DIY Busbar

[nebster]

+1 for real life experience it makes things so much more simple. One question I have though is if the battery is in open air or the cable will it reach a max heat or could it theoretically heat up forever. Lets just pretend the ambient temp is a constant 70 degrees. Also pretend the heat radiated from the bus bar will not increase the rooms temperature.

Usually the ampacity tables are built based on an assumption of "still air", which means convective flow is allowed. The air around the wire will heat, which will make it float away to be replaced by more ambient air at the "constant" 70deg.

If your wires are in a small compartment that's insulated, all bets are off: the air will just keep on heating up, until the airflow and conductive losses to the enclosure increase enough to reach a new equilibrium.

In practice in our leaky battery compartments and decent-sized rooms and vehicles, I don't think that's typically something we have to worry too much about. Other factors probably heat up or chill the enclosure faster than our electronics. But if your spidey instincts suggest to you that your situation might be special, it's definitely worth doing some more that real-life testing and measuring. IMHO.

 

[Airtime]

Good idea, do some math. But I think you may have missed some important things.

First of all, in a uniform wire, length is irrelevant. A longer 4/0 will have more resistance, and a shorter 4/0 will have less resistance, but also each will have different mass and different radiating surface area. This is why ampacity tables aren't parameterized on wire length -- it's irrelevant, at any normal lengths anyway (including 3 inches, in all likelihood).

Second of all, 2.5W is plenty of power to accumulate heat over time. You could boil a gallon of water with 2.5W if you had enough insulation. You simply must model the radiative and conductive losses on the item of interest to understand where it will end up.

But that modeling is tricky, you will definitely have to involve partial differential equations and start making estimates, and more importantly, that is what the ampacity tables do for us! Empirically, and probably analytically since it is such a common and interesting problem, people have determined the steady-state thermal balance for these common cross-sections. They are telling us, at a minimum, what continuous current will produce thermal rise below the insulation safety threshold, starting at standard ambient conditions. (Many tables also have correction factors for higher ambient, but if you do the math there you will find that their corrections correspond very closely to a simple I^2 ohmic adjustment.)



Great, even more heat! And even more reason to measure the temperature on your actual prototype conductor under the actual max-continuous for your battery, if you're going not going to take the overkill route or use a proven design out of the box.

I'll give you a practical example. I am commissioning a small 4p4s 12V battery right now, for a friend, on the bench. I am literally standing right next to it, right now, typing this while it is undergoing its first full discharge. It is built from the EVE 280Ah cells many of us are using. It is using the Xuba-provided, 30mm^2 cross sectional nickel-plated copper bus bars that are, frankly, a little wimpy. I have two of those, stacked, as the sole connection between each 4p module of cells.

We're discharging today at 130A. An ambient bus bar sitting on the table right next to me is currently measuring 69F. The double-stacked (60mm^2) bus bar in the circuit is sitting stable at 93F after several hours of steady current. So the rise is 24F, and I didn't have to do any math --thank goodness; math is hard!

Our design limit is 240A continuous. At 240A, I expect ohmic heating in the double bus bar to follow the square curve, or (240/130)^2 = ~3.4. The temp rise at our design goal would be 82F. In ambient conditions, that bus bar might then be at something like 157F. That is too hot for our application and for my taste with lithium batteries for the reason I mentioned earlier. So, we will use a substantially larger cross section when we deploy the battery in real life.

I understand resistance vs length. My point is that models may assume a long bar as a simplifying (and conservative) assumption. However, when you have a 3" bus bar heat sinking into large battery bank, those models will overstate temperature rise.

In the real world, 2.5W will feel hot to your finger if concentrated into a small part like a resistor. But you can comfortably wrap your hand around a 2.5W LED bulb. When heat sinking into a 100+lb battery pack, 2.5W will not feel hot. Unless of course you surround it with perfect insulation and no airflow so you can boil a gallon of water after 6 days

No need to use differential equations or run Flotherm. Just use the ABYC tables, go up a size if you want to be conservative, and yes test it under load.

So what size bus bars are you going to use? Looks like you're going to select it by prototyping and measuring ?

 

[nebster]

I understand resistance vs length. My point is that models may assume a long bar as a simplifying (and conservative) assumption. However, when you have a 3" bus bar heat sinking into large battery bank, those models will overstate temperature rise.

