Victron - The only choice.

[IanD]

1 hour ago, dmr said:

It's the DC screw terminals that concern me. I am thinking about a Victron DC-DC thingy as a better way combining the starter and domestic alternators but putting 30 amps through a screw terminal mounted on a PCB does feel like the upper end of sensibility.

The problem is not the current as such, it's perfectly possible to get much higher currents than that from a terminal into a PCB e.g. in the BMS inside LFP batteries.

 

The problem is either strain relief to stop stress on the stiff cable or during screw (over-) tightening damaging the PCB or joint, or not using enough torque to really tighten the terminal down properly -- one way or the other (too much torque or too little, or poor strain relief) it's often user error that is to blame.

[dmr]

What is the figure for the safe current on a PCB?.

The current handling of those BMS PCB's does worry me. It always feels odd that people are installing 70mm cables and m8, or even m10 connecting bolts, but are happy to see that same current running through a PCB track.

Its many years since I have done any "significant current" PCB design but at 30 amps I would have used double thickness copper and run thick solder on the top, but then I was building one-offs that needed high reliability.

Copper bus bars are specified at a bit more than 1 amp per mm2 of CSA, so a PCB track is well beyond this.

 

Ive just spent the last few days re-doing the high current part of our 12 volt electrical system and have a half melted shunt right in front of me now so I am hyper aware of what too many amps can do 😀

[rusty69]

I wonder how critical cable selection choice, both size and type and the use of bootlace ferrules affects the clamping ability. 

 

I wasn't aware there was a torque setting, perhaps I should read the manual closer. If that is the case I can see a torque screwdriver on this year's Christmas list. 

Come to think of it, I have a small torque wrench I bought for the lithium terminal bolts. What's to stop one sticking a screwdriver bit in it?! 

[IanD]

12 hours ago, dmr said:

What is the figure for the safe current on a PCB?.

The current handling of those BMS PCB's does worry me. It always feels odd that people are installing 70mm cables and m8, or even m10 connecting bolts, but are happy to see that same current running through a PCB track.

Its many years since I have done any "significant current" PCB design but at 30 amps I would have used double thickness copper and run thick solder on the top, but then I was building one-offs that needed high reliability.

Copper bus bars are specified at a bit more than 1 amp per mm2 of CSA, so a PCB track is well beyond this.

 

Ive just spent the last few days re-doing the high current part of our 12 volt electrical system and have a half melted shunt right in front of me now so I am hyper aware of what too many amps can do 😀

 

The current density for PCB traces can be considerably higher than for cables or busbars, because the things which limit the current are different. For thick cables/busbars you can't use high current densities because they get hot (heat escaping from inside -- or perimeter/area ratio) and the voltage drop becomes too high. PCB traces are better at getting rid of heat (thin flat copper) and can run hotter with bigger voltage drop per unit length because they're usually much shorter.

 

https://www.protoexpress.com/blog/trace-current-capacity-pcb-design/

 

When getting current into a PCB from a terminal, if it's bolted to the PCB with washers on both sides (or terminal on top and washer underneath) -- recommended anyway for strength under cable stress -- or soldered down (preferably both sides) then the current spreads out from this, there will be a local hot-spot but the current density drops rapidly as it spreads out. Then the heat and voltage drop getting across the PCB to the circuits (e.g. BMS FETS) is normally less of a problem as the current spreads out over a much bigger area.

 

2oz copper (standard for high-current PCBs) has a resistance of about 0.25milliohms per square, or half this (0.12mohms/square) if copper on both top and bottom is used (common practice). If you take the washer diameter as 15mm then the resistance for the current spreading out into the PCB (copper on both sides) will be something like 0.1milliohm, which will give a voltage drop of 1mV/A, or 10mV at 100A. This would also dissipate about 1W which is no problem. If you can put up with bigger voltage drop or more heating (or a bigger diameter terminal/washer) then even higher currents are OK, but 100A per terminal is a safe figure for a power PCB.

 

What probably matters more than worrying about the current capability of the PCB -- assuming it was properly designed in the first place! -- is the terminal and how it's attached to get reliably low resistance, and how effective any cable clamp/ferrule is at getting and keeping a low resistance connection over time. As soon as there's any oxidation or bad contact the resistance goes up rapidly, which makes it get hotter, which increases the resistance and oxidation, which makes it get hotter still... 😞

 

The problem is similar to what we see with high-power chips drawing hundreds of amps -- and at low voltage, so any drop is *really* important -- where we've seen thermal runaway like this in the past, ending up with an expensive chip melting an even more expensive socket into a *really* expensive PCB just because of a poor contact.

Edited January 16 by IanD

[magnetman]

16 hours ago, nicknorman said:

Is there some cross purpose discussion conflating the DC connectors with the mains AC connectors?

