Lithium Batteries installation

[nicknorman]

8 hours ago, IanD said:

 

You might want to bone up on Ohm's Law. Unless the voltage difference is huge -- which it won't be in the case you mentioned -- you won't get large currents flowing down thin wires because of their resistance.

 

😉

And there was me thinking you were some sort of design engineer! I did explain the consequences of the differing cable resistances. I also explained the potential for a bad connection to result in excessive current flowing in the thin wires / pcb tracks. What I didn’t explain was that the BMS measures cell voltages with a resolution of 1mV and an accuracy of a couple of mV and even a small current flowing through the BMS 0v wire to the cells is going to introduce a significant error. It would simply be bad design to create an alternative current path.

 

Actually I did think of connecting only the Tx and Rx VE.Direct wires but not the 0v, but decided this was a bit gash so used optoisolators instead.

 

[IanD]

1 hour ago, nicknorman said:

And there was me thinking you were some sort of design engineer! I did explain the consequences of the differing cable resistances. I also explained the potential for a bad connection to result in excessive current flowing in the thin wires / pcb tracks. What I didn’t explain was that the BMS measures cell voltages with a resolution of 1mV and an accuracy of a couple of mV and even a small current flowing through the BMS 0v wire to the cells is going to introduce a significant error. It would simply be bad design to create an alternative current path.

 

Actually I did think of connecting only the Tx and Rx VE.Direct wires but not the 0v, but decided this was a bit gash so used optoisolators instead.

 

Why don't you leave off the ad hominem attacks and stick to the facts and good old Ohm's and Kirchoff's laws?

 

Diversion tactics here -- you weren't talking about small errors in the BMS voltages which can obviously happen with ground voltage drops if the BMS is remote, you were talking about large currents flowing in thin ground wires like CANbus and damaging/melting them -- which simply isn't going to happen with the voltage differences here, regardless of how large the currents in the thick wires are.

 

If you get a high resistance ground connection in a high current path -- causing a ground voltage difference -- then I^2R heating due to voltage drop will damage the connection/cable in that ground path, not any other thin non-isolated ground wires between the two grounds.

 

Draw out an equivalent circuit and put some realistic numbers in and this becomes obvious... 😉

Edited February 12 by IanD

[magnetman]

This was an interesting comment on another forum in a thread about using LTO batteries as starter for a vehicle. 

 

Obviously the internet is like liquorice and there are ultracrepidarians all over the place but the M6 rod thing did pique my interest and raise an eyebrow. 

M6 is a thread not a diameter.

So presumably the suggestion is to somehow connect a length of ss M6 studding in the charge line from the alternator. 

 

would this really do anything ? 

If it might work then one could use ring crimps with 6mm holes and nyloc or binx nuts to clamp the wires to each end. 

 

Extending the idea a little more one could have a 1m length of the said studding mounted onto Stauff clamps (nylon hydraulic hose clamps) and experiment with clamping one of the wires at different positions. 

 

A variable resistor? 

[PeterF]

13 hours ago, magnetman said:

 

So why does the Victron 30A B2B have an isolated option but the 50A version doesn't? 

 

It can't be an oversight because the 30A version is available with or without. 

 

If anything one could think that the higher current device might want it more than the lower current device. 

 

Is it something which adds to the cost significantly ? 

 

 

The Orion XS DC to DC charger is a new product line using better electronics than the existing Orions and will be available in March/April in the UK. The 50A 12V to 12V non-isolated version is the first to be released. This will be followed by different voltage / current / step up / step down and isolated versions. There is also talk that the hardware is bi-directional but the firmware will not be ready for this. They will also be controllable by DVCC from the Cerbo taking the BMS directions, firmware for this available March.

Datasheet-Orion-XS-12-12-50A-DC-DC-battery-charger-EN.pdf

[nicknorman]

1 hour ago, IanD said:

Why don't you leave off the ad hominem attacks and stick to the facts and good old Ohm's and Kirchoff's laws?

 

Diversion tactics here -- you weren't talking about small errors in the BMS voltages which can obviously happen with ground voltage drops if the BMS is remote, you were talking about large currents flowing in thin ground wires like CANbus and damaging/melting them -- which simply isn't going to happen with the voltage differences here, regardless of how large the currents in the thick wires are.

