Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Connecting battery to Circuit breaker
#11
(03-10-2019, 09:00 AM)Sean Wrote: Resistance isnt the reason why inverter battery cables need to be the same length, and run parallel, bound together where possible.

Inductance is ..... 

Inductance, and the associated ripple in DC inverter battery cables reduce efficiency and will lead to the premature demise of your inverters input stage.

It's long been best practise to keep all aspects of battery to battery, and battery to inverter cabling as symmetrical as possible.

In a system using cascaded, or inverters in parallel, identical battery cable length is vital.

Those cheap high output inverters, that folks seem to favour on here, really dont like DC ripple.

Can you point to data supporting this?
The DC input ripple is general caused in the inverter (by the design) and is independent of the cables.
There would be so little inductance and the distances and because they are running in parallel you won't see induced currents that would cause any problem that isn't already a result of a poor inverter design.

The real reason is increase in resistance that can cause one cable to heat differently.  But hey that is what sizing is for.  Longer runs should have heavier gauge wires to keep the resistance down.

The cables need to be sized to over-current substantially but have proper fusing.  That's what fuses are for.  To fail first well in advance of the cables and power source.

Am I wrong?
Reply
#12
(03-10-2019, 01:36 PM)Bubba Wrote: Can you point to data supporting this?
The DC input ripple is general caused in the inverter (by the design) and is independent of the cables.
There would be so little inductance and the distances and because they are running in parallel you won't see induced currents that would cause any problem that isn't already a result of a poor inverter design.

Am I wrong?

You are mostly correct, based on the DC feeds running mostly parallel (as per my bound together comment) - obviously running parallel implies equal lengths.

Induced current is opposed to current flow, and often results in DC ripple voltages transients in excess of the connected equipments component voltage ratings - this is especially true during rapid load fall offs. This is likely the reason less well engineered equipment fails without apparent reason.

I'm not generally keen on spoon feeding folks snipits of info as it's often applied incorrectly, without an understanding of why .... but I'm allways happy to encourage folks to read around a topic.

Here are two links - one is slightly longer than a snippet, the other somewhat longer ... being Daniel Fahlgrens MSc work with Volvo 

https://www.jamestowndistributors.com/us...docId=1018

(03-10-2019, 01:36 PM)Bubba Wrote: The DC input ripple is general caused in the inverter (by the design) and is independent of the cables.

I omitted to directly address this point - whilst the input ripple is an artifact of the inverters design, the magnitude of it (both in terms of voltage and current) can be greatly exacerbated or ameliorated by input cable length (mismatch) and routing.
Reply
#13
If you have no inverter and are running lights, boost or buck units etc. forget about matching cable lengths.

If you have a single properly designed inverter with sufficient internal capacitance, forget about matching cable lengths.

If you have multiple parallel inverters with a poor design and non isolated communication links between inverters then basic cable length imbalance can be a massive issue because it will crate a DC offset in the communication links, potentially ending up with the wrong level being sent between inverters resulting in oscillation or incorrect overall output level, as happens with some.


I fail to see, techincally, how any small inline cable inductance for a properly designed inverter would cause any form of damage as the inductance on a relative basis is extremerly small, and probably not detectable with the average digital multimeter. The ripple caused by the battery resistance will normally be at least an order of magnitude larger.


Lets take a 2m length (imbalance) of 30mm2 cable (6mm diameter) carrying 100A pulsed at 100Hz (as would be seen by the battery if there where no smooting capacitors at all in an inverter - really bad design that even the Chinese are yet to match).

The inductance of that cable is then 2.58uH
https://www.eeweb.com/tools/wire-inductance

Next up we apply the formula to work out the effective resistance
2 x PI x L x f = 2 x 3.141 x 0.00000258H x 100Hz = 0.00162 Ohms

Applying a 100A peak current the resulting voltage is 0.162 Volts and dont forget this is also out of phase... !!!!!!