The discussion is about the screw terminal blocks on Victron products such as their solar controllers.

 

Its not an AC or DC question its just a question of doing up the screws on the connectors.

[IanD]

1 minute ago, magnetman said:

The discussion is about the screw terminal blocks on Victron products such as their solar controllers.

 

Its not an AC or DC question its just a question of doing up the screws on the connectors.

The current is much higher on the DC connectors so it's more of a problem for them, but then the AC connectors are smaller and can still have a problem if not clean and tightened correctly.

 

Many years ago we had a massive tester which filled most of a room and consumed something like 30kW, and being sourced from the USA it ran from 110V (or the 3-phase equivalent) so we had to connect it via a huge stepdown transformer about the size of a washing machine. One day the smoke alarm in the tester room -- closed off due to the deafening fans and aircon -- went off, and we found that one of the mains connections on the transformer had got so hot it had melted the cable insulation next to the connector and then set it on fire... 😞

[Js north]

I would only ever recommend victron 

[dmr]

On 16/01/2024 at 10:57, IanD said:

 

The current density for PCB traces can be considerably higher than for cables or busbars, because the things which limit the current are different. For thick cables/busbars you can't use high current densities because they get hot (heat escaping from inside -- or perimeter/area ratio) and the voltage drop becomes too high. PCB traces are better at getting rid of heat (thin flat copper) and can run hotter with bigger voltage drop per unit length because they're usually much shorter.

 

https://www.protoexpress.com/blog/trace-current-capacity-pcb-design/

 

When getting current into a PCB from a terminal, if it's bolted to the PCB with washers on both sides (or terminal on top and washer underneath) -- recommended anyway for strength under cable stress -- or soldered down (preferably both sides) then the current spreads out from this, there will be a local hot-spot but the current density drops rapidly as it spreads out. Then the heat and voltage drop getting across the PCB to the circuits (e.g. BMS FETS) is normally less of a problem as the current spreads out over a much bigger area.

 

2oz copper (standard for high-current PCBs) has a resistance of about 0.25milliohms per square, or half this (0.12mohms/square) if copper on both top and bottom is used (common practice). If you take the washer diameter as 15mm then the resistance for the current spreading out into the PCB (copper on both sides) will be something like 0.1milliohm, which will give a voltage drop of 1mV/A, or 10mV at 100A. This would also dissipate about 1W which is no problem. If you can put up with bigger voltage drop or more heating (or a bigger diameter terminal/washer) then even higher currents are OK, but 100A per terminal is a safe figure for a power PCB.

 

What probably matters more than worrying about the current capability of the PCB -- assuming it was properly designed in the first place! -- is the terminal and how it's attached to get reliably low resistance, and how effective any cable clamp/ferrule is at getting and keeping a low resistance connection over time. As soon as there's any oxidation or bad contact the resistance goes up rapidly, which makes it get hotter, which increases the resistance and oxidation, which makes it get hotter still... 😞

 

The problem is similar to what we see with high-power chips drawing hundreds of amps -- and at low voltage, so any drop is *really* important -- where we've seen thermal runaway like this in the past, ending up with an expensive chip melting an even more expensive socket into a *really* expensive PCB just because of a poor contact.

 

Interesting stuff, thanks, busy for a couple of days but will respond when I have read that link in more detail (and got my head round all those imperial units 😀)

[IanD]

Just now, dmr said:

 

Interesting stuff, thanks, busy for a couple of days but will respond when I have read that link in more detail (and got my head round all those imperial units 😀)

 

Blame the Yanks, they insist on using "mils" (1/1000") everywhere, which causes huge fun when everyone else uses mm or microns. Not just PCBs, it causes havoc with component sizes too -- an 0603 SMT capacitor in the US is 1608 elsewhere, and an 0603 elsewhere is 0201 in the USA. Guess what happens when you order a reel of 10000 0603 100nF capacitors... 😞

 

https://olimex.wordpress.com/2016/02/15/smt-component-sizes-and-metric-imperial-confusion/

[dmr]

1 hour ago, IanD said:

 

Blame the Yanks, they insist on using "mils" (1/1000") everywhere, which causes huge fun when everyone else uses mm or microns. Not just PCBs, it causes havoc with component sizes too -- an 0603 SMT capacitor in the US is 1608 elsewhere, and an 0603 elsewhere is 0201 in the USA. Guess what happens when you order a reel of 10000 0603 100nF capacitors... 😞

 

https://olimex.wordpress.com/2016/02/15/smt-component-sizes-and-metric-imperial-confusion/

 

I confess I had to look up mils, its just another name for a thou 😀 I assume its milli-inches ????? so a strange combination of metric and imperial.