 

If you get a high resistance ground connection in a high current path then heating due to voltage drop will damage the connection/cable in that ground path, not any other thin non-isolated ground wires.

 

Draw out an equivalent circuit and put some realistic numbers in and this becomes obvious... 😉

You started the ad hominem attack, but of course when you do it, it doesn’t count. In your head.

As I said at the outset, under normal circumstances the stray found current would not be problematic for heating/damage/fire risk. However in the event of a poor or bad connection it could be significant. And of course you would only be one human error away from disaster - someone decides to disconnect the battery negative and then you are left with the thin wires and PCB tracks to pass the entire system current.

 

It is fundamentally bad design to create that sort of alternative current path gotcha. Same applies to alternative current paths created eg by a radio aerial earth connected to hull. It is probably not going to be an issue due to the differing current path resistances, but it could be under unusual circumstances.

 

And while I think of it, I had similar thoughts on the CANBUS interface. In the end I had 2 pads on the PCB, CANBUS ground and system ground. Because CANBUS is a differential system I decided that most likely it would not be necessary to connect the 2 grounds together and this has shown to be correct, so another alternative current path was eliminated. A good design engineer thinks about these things and tries to eliminate these sort of issues that lie only just below the surface of a correctly functioning and correctly operated system.

 

 

[IanD]

11 minutes ago, PeterF said:

The Orion XS DC to DC charger is a new product line using better electronics than the existing Orions and will be available in March/April in the UK. The 50A 12V to 12V non-isolated version is the first to be released. This will be followed by different voltage / current / step up / step down and isolated versions. There is also talk that the hardware is bi-directional but the firmware will not be ready for this. They will also be controllable by DVCC from the Cerbo taking the BMS directions, firmware for this available March.

Datasheet-Orion-XS-12-12-50A-DC-DC-battery-charger-EN.pdf 650.87 kB · 0 downloads

Thanks for the detailed explanation 🙂

 

Given how non-isolated buck-boost converters work (same nominal voltage on both sides -- 12V in this case), they are inherently capable of power transfer in both directions, and in fact the control software has to take care to stop reverse power transfer if that's not desirable -- I should know, I've designed one for on-chip use and had to stop this happening... 😉

 

But there are a lot more settings and options needed for bidirectional use to deal with all the possible use cases -- including dumb users programming in settings that could cause damage to their system! -- so the firmware is more complex and takes longer to debug, I'm not surprised that the initial release will be unidirectional only... 🙂

[nicknorman]

30 minutes ago, magnetman said:

This was an interesting comment on another forum in a thread about using LTO batteries as starter for a vehicle. 

 

Obviously the internet is like liquorice and there are ultracrepidarians all over the place but the M6 rod thing did pique my interest and raise an eyebrow. 

M6 is a thread not a diameter.

So presumably the suggestion is to somehow connect a length of ss M6 studding in the charge line from the alternator. 

 

would this really do anything ? 

If it might work then one could use ring crimps with 6mm holes and nyloc or binx nuts to clamp the wires to each end. 

 

Extending the idea a little more one could have a 1m length of the said studding mounted onto Stauff clamps (nylon hydraulic hose clamps) and experiment with clamping one of the wires at different positions. 

 

A variable resistor? 

You have to consider the heat dissipated in the “resistor”. A short length of rod with resistance 1 ohm and 100A is going to dissipate 10kw! Apart from anything else, 1 ohm is far too much as it would drop 100v. More realistic would be 0.01ohm which would drop 1v at 100A. But that is still going to dissipate 100w and will get extremely hot. This is why a long wire is the preferred choice - the longer the wire the more scope for dissipating power without getting too hot.

[IanD]

9 minutes ago, nicknorman said:

You started the ad hominem attack, but of course when you do it, it doesn’t count. In your head.

As I said at the outset, under normal circumstances the stray found current would not be problematic for heating/damage/fire risk. However in the event of a poor or bad connection it could be significant. And of course you would only be one human error away from disaster - someone decides to disconnect the battery negative and then you are left with the thin wires and PCB tracks to pass the entire system current.

 

It is fundamentally bad design to create that sort of alternative current path gotcha. Same applies to alternative current paths created eg by a radio aerial earth connected to hull. It is probably not going to be an issue due to the differing current path resistances, but it could be under unusual circumstances.