If you share a common -ve return path between connected electronics (fed by different power cables) and there is a DC bias between the -ve lines then this "may" create problems for the electronics involved if the bias is large enough. On typical small wires at currents in the order of 1A the bias can effectively be treated as zero. Start moving 50A around and it shows up.

Inductance, treat as zero, unless proven otherwise.
Bubba likes this post
If you can't quantify how much they cost, it's a deal, I'll buy 5 of them for 3 lumps of rocking horse ......
Reply
#14
(03-10-2019, 05:22 PM)completelycharged Wrote: If you have a single properly designed inverter with sufficient internal capacitance, forget about matching cable lengths.

Are you able to define what "sufficient internal capacitance" is ?

(03-10-2019, 05:22 PM)completelycharged Wrote: If you have multiple parallel inverters with a poor design and non isolated communication links between inverters then basic cable length imbalance can be a massive issue because it will crate a DC offset in the communication links, potentially ending up with the wrong level being sent between inverters resulting in oscillation or incorrect overall output level, as happens with some.

You are conflating non isolated comms circuits with DC input stage ripple.

If at least we can agree that Victron are proven competent designers and manufacturers, why is the following statement a requirement ?

"DC and AC wiring

Both the DC and AC wiring needs to be symmetrical per phase: use the same length, type and cross-section to every unit in the phase. To make this easy, use a bus-bar or power-post before and after the inverter/chargers"


https://www.victronenergy.com/live/ve.bu...se_systems

(03-10-2019, 05:22 PM)completelycharged Wrote: I fail to see, techincally,  how any small inline cable inductance for a properly designed inverter would cause any form of damage as the inductance on a relative basis is extremerly small, and probably not detectable with the average digital multimeter. The ripple caused by the battery resistance will normally be at least an order of magnitude larger.

That's correct, even £250  worth of Fluke wont visibly detect the spikes, but that doesnt mean they dont occur. A fairly cheap storage scope might change your opinion ....  someone on ES has modelled using what looks like Spice, clearly showing what will damage your input stage (in a badly designed system)

Here's a little more reading, if anyone is still interested 

https://endless-sphere.com/forums/viewto...=2&t=19873

(03-10-2019, 05:22 PM)completelycharged Wrote: Inductance, treat as zero, unless proven otherwise.

The ability to prove its existence depends entirely on you (and your test equipments) ability to measure it.

If folks are happy ignore its existence,  and the likely implications, then so be it.
Reply
#15
Sorry I don't do spoon feeding.
If you can't quantify how much they cost, it's a deal, I'll buy 5 of them for 3 lumps of rocking horse ......
Reply
#16
(03-10-2019, 07:15 PM)completelycharged Wrote: Sorry I don't do spoon feeding.

Nice to see you've taken me off your ignore list  Smile

I'm not expecting you to believe anything I say, if you opt to ignore manufacturers instructions, so be it.
Reply
#17
When misleading, false information is posted, that information needs correcting.

"You are conflating non isolated comms circuits with DC input stage ripple." - you obviously have no idea what your talking about.

"That's correct, even £250 worth of Fluke wont visibly detect the spikes," - now your are seemingly pointing toward the 1-5nS FET switching induced voltage spikes (at around 20kHz) that are actually filtered by inline LC filter in a normal inverter ?

"A fairly cheap storage scope might change your opinion" - thanks, I already have a not so cheap storaeg scope.

"I'm not expecting you to believe anything I say" I do actually, some of what you have input is actually quite good.

Read through this to get an idea on switching issues, harmonics and some behaviours of inverters (and the lack of damage likely from stray input cable inductance)

https://www.thebackshed.com/forum/forum_...=9409&PN=1

In there are the scope traces for various spikes, timings and solutions. A little bit more technically backed up than a web page for a boat builder.
If you can't quantify how much they cost, it's a deal, I'll buy 5 of them for 3 lumps of rocking horse ......
Reply
#18
Were the Victron instructions incorrect ?

Was the Xantrex data wrong ? (that's the boat builders link I posted)

Did Daniel Fahlgren working with Volvo get it wrong ?