Indeed, they dump the heat into the battery for sure. But... that is going into the one cell. Not a "100 pound pack." Right?

In the real world, 2.5W will feel hot to your finger if concentrated into a small part like a resistor. But you can comfortably wrap your hand around a 2.5W LED bulb. When heat sinking into a 100+lb battery pack, 2.5W will not feel hot. Unless of course you surround it with perfect insulation and no airflow so you can boil a gallon of water after 6 days

A 2.5W LED bulb has a huge surface area for your hand to soak heat out of. Try holding a 2.5W LED running at rated power. It will be blazing hot. Now, put one of those on each of your battery cells. It's not necessarily intuitive, nor is there a single recipe that will be right for all cases, as usual.

No need to use differential equations or run Flotherm. Just use the ABYC tables, go up a size if you want to be conservative, and yes test it under load.

I disagree. I think one size up is not enough for a continuous load at the rating. 10 or 20C is about as much as I would find acceptable in a living setting with lithium, and one size up is still a +35C rise at the limit.

Most people won't run continuous to the rating, and so there will be a huge safety buffer there. But by the letter of the table... it's still too much for a safe application.

So what size bus bars are you going to use? Looks like you're going to select it by prototyping and measuring ?

I'm going with 1/4" x 1" aluminum, or a 160mm^2 cross section. I went with that because it happens to sit on terminals nicely, it's what I can machine comfortably and relatively easily, and I can get it quickly and with rounded edges. And, yeah, I'll be putting the thermal camera on the pack when it's built up, just like always. Moreso to verify everything is cleaned and torqued properly than to verify the heating, because those bars are borderline overkill for this application.

For my own 16s pack, I can get away with using much less material for the same temps, but the really small stuff is not much cheaper, and I'm no machinist!

Cheers.

 

[Airtime]

A 2.5W LED bulb has a huge surface area for your hand to soak heat out of. Try holding a 2.5W LED running at rated power. It will be blazing hot. Now, put one of those on each of your battery cells. It's not necessarily intuitive, nor is there a single recipe that will be right for all cases, as usual.

That's exactly my point. A cell is a lot bigger with more mass and more radiating surface than an LED. It will feel less warm than a 2W LED bulb much less an LED.

I disagree. I think one size up is not enough for a continuous load at the rating. 10 or 20C is about as much as I would find acceptable in a living setting with lithium, and one size up is still a +35C rise at the limit.

Most people won't run continuous to the rating, and so there will be a huge safety buffer there. But by the letter of the table... it's still too much for a safe application.

Well we'll just disagree. A marine environment is the most hostile and unforgiving there is, outside of spacecraft. I'm comfortable with the level of safety defined in ABYC standards.

I'm going with 1/4" x 1" aluminum, or a 160mm^2 cross section. I went with that because it happens to sit on terminals nicely, it's what I can machine comfortably and relatively easily, and I can get it quickly and with rounded edges.

Now we agree ?. Pretty much the thought process I followed in selecting my 1/4" x 1" tin plated copper bus bars. Actually I selected that busbar material when I was still planning a 12V pack. I later changed to 24V 1P8S to make the cables smaller and fully use my Victron Multiplus 3000 capability. So now I have overkill^2. 3 inches of 160mm^2 Cu at 140A continuous (0.5C) is about 0.17W. I think we can both agree that will not get hot

 

[nebster]

Well we'll just disagree.

Sure, no worries; your stance is not unreasonable.

You didn't respond to my important clarifying question, though: we are talking about multiple bus bars, not a single one dumping its heat into a massive battery pack. If you're going to leave an order-of-magnitude factor hanging out there, I want to understand why you think it works that way. I want to make sure I'm not missing something important.

A marine environment is the most hostile and unforgiving there is, outside of spacecraft.

I can think of several that are substantially more hostile than marine. Overhead/caves. Salt flats, with their conductive alkali dust that can get into every crevice and eat electronics for breakfast, lunch, and dinner. Thermal vents on the ocean floor. But, sure, your point is taken.

I'm comfortable with the level of safety defined in ABYC standards.