 

And while I think of it, I had similar thoughts on the CANBUS interface. In the end I had 2 pads on the PCB, CANBUS ground and system ground. Because CANBUS is a differential system I decided that most likely it would not be necessary to connect the 2 grounds together and this has shown to be correct, so another alternative current path was eliminated. A good design engineer thinks about these things and tries to eliminate these sort of issues that lie only just below the surface of a correctly functioning and correctly operated system.

 

 

 

Many CANbus systems are isolated, including some in Victron gear, to cope with large ground voltage differences in large systems. Many comms systems including CANbus are differential, but with transmitters at both ends you can't follow the usual "ground cable shield at TX, isolate at RX" recommendation because there are TX at both ends, so the cable shield has to connect to GND at both ends -- a completely isolated interface usually has a floating GND for the transceiver and internally uses optical/transformer isolation to get power and data to and from the equipment.

 

If you want to cope with things like disconnecting a battery negative then you need rather more than a DC-DC converter with isolated grounds, you need to design the entire system to cope with this -- and as you say there's nothing wrong with doing that if it's feasible. Which it often isn't... 😞

 

None of which is anything to do with the question "does a B2B charger/DC-DC converter on a boat need isolated grounds?" -- to which the answer is, no it doesn't.

Edited February 12 by IanD

[nicknorman]

17 minutes ago, IanD said:

 

Many CANbus systems are isolated, including some in Victron gear, to cope with large ground voltage differences in large systems. Many comms systems including CANbus are differential, but with transmitters at both ends you can't follow the usual "ground cable shield at TX, isolate at RX" recommendation because there are TX at both ends, so the cable shield has to connect to GND at both ends -- a completely isolated interface usually has a floating GND for the transceiver and internally uses optical/transformer isolation to get power and data to and from the equipment.

 

 

 

With CANBUS being a differential system, generally there is no need for cable screening. Certainly not for boat systems. At a pinch you can use a twisted pair but it's not necessary at the 250kB/s rate that Masterbus uses. In fact thinking about it, probably the only reason for Mastervolt including a ground in their connectors/cabling is because some of their devices need to receive power via the cable to run the CANBUS interface. But if I had been careless of thought I might have connected the gnd on the CANBUS system to the gnd on the BMS and Alternator controllers and that would have been a mistake.

 

21 minutes ago, IanD said:

 

None of which is anything to do with the question "does a B2B charger/DC-DC converter on a boat need isolated grounds?" -- to which the answer is, no it doesn't.

 

My point was that was too blanket a statement. Often no, but sometimes yes especially if there are systems at different supply voltages (eg 12v, 24v, 48v systems) which does sometimes happen on boats.

[magnetman]

46 minutes ago, nicknorman said:

You have to consider the heat dissipated in the “resistor”. A short length of rod with resistance 1 ohm and 100A is going to dissipate 10kw! Apart from anything else, 1 ohm is far too much as it would drop 100v. More realistic would be 0.01ohm which would drop 1v at 100A. But that is still going to dissipate 100w and will get extremely hot. This is why a long wire is the preferred choice - the longer the wire the more scope for dissipating power without getting too hot.

Threaded rod seems interesting as it is sort of formed like a natural heatsink. 

Maybe a rod through a cylinder containing water. 

[nicknorman]

8 minutes ago, magnetman said:

Threaded rod seems interesting as it is sort of formed like a natural heatsink. 

Maybe a rod through a cylinder containing water. 

What happens when the water has all steamed off?

[magnetman]

4 minutes ago, nicknorman said:

What happens when the water has all steamed off?

A tiny circulation pump and take it to a little calorifier for hot water to wash hands. 

[IanD]

29 minutes ago, nicknorman said:

 

With CANBUS being a differential system, generally there is no need for cable screening. Certainly not for boat systems. At a pinch you can use a twisted pair but it's not necessary at the 250kB/s rate that Masterbus uses. In fact thinking about it, probably the only reason for Mastervolt including a ground in their connectors/cabling is because some of their devices need to receive power via the cable to run the CANBUS interface. But if I had been careless of thought I might have connected the gnd on the CANBUS system to the gnd on the BMS and Alternator controllers and that would have been a mistake.

