Voltage drop/ Cable size query

[TommyBoy]

Hi All,

Quick brain pick.

If I don't care about voltage drop (eg. all components on the circuit will run between 100-250V) and current capacity is within limits, are there any other considerations when deciding to use a smaller than 'normal' cable size?

200m run - 10 x 300w loads equally spaced. Cable calcs say 10mm (SWA) but a 4mm could take the load. This is not connected directly to the system - this is a chain of architectural LED lamps coming out of a 16A Commando.

Whadyathink?

 

[Marvo]

I'd suggest you do a live test on the load to establish the exact current draw using an RMS clamp meter. Given the wide voltage range I'd guess it's supplied by a driver or PSU of some sort. If the PSU is rated at 300w output if could be as high as 450 watts on the input side so all of a sudden your 600watts of drivers is actually a 900watt load which can very easily affect the circuit design.

 

[TommyBoy]

Thanks for your reply Marvo,

Unfortunately I've sent the test units back so all I have is the tech spiel but it does say the rated 110W is the max load on the supply side. If this is the case would you think I'd have to worry about the drop in voltage so much?

 

[Richard Burns]

You have a variance between 10 x 300W loads and a maximum load on the supply of 110W which does not seem compatible.

If the equipment is rated to be supplied by a lower voltage, so long as it is part of the product standard, there should be no particular reason to increase the cable size based on the 3% voltage drop limit in BS7671.

The cable must be able to take the design current of the installed equipment.

However this does become somewhat blurred once it is a plugged in piece of equipment, but considering the nature of the installation I think compliance with the wiring regulations would be suitable.

 

[TommyBoy]

Apologies, it should be 330W but for simplicity I wrote 300. To explain the variance i may as well describe the installation.

I have two circuits with 16A commando sockets at a building at one end of a park, there are two groups of trees. The largest of the groups has 10. Each tree is to have three 110W RGBW DMX controlled uplighters (with 100-250V PSU/Driver/DMX decoder inbuilt) fixed around the crown. Cables to be run from one to the next at a height of around 15m to enclosures to house a circuit breaker/fuse to protect each trio of fittings. This is a semi-permanent install, it will go up for around two months starting in November and then be replaced at the same time on consecutive years.

A 6mm cable could take the design current. 10mm will be too obvious, since it's quite a prestigious location I'd like to make it as discreet as possible (since they are a large land owner it's vital that this first install impresses for future business)

 

[Richard Burns]

Wow, talk about making it complicated and inconvenient!!

Anyway, your total design current is 14A; over the arrangement you describe your volt drop would be about max 5% to about 218V.

Even if you used a 1.5mm² cable, that can also take the load, then your volt drop would be about 20% down to 185V.

In each case the PSU would not be troubled by this, I am not sure how integrated the rest of the kit is but if the voltage range on input is as specified then I cannot see a problem.
Is there not fault protection built into each unit?

 

[Lucien Nunes]

Switched-mode drivers run at constant input power, demanding more current as the voltage falls. If you are going to allow a non-trivial amount of drop, the currents will increase by a non-trivial amount and increase the drop yet further. It becomes a load-line calculation or you will need to solve simultaneous equations for each tap-off point as the current drawn by each driver will be different.

Because LED drivers often have significant inrush, you might need to consider the VD at that inrush current, if all drivers are to be able to start up at once. Otherwise the voltage at the far end might fall so low that the units don't start up correctly, or repeatedly shut each other down by pulling the voltage below the minimum rating, or interact in some annoying way.

 

[TommyBoy]

Richard, tell me about it! This project is reaching a while new level of inconvenience, I'm not even going to go in to the planning permission!! You are correct, the device will have a thermal cut out but is that sufficient? They are rather expensive, and I think I'd rather one tree go out in case of fault rather than half the park. I could be going OTT though...

Lucien, I'm expecting the drivers to be fed 24/7 and switch them via DMX. Perhaps I should write a program so that each light 'wakes up' in turn at the beginning of that day's duty just to be safe.

Thank you all for your advice, much appreciated. Perhaps the 4mm is back on the table!!

 

[Richard Burns]

The circuit providing the 16A commando sockets must have a level of protection, presumably each socket for each group of trees are own their own circuit.
You would need to ensure that any protective device in a tree (at 15m high and inaccessible?) would discriminate from the circuit protective device other wise you would lose the group anyway.