Its matters not if you don't believe me .... 

(03-10-2019, 08:00 PM)completelycharged Wrote: Read through this to get an idea on switching issues, harmonics and some behaviours of inverters (and the lack of damage likely from stray input cable inductance)

https://www.thebackshed.com/forum/forum_...=9409&PN=1

In there are the scope traces for various spikes, timings and solutions. A little bit more technically backed up than a web page for a boat builder.

That "boat builder" is actually an equipment distributor, using Xantrex data, you'd agree they are competent ?

I read the first 10 pages, and the last 2 of that link - have they resolved what's causing FETS to self destruct ?
Reply
#19
A lot of the test that are being carried out are quite interesting to actually see the real timing duration of events. You need to read the whole lot if your interested as it will show switching spikes and filtering of them. Skip reading to the last 2 pages will tell you nothing. It's a bit heavy at times to really understand what is actually happening / being displayed in the data as the associated commentry can be wrong (prime example the case of overheatting FET test, where a 180 degree about face was performed after a few pages)

Yes, there are a few different reasons for failure though (and resolved), some solutions are not in that set of posts.

oztules inverter works fine, EGS002 units work fine, put the wrong components in, expect a random or spectacular failure.

I could be a distributor for X-ray machines, it does not make me a nuclear scientist. I can read and repeat anything given to me. Put the uH value in real world context, what does a 5uH line inductance on a wire to an inverter with input capacitors really mean ? Means nothing. Some manufacturers will quote whatever they need to sell things.

I agree the "boat builder" is competent at building boats.

Too many people parrot internet myths over and over until they seemingly become gospel. Take the 2800mm rule of thumb from the inverter builds, work the actual number out and that figure is based on a core flux density of 1.6 Tesla, which is above normal commercial designs (1.5 Tesla) and means your idle losses may be around 15% higher than what could be achieved (based on 0.85W/kg at 1.5T vs 1.0W/kg at 1.6T provided by a transformer manucafturer). But, hey it becomes internet gospel and re-posted about 150 times. It works so must be right, I just miss out on an extra saving of 15% by taking someone elses word for something I am too lazy to work out myself....

Helps a bit having read through the previous few hundered pages of various post threads starting back in from the original post in 2011.
If you can't quantify how much they cost, it's a deal, I'll buy 5 of them for 3 lumps of rocking horse ......
Reply
#20
(03-10-2019, 09:57 PM)completelycharged Wrote: Too many people parrot internet myths over and over until they seemingly become gospel. 

I'm providing you links to Inverter manufacturers instructions ....

That "boat builder" is using data from the Xantrex/Trace/Scheider SW series installation manual (page 120 ish) - link. (that's 3 times I've reminded you of that)

The Victron instructions echo what I've previously posted, we are going around in circles.

Perhaps you know better than the Manufacturers .....
Reply


Who read this thread?
66 User(s) read this thread:
daromer, Sean, mike, opus40, Korishan, PAF, Mike C, wim, watts-on, Dan Lim, Stefaan De Ridder, Chiptosser, Geek, SWEwall, Jon, CarelHassink, AndyMc1970, Charly144, djuro, sunnyeveryday, Crimp Daddy, Redpacket, completelycharged, Ibiza, Oliman, arhunter, choncy12, Bubba, Walde, Riplash, 100fuegos, bukvaahmed@gmail.com, chuckp, emuland-metroman, Truth, Leipi35, Wolf, Bastler_Ro, camthecam, Dallski, Speicher, ajw22, TimBelgium, kevinjaye, OffGridInTheCity, bchur83, fannel, Richard, Dragon, Church1182, beserker786, winny, Havskum, strat24v, Ebarnett, allend83, 100kwh-hunter, Bilabob, JimU, energy_man, zoneblue, loyd, tachion, sgrevskott, Werner2, offgrid2020

Forum Jump:


Users browsing this thread: 1 Guest(s)