It's really not even about safety as much as it's about maximizing battery lifespan. ABYC has demonstrated ignorance (and sluggishness) with respect to battery, especially lithium. Half the people on the prior year's working committee have relationships or work for companies that make battery products that have huge, glaring flaws in their designs. It's just a tricky space without a lot of stability, and ABYC does better with low-inertia concepts. That's pretty common for safety standards organizations.

Put another way, I wouldn't expect ABYC's ampacity table to expose the optimal strategy for making lithium iron phosphate battery interconnects run best. It looks a lot like a plain NEC-derived one if you ask me.

Now we agree ?. Pretty much the thought process I followed in selecting my 1/4" x 1" tin plated copper bus bars. Actually I selected that busbar material when I was still planning a 12V pack. I later changed to 24V 1P8S to make the cables smaller and fully use my Victron Multiplus 3000 capability. So now I have overkill^2. 3 inches of 160mm^2 Cu at 140A continuous (0.5C) is about 0.17W. I think we can both agree that will not get hot

Yes, both of our interconnects are going to stay super cool, but not because of the lower watts alone. It's the watts distributed across the radiative surface that makes them so effective!

I'm curious why you say the MP 3KVA wasn't fully usable at 12V. Can you share more about that?

p.s. Did you know that insulated wires stay cooler than bare ones of the same cross-section, at the same current? Or that vertical bus bars have greater ampacity than horizontal ones?

 

[Airtime]

Sure, no worries; your stance is not unreasonable.

You didn't respond to my important clarifying question, though: we are talking about multiple bus bars, not a single one dumping its heat into a massive battery pack. If you're going to leave an order-of-magnitude factor hanging out there, I want to understand why you think it works that way. I want to make sure I'm not missing something important.

Yes that's right, in my 1P8S setup there are 7 bus bars so there would be a total of 17.5W of heat sinking into my 100 lb pack, about 1 cubic foot, if I was actually running 445A continuous. 17.5W in that volume is not that much. Unless of course you put it in a hermetically sealed highly insulated box, and why would you do that?

I can think of several that are substantially more hostile than marine. Overhead/caves. Salt flats, with their conductive alkali dust that can get into every crevice and eat electronics for breakfast, lunch, and dinner. Thermal vents on the ocean floor. But, sure, your point is taken.

No sane person lives on the salt flats for years. Thermal vents and caves are silly examples and not relevant. Marine applications are continuous for many years, with humidity and corrosion, pounding seas, close proximity to living quarters, and if there is a fire you have few options to escape.

It's really not even about safety as much as it's about maximizing battery lifespan. ABYC has demonstrated ignorance (and sluggishness) with respect to battery, especially lithium. Half the people on the prior year's working committee have relationships or work for companies that make battery products that have huge, glaring flaws in their designs. It's just a tricky space without a lot of stability, and ABYC does better with low-inertia concepts. That's pretty common for safety standards organizations.

Yes, I've heard some similar things. But we're talking about basic things here like ohm's law and heating. Here's a good experiment, take a 17W heater and attach it to a 100 lb 1 cubic foot disconnected battery bank and let it come to equilibrium. It will not be very hot.

Yes, both of our interconnects are going to stay super cool, but not because of the lower watts alone. It's the watts distributed across the radiative surface that makes them so effective!

Yes, again exactly my point!!!

I'm curious why you say the MP 3KVA wasn't fully usable at 12V. Can you share more about that?

Victron specifies a 400A fuse for 12V and 300A fuse for 24V. Don't know why, they really should have a 2X relationship. I'm guessing that it's because practically speaking you can't get things like Class T fuses higher rated than 400A. Or maybe internal wiring ampacity.

But anyway 400A at 12v is of course 4800W, which is a bit shy of the 6000W 2 minute surge that the Multiplus is capable of. At 24V, 300A is 7200W. So at 24V you can actually use the rated surge capability without risking blowing that big fuse.

I don't really anticipate testing that limit, but... I'm kind of a geek and it pains me to have that capacity and not be able to access it. Who knows, my wife or I may sometime have the induction cooktop on high, microwave on, hot water heater running, and decide to make an espresso . Some things are just not rational.

p.s. Did you know that insulated wires stay cooler than bare ones of the same cross-section, at the same current? Or that vertical bus bars have greater ampacity than horizontal ones?