 

 

My point was that was too blanket a statement. Often no, but sometimes yes especially if there are systems at different supply voltages (eg 12v, 24v, 48v systems) which does sometimes happen on boats.

 

CANbus is differential but even so it's often screened on boats, because people will insist on doing things like running the cables right past things like SMPS that radiate a lot of switching noise. Yes I know that people like you and I wouldn't dream of doing this, in the same way we wouldn't run microphone/instrument cables on a stage right next to the mains cables, but that doesn't stop people doing this... 😞

 

The need for an isolated GND on B2B is still only true if the low rail is not 0V on the higher-voltage systems, which is not common -- for example my boat has 12V and 24V and 48V systems and all are ground-referred to 0V. If one GND is not the same voltage as the other then obviously you do need an isolated B2B, but I doubt that many boats are like this because it's asking for trouble with wiring -- because people think GND means GND and there's a temptation to connect them together... 😉

[nicknorman]

15 minutes ago, IanD said:

 

CANbus is differential but even so it's often screened on boats, because people will insist on doing things like running the cables right past things like SMPS that radiate a lot of switching noise. Yes I know that people like you and I wouldn't dream of doing this, in the same way we wouldn't run microphone/instrument cables on a stage right next to the mains cables, but that doesn't stop people doing this... 😞

 

The need for an isolated GND on B2B is still only true if the low rail is not 0V on the higher-voltage systems, which is not common -- for example my boat has 12V and 24V and 48V systems and all are ground-referred to 0V. If one GND is not the same voltage as the other then obviously you do need an isolated B2B, but I doubt that many boats are like this because it's asking for trouble with wiring -- because people think GND means GND and there's a temptation to connect them together... 😉

Mastervolt's own Masterbus/CANBUS cables are not screened. So I don't know where you are getting your "often screened on boats" from. How many boats with CANBUS are you aware of that use screened cable? Please cite them. And I don't mean eg naval warships! 120 ohm impedance and voltage differentials of several volts between hi and lo means that the chances of some SMPS radiating interferece that affects the data is effectively zero. More so (better than zero😁) if you use a twisted pair.

 

I am personally aware of 2 boats that use an isolated B2B, in both cases they are hybrids with 48v LA drive batteries and 24v engine electrics along with 2 B2Bs for charging the 48v bank from the engine alternator/battery. The upper one at least must be isolated.
 

[IanD]

1 minute ago, nicknorman said:

Mastervolt's own Masterbus/CANBUS cables are not screened. So I don't know where you are getting your "often screened on boats" from. How many boats with CANBUS are you aware of that use screened cable? Please cite them. And I don't mean eg naval warships! 120 ohm impedance and voltage differentials of several volts between hi and lo means that the chances of some SMPS radiating interferece that affects the data is effectively zero. More so (better than zero😁) if you use a twisted pair.

 

I am personally aware of 2 boats that use an isolated B2B, in both cases they are hybrids with 48v LA drive batteries and 24v engine electrics along with 2 B2Bs for charging the 48v bank from the engine alternator/battery. The upper one at least must be isolated.
 

 

The longer comms cables on my boat are screened, for example from the Vetus Pro BT controller at the stern to the BT. Victron also recommend this for long cable runs in noisy environments, which is exactly what I said. The shorter ones are unscreened.

 

If you want to charge a battery from a system with a different DC voltage on the GND terminal then you need an isolated B2B -- the case you quoted which has 2 B2B with parallel inputs (0-24V) and series outputs (0-24V and 24V-48V) needs one for the upper B2B, just like I said. Though this does raise potential balancing issues with the centre tap of the 48V battery, which I assume is connected to the series centre tap of the B2B outputs.

 

The fact remains that this is an unusual case for a boat, and in the normal case under discussion -- 12V in to 12V out -- there's no need for isolation, or indeed any point in it.

 

The same applies for conversion between different voltages where the low rail is a common 0V GND in all cases, which again is the most common case -- for example my boat has 12V for the generator start battery, 24V for the boat lighting etc, 48V for propulsion, all referred to 0V, and with non-isolated DC-DC converters for 24V (48V input, no 24V battery) and to trickle-charge the genny battery.

[magnetman]

LTO. 

 

Brilliant ! 

 

Cuts out the problems 

 

I wish I had bought all of them except I have no use for more than about 10. 

