally, I sized my load's supply lines to accept the biggest commonly-available bulb (there are 100W H3s on the shelf at every parts store), so if/when I change out for those, I will then have 200W, divide by 12 = 16.6A. I'll throw a 20A fuse in there and be done. Now in the event of a motor, there's Inrush Current to consider- among other things, you have to overcome the inertia of the motor, which requires a huge spike- but that's not what we're doing here. The only possible application (other than OEM starter motor) would be a winch. I don't know any manufacturer that fuses the car's starter, nor any winch manufacture that advises it. I've never seen a fuse on a winch. Another story altogether, though. Best guess is they're used so seldom, and are built for the abuse, that the fuse just isn't necessary. Yes you can get MUCH deeper into the science of this subject- but it's not rocket science. We're building a truck here, not a Tactical Data System on a Nuclear Aircraft Carrier. (I should know ;) ) "Always listen to experts. They'll tell you what can't be done, and why. Then do it. - Robert A. Heinlein"

BTW, I really like your work here. I might have done some things differently, but that's just in keeping with habit, nothing wrong with what you've done. In fact, I am going to steal an idea from you- the clear shrink over diodes. I've done some unnecessary rework on the present build simply because I couldn't remember if that was a solder joint or a diode under the heatshrink - I cut my diode leads short, and once my solder joint is made and under shrink it's impossible to tell (without a meter, which needed a battery) if there's a diode under there. I'll continue to use opaque shrink or tape on my connections, but the diodes will be shrunk with clear shrink to make this a no-brainer."Always listen to experts. They'll tell you what can't be done, and why. Then do it. - Robert A. Heinlein"

Actually, if I may- I would have done one thing differently: The solder and shrink trick to split the main feeder cable to the 4 loads is neat, and if space is an issue it sometimes has to be done- but in a vehicle there's always an alternative- move the split upstream to a physical location with room for the following, or to the battery terminal: I'd have installed an insulated stud (the sort I posted a link to earlier, from Blue Sea) and used ring terminals on all the connections, then bolted them together. This way you KNOW there is sufficient cable cross-section for you to just look it up on an ampacity chart and figure out the maximum current it can take. As you've done it, even if you were very scientific in dividing the cable into sections for the soldering, you still have some variance. I suspect that you eyeballed it, which is fine but you'll have yet more variance. While probably not an issue, it is possible that you may be exceeding the ampacity of the junction you've made. Would I lose sleep over it? Not really. But I wouldn't do it again. "Always listen to experts. They'll tell you what can't be done, and why. Then do it. - Robert A. Heinlein"

Oldspark mentioned designing for no more than a .5V drop across his lights. I like that approach, definitely agree. I went through this with my headlights several years ago- and realized that the factory runs the headlights from battery, through engine compartment into cabin, through the switch and back to the headlights- all via 16GA wire! When I realized this I backprobed the bulb when in use- I was only getting 9.5V to the bulb! This got me started on reworking the heavy loads in my vehicle, as the factory had done a sub-par job. I now have each filament of the OEM lights on its own relay and fuse. I could have ganged the low and high beams each on one relay- but then if I lose a relay I lose an entire circuit (thus, losing headlights while I need them. Yikes!) I definitely like a good CB, but for the expense and packaging concerns, a fuse works just fine."Always listen to experts. They'll tell you what can't be done, and why. Then do it. - Robert A. Heinlein"

Burnkat, With the ohms law, Oldspark is dividing by 10 instead of 12. Then he is sizing the fuse after that result which results in a value that is 20% higher. Makes the math simpler, and ending up with pretty much the same result. You take 110W divided by 12..get 9.2, fuse at 10A. He would take 110W divide by 10..get 11A, fuse at 15A. You are sizing your fuse close to the rated load. Look into a modern vehicle fusebox, you'll see circuits "overfused" according to you. Now we know a fuse will carry a bit more current before it blows. As long as the wire is of proper size, I see nothing wrong. Now I'd fuse the lighting circuits independently, but that's me. In the splitting of the feeder cable, 1-8ga = 2-11ga = 4-14ga. Now say you erred in splitting them evenly and one ended up 16ga. It's less than a 3" length of 16ga! I doubt you'd measure any significant voltage drop on that length at 30A. On a construction note somewhat related. For another trick with using the diodes especially when using individual terminals. Bend the diode leads in a "n" shape with the legs as wide as the outside of the coil terminals. Lightly scuff the outside surface of the coil terminals with small file. Do this near the base of the relay. Locate diode across coil leads (Convention says term 86 is more positive than term 85). Solder the diode in place, trim excessive length. This can be tried with socketed relays, but there may not be clearance when the relay is reinserted into the socket. Just a few thoughts. Mark