No and no. I'm guessing it's because the insulation surface has higher emissivity than bare copper, and vertical bus bars set up a chimney-like convection that cools them better. Did I pass?

 

[nebster]

Yes that's right, in my 1P8S setup there are 7 bus bars so there would be a total of 17.5W of heat sinking into my 100 lb pack, about 1 cubic foot, if I was actually running 445A continuous. 17.5W in that volume is not that much. Unless of course you put it in a hermetically sealed highly insulated box, and why would you do that?

Lots of people insulate their batteries for cold. My battery bay in my RV is fully insulated, for example, to the point where I have forced-air cooling routed down to it. It's well within the bounds of reasonableness for that to be the case for someone's DIY pack. (Btw, I heat that bay with a 40W reptile heat pad in the dead of winter at 0F -- a condition very few marine pleasure vessels ever experience or design for, btw.)

No sane person lives on the salt flats for years. Thermal vents and caves are silly examples and not relevant. Marine applications are continuous for many years, with humidity and corrosion, pounding seas, close proximity to living quarters, and if there is a fire you have few options to escape.

The thermal vents were tongue-in-cheek, but lithium battery systems are most definitely engineered for caving, and it only takes 24 hours on the wrong salt flat in the wind with the wrong design and you can wreck some very expensive electronics (and motors for that matter). They are incredibly aggressive environments, particularly for electrical systems, and I would argue much more challenging to design for than an average day on the typical milk run route.

Yes, I've heard some similar things. But we're talking about basic things here like ohm's law and heating. Here's a good experiment, take a 17W heater and attach it to a 100 lb 1 cubic foot disconnected battery bank and let it come to equilibrium. It will not be very hot.

I guess you are assuming the battery bank perfectly dissipates the heat with no gradient or local heating effects? Surely that's not the case in real life.

I'm standing here, next to the same battery bank I mentioned up thread, running the same 130A through the same busbar I mentioned before. The bar is sitting at 93F right now. The air is 70F, and the plastic surrounding the aluminum cell shells is at 73F (well, actually, 72F on the outside cells and 74F on the inside ones!).

The gradient between the bus bar and the plastic wrapper must include all temperatures between 93 and 73, right, by the second fundamental theorem of calculus? So how is it that we can be sure that the top inch of "jelly roll" under the hot bus bar is not at 92F, or 90F, or something substantially greater than 73F?

Now, double the current. The new ohmic heating is something capable of pushing the bar closer to 160F. (Yes, that comes close to matching the data in the vaunted ABYC table.) The problem is, what is the temperature of the battery anode, cathode, and electrolyte inside the case, right below the terminal?

It defies reason that it will be at an acceptable level for long-term health of at least part of that cell.

I feel like I'm somehow not making the concern clear, so I don't know if this example muddies the water or clarifies what I'm getting at. But this is not just about safety, it's about best practice for managing lithium chemistries.

But anyway 400A at 12v is of course 4800W, which is a bit shy of the 6000W 2 minute surge that the Multiplus is capable of. At 24V, 300A is 7200W. So at 24V you can actually use the rated surge capability without risking blowing that big fuse.

Ah, got it. Yeah, the theoretical surge capability of the coil is impressive, indeed, and it's interesting that they basically end up derating their own unit in the 12V model. I bet it's because of that fuse practicality, too.

Anyway, that is only "recommended" fusing, not required, so presumably you could go bigger with sufficient wiring (or sufficiently-short peak draws). Performance envelopes are complex and multivariate, though; a simple "we can do this many KVA for n seconds" grossly oversimplifies it. My experience with Victron inverters is that they perform well below what their marketing sheets state in some cases. (I wouldn't be surprised if you could get 6KVA for far less time if you don't have the unit installed in a freezer, for example.)

No and no. I'm guessing it's because the insulation surface has higher emissivity than bare copper, and vertical bus bars set up a chimney-like convection that cools them better. Did I pass?

You got 50%, enough for a... silver star?

The vertical bars indeed do facilitate faster convective action.

The insulated wire performs a bit better because the exposed surface area is larger.

 

[Airtime]

Lots of people insulate their batteries for cold. My battery bay in my RV is fully insulated, for example, to the point where I have forced-air cooling routed down to it. It's well within the bounds of reasonableness for that to be the case for someone's DIY pack. (Btw, I heat that bay with a 40W reptile heat pad in the dead of winter at 0F -- a condition very few marine pleasure vessels ever experience or design for, btw.)