 

[shanky887614]

cant you bury the cable and loop through adaptable boxes?

you could replace the swa hole with metal stop ends each year (use piranha nuts to make it neater)

 

[TommyBoy]

Richard, I think you're right on the money there. Although I'm pretty used to call backs with a cherry picker in tow, this area has particularly difficult access; perhaps half the park is doable.

Shanky, unfortunately I can't do anything to the ground. I can't even park the truck-mount on the grass!

 

[shanky887614]

Richard, I think you're right on the money there. Although I'm pretty used to call backs with a cherry picker in tow, this area has particularly difficult access; perhaps half the park is doable.

Shanky, unfortunately I can't do anything to the ground. I can't even park the truck-mount on the grass!

get the client to pay for a tree surgion to climb the tree with chains and fit them then.

if they want it but wont allow access equipment up then they can pay through the nose for it

 

[davesparks]

Have you considered the EFLI of the circuit?

 

[TommyBoy]

Have you considered the EFLI of the circuit?

Since everything I'm putting in place connects via commando socket it won't be classed as part of the fixed installation. The sockets themselves are probably only 10m from the DB, which in turn is fed from a substation beneath the park. I should hope the readings are fairly healthy.

 

[TommyBoy]

get the client to pay for a tree surgion to climb the tree with chains and fit them then.

if they want it but wont allow access equipment up then they can pay through the nose for it

Truck mounted cherry pickers from the paths and pavement surrounding. It's a big square with a fountain in the middle so with a decent outreach I'll be fine. We've done installations there before but nothing with the kind of load or complexity as this. The environmental officer is pretty hot on us too so everything has to be 'kind' to the trees - all material straps, no fixing with anything tight that could strangle the branches as it grows etc. And definitely no machines on the grass, made that mistake last time [emoji51].

 

[marconi]

Apologies, it should be 330W but for simplicity I wrote 300. To explain the variance i may as well describe the installation.

I have two circuits with 16A commando sockets at a building at one end of a park, there are two groups of trees. The largest of the groups has 10. Each tree is to have three 110W RGBW DMX controlled uplighters (with 100-250V PSU/Driver/DMX decoder inbuilt) fixed around the crown. Cables to be run from one to the next at a height of around 15m to enclosures to house a circuit breaker/fuse to protect each trio of fittings. This is a semi-permanent install, it will go up for around two months starting in November and then be replaced at the same time on consecutive years.

A 6mm cable could take the design current. 10mm will be too obvious, since it's quite a prestigious location I'd like to make it as discreet as possible (since they are a large land owner it's vital that this first install impresses for future business)

For my own amusement - I will not be posting the working out - could you tell me the lengths and of the cable runs:

a. DB to Commando sockets?
b. Commando socket to first tree?
c. Distance between trees?
d. CSA of cable run between DB to Commando sockets?
e. Will the lamp fittings be all turned on at the same time or in a sequence?
f. Once the lamps are on are they kept on or is some form of light show intended?

What nags in my mind is ensuring the long cable run has a sufficiently low impedance so that amalgam of all the 'top hat' current waveform of the power supplies does not create problems - such as an odd lighting effect on the observer's eye when the trees are viewed together. It would be a risk I think to do the installation without confirming not.

No promises by me to come up with a result but if I do I will post something.

 

[davesparks]

Since everything I'm putting in place connects via commando socket it won't be classed as part of the fixed installation. The sockets themselves are probably only 10m from the DB, which in turn is fed from a substation beneath the park. I should hope the readings are fairly healthy.

So what? You still need to ensure that your OCPD will disconnect in the correct time. You are working to Bs7671 and bs7909 aren't you?

 

[TommyBoy]

Marconi, some numbers for you. Please excuse the delay, I wanted to get them as accurate as i could.

a. DB to Commando sockets?
b. Commando socket to first tree?
c. Distance between trees?
d. CSA of cable run between DB to Commando sockets?
e. Will the lamp fittings be all turned on at the same time or in a sequence?
f. Once the lamps are on are they kept on or is some form of light show intended?

A. 15m
B. 18m
C. 26, 24, 16, 27, 26, 27, 18, 20, 18. (This is a maximum distance as I have allowed 10m extra per tree to dress from the anchor point to the adaptable box and back)
D. 10mm
E. As previously advised, I will write a program to start each fitting in turn at the beginning of the day.
F. There will be a light show, usually smooth fades. Since they are RGBW they will only ever be at 3/4 power at the most - All three colours and white will not be used together.