We could also have km as kilomiles to really cause confusion.

[Alan de Enfield]

1 minute ago, dmr said:

 

I confess I had to look up mils, its just another name for a thou 😀 I assume its milli-inches ????? so a strange combination of metric and imperial.

We could also have km as kilomiles to really cause confusion.

 

But then of course we have the 'military' Mil which is used by pretty much all of the major armies.

 

 

Mils:
Another unit of measure, the radian, is used mainly by militaries in artillery, tank, and mortar gunnery.
There are 2 PI radians in a circle. PI is a constant of approximately 3.1416. That is 2 * 3.1416, or 6.283 radians. Divide each radian into 1000 mil-radians and you see there are 6283 mil-radians in a circle. Mil-radians are called mils for short.
17.78 mils equal 1 degree.

Compass use of mils typically rounds 6283 to 6400 for simplification. Some foreign militaries have simplified the other direction and divided the compass face into 6000 units, exactly like the face of a watch, with 100 units the same angle as a minute on the watch face.

Using mils, the actual size of an object observed in the field can be estimated. An object that appears to be n mils wide when it is 1000 units away from you, is actually n units wide - the units used does not matter, feet, yards, meters, miles.

A train that appears to be 300 mils long and is 1000 feet distant is actually 300 feet long. Or, two vehicles that appear to be 200 mils apart and are 1000 meters away, are actually 200 meters apart.

 

My military sniper compass in Mils :

 

 

[dmr]

12 pence in a shilling was bad.

13 in a bakers dozen was really bad.

but 6 and a bit radians in a circle is just daft 😀

[Alan de Enfield]

1 minute ago, dmr said:

but 6 and a bit radians in a circle is just daft

 

Blame the foreigners - a radian is an SI unit

It is defined such that one radian is the angle subtended at the centre of a circle by an arc that is equal in length to the radius.

 

It is a 'standard' unit which means if an Angle does not show explicitly specified units they are generally assumed to be measured in radians, especially in mathematical writing.

[dmr]

2 minutes ago, Alan de Enfield said:

 

Blame the foreigners - a radian is an SI unit

It is defined such that one radian is the angle subtended at the centre of a circle by an arc that is equal in length to the radius.

 

It is a 'standard' unit which means if an Angle does not show explicitly specified units they are generally assumed to be measured in radians, especially in mathematical writing.

 

As an engineer I am totally familiar with radians but when I first met them I did find the concept of "6 and a bit" a little odd, but then thats pi for you.

It could be worse, I remember trying to explain the concept of imaginary numbers to a non mathematical friend and I just got laughrd at.

 

Actually radians make a lot more sense than 360 degrees in a circle 😀.

[BoatinglifeupNorth]

2 hours ago, Alan de Enfield said:

 

But then of course we have the 'military' Mil which is used by pretty much all of the major armies.

 

 

Mils:
Another unit of measure, the radian, is used mainly by militaries in artillery, tank, and mortar gunnery.
There are 2 PI radians in a circle. PI is a constant of approximately 3.1416. That is 2 * 3.1416, or 6.283 radians. Divide each radian into 1000 mil-radians and you see there are 6283 mil-radians in a circle. Mil-radians are called mils for short.
17.78 mils equal 1 degree.

Compass use of mils typically rounds 6283 to 6400 for simplification. Some foreign militaries have simplified the other direction and divided the compass face into 6000 units, exactly like the face of a watch, with 100 units the same angle as a minute on the watch face.

Using mils, the actual size of an object observed in the field can be estimated. An object that appears to be n mils wide when it is 1000 units away from you, is actually n units wide - the units used does not matter, feet, yards, meters, miles.

A train that appears to be 300 mils long and is 1000 feet distant is actually 300 feet long. Or, two vehicles that appear to be 200 mils apart and are 1000 meters away, are actually 200 meters apart.

 

My military sniper compass in Mils :

 

 

You were seen off there if you paid extra for a “Sniper” compass, just a standard military prismatic compass.

Edited January 17 by BoatinglifeupNorth

[MtB]

3 hours ago, dmr said:

Actually radians make a lot more sense than 360 degrees in a circle 😀.

 

ISTR when I was at skool there was a body of opinion holding that we should change to having 400 degrees in a circle. 

 

 

A bit daft as I bet 360 is divisible by more integers than 400. Just a guess though. 

[dmr]

7 minutes ago, MtB said:

 

ISTR when I was at skool there was a body of opinion holding that we should change to having 400 degrees in a circle. 

 

 

A bit daft as I bet 360 is divisible by more integers than 400. Just a guess though. 