The thermal vents were tongue-in-cheek, but lithium battery systems are most definitely engineered for caving, and it only takes 24 hours on the wrong salt flat in the wind with the wrong design and you can wreck some very expensive electronics (and motors for that matter). They are incredibly aggressive environments, particularly for electrical systems, and I would argue much more challenging to design for than an average day on the typical milk run route.



I guess you are assuming the battery bank perfectly dissipates the heat with no gradient or local heating effects? Surely that's not the case in real life.

I'm standing here, next to the same battery bank I mentioned up thread, running the same 130A through the same busbar I mentioned before. The bar is sitting at 93F right now. The air is 70F, and the plastic surrounding the aluminum cell shells is at 73F (well, actually, 72F on the outside cells and 74F on the inside ones!).

The gradient between the bus bar and the plastic wrapper must include all temperatures between 93 and 73, right, by the second fundamental theorem of calculus? So how is it that we can be sure that the top inch of "jelly roll" under the hot bus bar is not at 92F, or 90F, or something substantially greater than 73F?

Now, double the current. The new ohmic heating is something capable of pushing the bar closer to 160F. (Yes, that comes close to matching the data in the vaunted ABYC table.) The problem is, what is the temperature of the battery anode, cathode, and electrolyte inside the case, right below the terminal?

It defies reason that it will be at an acceptable level for long-term health of at least part of that cell.

I feel like I'm somehow not making the concern clear, so I don't know if this example muddies the water or clarifies what I'm getting at. But this is not just about safety, it's about best practice for managing lithium chemistries.



Ah, got it. Yeah, the theoretical surge capability of the coil is impressive, indeed, and it's interesting that they basically end up derating their own unit in the 12V model. I bet it's because of that fuse practicality, too.

Anyway, that is only "recommended" fusing, not required, so presumably you could go bigger with sufficient wiring (or sufficiently-short peak draws). Performance envelopes are complex and multivariate, though; a simple "we can do this many KVA for n seconds" grossly oversimplifies it. My experience with Victron inverters is that they perform well below what their marketing sheets state in some cases. (I wouldn't be surprised if you could get 6KVA for far less time if you don't have the unit installed in a freezer, for example.)



You got 50%, enough for a... silver star?

The vertical bars indeed do facilitate faster convective action.

The insulated wire performs a bit better because the exposed surface area is larger.

Dang, I thought about the surface area! Actually I meant total emission factoring in area! Er...

Anyway, not trying to be too flippant. I do get that there are insulated boxes and in true continuous high current applications you want to be careful. In my own application I'm about 4X oversized, as I said overkill. 160mm^2 copper bus bars for maybe 150A max in real life and not that long a duration.

Edit--I'm going to claim partial credit, clean Cu emissivity is 0.02 - 0.06 and plastics are 0.9 - 0.97. For thick cables insulation is a small percentage of diameter and emissivity dominates ?‍?
https://www.engineeringtoolbox.com/emissivity-coefficients-d_447.html

 

[ghostwriter66]

and yet - after all the math and engineering -- I still somehow just stop by Home Depot and get 2" copper plumbing pipe - smash it flat - cut it to size - drill two holes -- and boom - busbars good enough for what I do ... LOL

BUT all joking aside -- this was a very good discussion thread....

 

[Luthj]

Lots of folks use "looks big enough" CCA wire for everything. Sure it works fine for intermittent loads, but some applications will suddenly need the rated current of the circuit. Then the crap hits the fan.

 

[nebster]

Anyway, not trying to be too flippant. I do get that there are insulated boxes and in true continuous high current applications you want to be careful. In my own application I'm about 4X oversized, as I said overkill. 160mm^2 copper bus bars for maybe 150A max in real life and not that long a duration.

Yes, for most of us, our design peak currents will not be continuous and/or may never be achieved in real life. And many of us will just overkill since we are not often cost-engineering here.

Edit--I'm going to claim partial credit, clean Cu emissivity is 0.02 - 0.06 and plastics are 0.9 - 0.97. For thick cables insulation is a small percentage of diameter and emissivity dominates ?‍?