To address some of burntkat's comments... I explained why the 10 factor is used instead of 12. But it is also a simple initial design approach which often turns out to be quite accurate in practice. For the exact design you would not use 12 (I can't think of anyone that would use 12V as the design voltage for 12V automotive etc systems), but 12.7, or 14.4, or whatever minimum voltage is applicable (11.6V, 10.5V, 8V etc) and also factor in any inefficiency, and then whatever overhead you want. Plus allow for manufacturing tolerances, temperature, etc. All in all dividing by 10 (Volts) is easy and accurate in almost all situations. Only for borderline cases where f.ex the div-10 answer is close to a preferred value (eg, 14.5A or 15.5A) would you normally have to do the exact calcs to determine if a 15A cable & fuse will suffice, or if you have to go to 20A etc. FYI - Div-10 (or div-20 for 24V) is a common automotive design ROT and you should find it is quite common. It is not merely something I have concocted. And I think burntkat misunderstood what I wrote about not fusing within (say) 12" of the battery, else I expressed poorly. If burntkat is correct, then his starter motor cable has a fuse. (No?)    And burntkat will guarantee that the fuse in the link I posted is safer than no fuse....? (No?) And what distance will the "Rule" standardise on? I have often read legit sources that will claim 4" or 6" or 10cm etc (even if it means being on top of a battery!). I merely tried to qualify any so-called distance rule so that it/they are not blindly followed. (Did I mention physical security/protection? Starter? No?) And burntkat's first line omitted that the main feeder fuse is also to protect the feeder cable (though that should be clear from his later statements - if not mine). But usually cable protection is the sole reason for fusing because many projects like this use cables too small to be much threat to the battery whereas if that cable heats or melts, you're in big trouble. Re acceptable voltage drops, I was shocked to find that many automotive electrics references reckoned that "up to 3V" is acceptable (without any qualification). And that seems to be a traditional view... These days with HIDs and (ballasted aka SMPS regulated) LEDs etc it's not as important (provided their highest current at lowest voltage if a constant power load is taken into account). I may have decided my 0.5V max drop when rally driving where a 1V drop to halogens meant a significant drop in light output (far greater than for incandescents), but I use 0.5V as max desirable drop for heavy and such 'critical' loads. I might tolerate 1V when impractical to get less (not that I ever have). For things like plain LEDs or bulbs or fans, a 3V drop may be acceptable. But 3V means 9.5V from a typical loaded fully charged battery (~12.5V ), or ~11.5V from a typical max alternator voltage (14.4V). Is that acceptable? Re CBs - it costs under $10 for a (self resetting) ATS CB up to 30A (or mini-ATS CB up to 20A) which - being ATS - generally has no packaging (housing) concerns. Likewise under $10 for up to 50A but they are typically stud terminals with mounting flanges to bodywork etc. CBs larger than 50A can get expensive and awkward as burntkat suggested, but I have never needed them that big. My largest self resetting CB for lighting (where you don't want fuses unless redundancy or fallback is provided) is presently 30A because of my single relays feeding 2 filaments, but normally it'd be 10A or 15A for 12V 100W or 130W filaments. Otherwise my largest is 50A for my 2nd battery - I found that 30A ATS fuses would occasionally blow and - not being alarmed - meant risking a flattened 2nd battery or a warm fridge. And I'd only have to blow one flink to make a profit from that CB.    BTW - I prefer one relay per filament merely for alignment purposes, and sometimes for testing. Relays themselves are very reliable - moreso than filaments and most switches. And tho the self-resetting CBs may provide enough circuit recovery, my redundancy is mainly through source and signal splitting - eg highs & lows on separate main fuses and feeds (if using one relay for both sides); highs are the fallback for blown lows, and the flasher (pass) is the fallback for beam (hi & lo) signal failure. I've avoided details for brevity but hopefully you'll understand my clarifications and maybe find support for the div-10 ROT. PS Mark - I started my reply before your reply. Thanks for the "short bit" note on the split. I have oft pointed out that consideration yet overlooked mentioning it here. Bless clear thinkers and proof readers! And thanks for the rest.