I reckon get rid of those degrees alltogether cus they just get confused with temperature, and instead have the simple concept of a revolution, then use SI prefixes so we would normally work in millirevolutions  😀

 

Then again I suppose that could get confused with a very small uprising.

[MtB]

2 minutes ago, dmr said:

I reckon get rid of those degrees alltogether cus they just get confused with temperature, and instead have the simple concept of a revolution, then use SI prefixes so we would normally work in millirevolutions  😀

 

Then again I suppose that could get confused with a very small uprising.

 

Anyone caught planning a revolution should be given the third degree. 

 

 

 

[Alan de Enfield]

1 hour ago, BoatinglifeupNorth said:

You were seen off there if you paid extra for a “Sniper” compass, just a standard military prismatic compass.

 

No, the standard compass is 'degrees', the sniper used the 'mils' to establish distances. They were normally issued to the Artillery but snipers were the elite and could request pretty much anything they wanted,

 

I have the complete issued L42a (Falklands war) issue Sniper rifle and full CES (Complete Equipment Schedule = all the items issued with the rifle, watch, compass, binoculars etc)

 

I also have a WW2 No4T sniper rifle and CES, this was issued with a 'degree' compass.

 

Both of these are 'live'.

Edited January 17 by Alan de Enfield

[Slim]

On 28/12/2023 at 15:33, truckcab79 said:

I was talking more generally. Not many ‘doers’ on here, but you could well be right. 
All the best. 

Define 'doers'

[BoatinglifeupNorth]

54 minutes ago, Alan de Enfield said:

 

No, the standard compass is 'degrees', the sniper used the 'mils' to establish distances. They were normally issued to the Artillery but snipers were the elite and could request pretty much anything they wanted,

 

I have the complete issued L42a (Falklands war) issue Sniper rifle and full CES (Complete Equipment Schedule = all the items issued with the rifle, watch, compass, binoculars etc)

 

I also have a WW2 No4T sniper rifle and CES, this was issued with a 'degree' compass.

 

Both of these are 'live'.

I imagine the watch is a Rolex and the binos special issue Leitz also😂😂 nothing special about that compass standard military issue.

NSN something like W3W6? 6605-99-???-8805???? Off the top of my head.

Edited January 17 by BoatinglifeupNorth

[Alan de Enfield]

9 minutes ago, BoatinglifeupNorth said:

I imagine the watch is a Rolex and the binos special issue Leitz also😂😂 nothing special about that compass standard military issue.

NSN something like W3W6? 6605-99-???-8805???? Off the top of my head.

 

Those ???   ???? could be the difference between a tank and a pencil

 

I said it was standard military issue - for the Artillery, The infantry had 'degree' compasses.

 

See for yourself ...................

 

[BoatinglifeupNorth]

10 hours ago, Alan de Enfield said:

 

Those ???   ???? could be the difference between a tank and a pencil

 

I said it was standard military issue - for the Artillery, The infantry had 'degree' compasses.

 

See for yourself ...................

 

What do I need to see? I see a standard military prismatic compass.
  You should know what I mean about the number. Have a look at the NSN(That’s Nato Stock Number for civvies) on the back of the  compass and fill in the * at a guess. W* 6605-99-***-8805, that just popped into my head for some reason?

 Everyone uses the same compass in Mls, as everyone sings off the same song sheet. Far quicker and more accurate for the infantry bloke to call in artillery/mortar support if everyone’s using the same compass and is taught how to use the same compass. They do not use Degs.

 I take it you’ve never been an Infantry man?

The military prismatic compass below, if you Google I’m sure you’ll be able to learn how to use it accurately.

 

 

Edited January 18 by BoatinglifeupNorth

[BoatinglifeupNorth]

British military prismatic compass NSN read off an Ebay item, 6605-99-537-9034, not that I’m really bothered but some “Walts” maybe.

Edited January 18 by BoatinglifeupNorth

[Alan de Enfield]

9 minutes ago, BoatinglifeupNorth said:

British military prismatic compass NSN read off an Ebay item, 6605-99-537-9034, not that I’m really bothered but some “Walts” maybe.

 

I really do not know why you are persuing this 'argument' - I have agreed that the mils compass is is a standard military issue - just not for the Infantry.

 

You should also remember that the NSN numbering system was introduced in the USA in 1974 and was taken on board by all NATO members over the subsequent years. My rifle is dated 1971

My compass predates the use of NSN numbers by the British and is marked up as 'Compass MKIII' (No NSN numbers)

 

My 'degrees' compass (standard issue to the infantry) also predates the NSN system and is marked up as 'Compass MKIII'

 

The Binoculars are marked "Bino. Prism No2 MkII"

 

The watch is a standard "CWC" issue and does have the NSN on the rear.

 

If you wish to discuss further please take it to PMs.