Yes, but radiative losses are not the dominant heat transfer mechanism in most wire (unless it is glowing or you are in outer space!). Convection is the big contributor.

You lose your silver star for being obnoxious.

 

[nebster]

and yet - after all the math and engineering -- I still somehow just stop by Home Depot and get 2" copper plumbing pipe - smash it flat - cut it to size - drill two holes -- and boom - busbars good enough for what I do ... LOL

I actually took a piece of copper pipe and tried to make a bar out of it after I read your story a while back of doing this, since pipe is so easy to get and time is of the essence over here.

What I found was that I could not machine the through-holes with confidence. I need two edges, and preferably three, to be square so that I can fit the bar into the jig on the drill press to get precision spacing.

I have about 100 bars and ~250 holes to drill, so my time economics may be a little different than someone building a smaller pack.

 

[John Frum]

And many of us will just overkill since we are not often cost-engineering here.

^^^^^

 

[johntaves]

I found the bolts on ebay. I was surprised how difficult it was to find them.

Do you have a link to the ones you used?

I third this. Haugen, please do share.

 

[Ampster]

What I found was that I could not machine the through-holes with confidence. I need two edges, and preferably three, to be square so that I can fit the bar into the jig on the drill press to get precision spacing.

I am in the same process dilemma. I have a 2P16S pack so some of those buss bars have four holes. to further compliment things my pack is vertical rows and columns and some of the spaces between cells have threadrod so half of them have different spacing in one direction. I have cut some of the buss bar stock on my table saw in one dimension. I have procrastinated this part until the weather gets cooler.

 

[nebster]

I'm making a 3p16s "on end," so the parallel/cross bars have three holes each. The series bars are all 2-hole spans but the "long way". I fortunately don't need to cut any 4-hole bars, but I do need to set up 3 jigs/stops for the various configs I have.

I'm planning on drilling ~M8 holes to leave some slop for shifting/expansion and to accommodate the errors in my amateur machining operation. Hopefully those will be large enough.

 

[Craig]

I'm making a 3p16s "on end," so the parallel/cross bars have three holes each. The series bars are all 2-hole spans but the "long way". I fortunately don't need to cut any 4-hole bars, but I do need to set up 3 jigs/stops for the various configs I have.

I'm planning on drilling ~M8 holes to leave some slop for shifting/expansion and to accommodate the errors in my amateur machining operation. Hopefully those will be large enough.

I had to drill 10 holes in 11 seperate buss bars. I used a Jig for everything but still found I had to enlarge a few holes. Next time I will just use a larger bit to begin with. There is really no need to be so tight.

 

[Haugen]

I third this. Haugen, please do share.

This should get you started.
There are hex head versions available as well if you dig around.
Be sure to get the right length. The data sheet for your cells will tell you how deep the holes are, or you can measure it. Try to get your stack-up of bus bar, washer(s) figured out so that you have at least 3 threads of engagement. With the 1mm thread pitch, that means about 5mm.
You may need different length screws in certain places such as the end terminals if you don't use bus bar material to fill in the space.

Aluminum Alloy 6061 Allen Bolt Hex Socket Cap Head Screw M2.5/M3/M4/M5/M6/M8/M10 | eBay

Material: 6061 Aluminum Alloy.

www.ebay.com

 

[Bob B]

@Haugen .... did you use a flange nut with those or just the bolt?

 

[Haugen]

@Haugen .... did you use a flange nut with those or just the bolt?

I'm using the hex head bolt with aluminum washer.

ALUMINUM FLAT WASHERS 1/4" I.D. 11/16" O.D. 40 PCS. NEW | eBay

FOR SALE ARE 40 SOLID ALUMINUM FLAT WASHERS. 1/4" I.D. 11/16" O.D. NEW.

www.ebay.com

1/4" ID will work well for 6mm.
I would also consider using M6 threaded rod and a nut with the washer. It's a lot more expensive, because you have to get a nut also, and you have to rod protruding above the nut, so not an option I prefer.

1pcs M6 aluminum thread bar rod stick full tooth AL alloy cudgel 1.0mm pitch | eBay

Find many great new & used options and get the best deals for 1pcs M6 aluminum thread bar rod stick full tooth AL alloy cudgel 1.0mm pitch at the best online prices at eBay! Free shipping for many products!

www.ebay.com