Its great being a rebel with no BMS or balancer. 

 

Live life on the edge 

 

I don't even have a can bus ! 

 

Or a 0v reference. 

 

 

[nicknorman]

Maybe you should get some sleep, it must be incredibly tiring to always be right about everything.

[IanD]

1 minute ago, nicknorman said:

Maybe you should get some sleep, it must be incredibly tiring to always be right about everything.

Maybe as well as leaving out the snide digs, you should be wrong less often then... 😉

Edited February 12 by IanD

[nicknorman]

Cite a narrowboat or other small boat that uses screened canbus cables?:

"My boat. Therefore that must mean that it is often used on boats."

<Vetus installation manual doesn't mention any need for screened cable>

Examples given of boats needing isolate gnd B2S:

"Yes but my boat doesn't do that so clearly it doesn't need to exist elsewhere"

It really is all about you!

[magnetman]

55 minutes ago, nicknorman said:

Maybe you should get some sleep, it must be incredibly tiring to always be right about everything.

Being right just comes naturally to me.

 

Its no bother at all. 

[Iain_S]

2 hours ago, nicknorman said:

You have to consider the heat dissipated in the “resistor”. A short length of rod with resistance 1 ohm and 100A is going to dissipate 10kw! Apart from anything else, 1 ohm is far too much as it would drop 100v. More realistic would be 0.01ohm which would drop 1v at 100A. But that is still going to dissipate 100w and will get extremely hot. This is why a long wire is the preferred choice - the longer the wire the more scope for dissipating power without getting too hot.

Would 100A through 1 ohm not need 100V?

[IanD]

37 minutes ago, nicknorman said:

Cite a narrowboat or other small boat that uses screened canbus cables?:

"My boat. Therefore that must mean that it is often used on boats."

<Vetus installation manual doesn't mention any need for screened cable>

Examples given of boats needing isolate gnd B2S:

"Yes but my boat doesn't do that so clearly it doesn't need to exist elsewhere"

It really is all about you!

 

I really don't understand why you've turned a simple question about whether a B2B -- used as being discussed for the Victron Orion XS, 12V LA to 12V LFP -- needs isolation into an argument about CANbus screening, throwing out assorted strawman and diversionary arguments in the process, some of them just plain wrong -- or at least, wrong as you stated them before backpedalling... 😉

 

To repeat the answer to the original question -- no, there's no advantage to having an isolated B2B except in rare and unusual circumstances, 12V LA to 12V LFP isn't one of these, isolation costs more and increases losses to no benefit.

[nicknorman]

38 minutes ago, IanD said:

 

I really don't understand why you've turned a simple question about whether a B2B -- used as being discussed for the Victron Orion XS, 12V LA to 12V LFP -- needs isolation into an argument about CANbus screening, throwing out assorted strawman and diversionary arguments in the process, some of them just plain wrong -- or at least, wrong as you stated them before backpedalling... 😉

 

To repeat the answer to the original question -- no, there's no advantage to having an isolated B2B except in rare and unusual circumstances, 12V LA to 12V LFP isn't one of these, isolation costs more and increases losses to no benefit.

As I explained several times, you made a blanket statement that was flawed. I was merely pointing that out. But I realise my mistake now, it is against the laws of the universe for you to have said anything even slightly wrong or over-generalised. Sorry about that.

45 minutes ago, Iain_S said:

Would 100A through 1 ohm not need 100V?

Yes, as I said in the next sentence!

[IanD]

2 hours ago, nicknorman said:

As I explained several times, you made a blanket statement that was flawed. I was merely pointing that out. But I realise my mistake now, it is against the laws of the universe for you to have said anything even slightly wrong or over-generalised. Sorry about that.

Yes, as I said in the next sentence!

 

Given your fundamentally wrong statement about a bad high-current connection resulting in smoky melted low-current wires -- because the numbers simply don't add up, the thin wire current will be small, probably less than an amp with a 1V drop on the high-current connection which *will* melt *that* -- you might want to think twice before accusing me of over-generalising.

 

BTW, you still haven't pointed out exactly what I said that you claim was wrong, so please feel free to do so -- and if it was wrong, I'll happily admit it 🙂

[nicknorman]

Bored with pointless arguing with someone who is not capable of ever being even slightly wrong.