TechTip- Wire Current

[PRR]

How much current can this wire handle?

How big a wire do I need for this much current?

Here's a table from Littelfuse, aimed at automotive uses. Cars are low-voltage, vibration, and somewhat hot, a little like tube amps.

Note that we *never* have to worry about current when picking wires for plate supplies; worst-case a #20 can carry ten times more than the biggest B+ current we are likely to need.

Heater supplies can be an issue. Not for say a DeLuxe, 1.5 Amps, which is off the chart. However an SVT can be pulling 10 Amps of heater current in its big bottles, plus more for the little bottles. Say 12 Amps.

There's two types of wire insulation listed. Unless you are sure, assume the 90 C stuff. (Even THHN house-wire is not technically "cross-link".) So use the "(2)" columns, the "T" column not the "XL" column.

Note the severe de-rating for ambient temperature. I'm rewiring a tractor with a 50 Amp tap. Inside the cab I can almost use #12 (#12 GPT 47A at 25C), so I would use #10. But near the radiator could be over 100 C hot. #6 is a bit small (#6 GPT 47A at 80C), I really should get high-temp wire. In fact my tap is on the cooler end of an open bay, I have 150C rated #6 for the first foot and some tough #8 to run into the cab. Also my max-current (heater blower) will be in the coldest weather when it is zero C in the cab.

Guitar amp guts can run way over 25C but never 100C (we hope).

So for the SVT's 12A, #20 is OK when cool but #15 is needed at 80C. Assuming ventilation, #16 may be OK, #14 would be safer.

Since #14 is very awkward on Octal sockets, you might run the PT leads to lugs, run #16 to the big tube heaters, and #20 to the little tubes. Or even run two #20 to the big tubes in triplets.

[John]

Darn useful. Saved. Thanks!

[terminalgs]

thanks for posting.

[Ken Moon]

I did some calculations of capacities for teflon-jacketed copper wire at a length of 2 feet and a temperature range of 0 to 35 degrees celsius, based on MIL-STD-975, which is VERY conservative.

The bottom line is that 20 guage is plenty thick for heater wires, and 22 guage is fine for all signal wires.

The results:

22 guage - 4.42 amps
20 guage - 6.15 amps
18 guage - 8.56 amps

These figures are so conservative that they can easily be DOUBLED in electronic equipment applications (such as guitar amps) with no problems!

Bundled 20-ga wires in aircraft applications are usually limited in practice to 5 amps, but this is at vastly longer lengths in a much more demanding physical environment than guitar amps.

In a 4-foot loop, 20 gauge wire carrying 5 amps will have a voltage drop at the load of about 3.5% - not a problem.

If you want to be conservative, go with 20 gauge for heaters and 22 gauge for signal wires. You could, however, build your amp entirely with 22 gauge wire and you will never have a problem.

Note: Electrical code tables that show ampacity based on the 700 circular mils per amp rule do not apply to electronic equipment, but to power distribution.

here is a spreadsheet based on MIL-STD-975 that you can play with yourself:

[PRR]

> Electrical code tables that show ampacity based on the 700 circular mils per amp rule do not apply to electronic equipment, but to power distribution.

Which is why I posted more appropriate numbers.

For "transmission" and "chassis" numbers you can look at:
https://www.hagemeyerna.com/getdoc/93f497f0-29a2-4a84-bd81-8e529103ced1/Wire--basics-of-Ampacity-or-Copper-Wire-Current-Ca.aspx

That says #2 can carry 94A in transmission work. Here in Maine we use it for 100A service all the time; in Texas they may lean to a larger wire. (However a "100A" home rarely pulls 100A.)

It also says #14 can carry 6A for transmission. NEC for interior wire allows 15A, in cable. Similar for #12 which NEC allows at twice the 700 cm/A rule. So all our house walls are full of 1,400 cm/A, and rarely burn (about never, if NEC gauge and workmanship rules are carefully followed).

That table also shows "chassis wiring" and indeed #20 can do 11A; as you say, double your MIL-spec very-conservative computation. (However 35 deg C seems cool for a tube chassis, and some insulations must de-rate quickly for hot.)


Folks who absolutely can not use a slide-rule (or other calculating device) can *try* this tool:
http://www.powerstream.com/Wire_Size.htm

______________________________________________

This link posted at Harvard starts from a false premise then comes to useful results:
http://frank.harvard.edu/~coldwell/wire-gauge.pdf

He thinks wire gauge is a decibel function. In fact wire gauge is: you start with a rod and pull it through a slightly smaller hole, then a smaller hole, then a smaller hole. Anneal the hardness out, repeat repeat. #2 is nominally the 2nd draw from rod, #40 has been pulled through smaller holes 40 times. At each hole they take the maximum reduction possible, with reasonable breakage and die-wear.

The "max reduction" concept probably does lead to a quasi-log spacing. But wire gauges were set long before decibels or good resistance numbers existed.

If you can remember those formulas you don't have to carry a wire-table around.

Calling circular mils "perverse" is strange.

[Ken Moon]

You certainly know your stuff 

I didn't mean to say you presented incorrect data, but reading that sentence now, I see why you could easily take it that way.  Sorry.

It's always interesting to see how people from other fields approach things.  All my experience is with military aircraft wire bundles, So Mil-STD-975 is what I lived by.

It's good to see that the bottom line is the same - 20 gauge wire can carry any currents encountered in a guitar amp.

Best regards, Ken

BTW - Harvard guys are just wierd. If the definition of circular mil is perverse, he should propose a more elegant definition that is up to Haavaad standards 

[PRR]

> Harvard guys are just wierd.

Moreover, he's really a "software engineer".

[eleventeen]

"That table also shows "chassis wiring" and indeed #20 can do 11A;"


Dunno about any darned table but I would be VERY uncomfortable running 11 A thru AWG 20 wire. I flat out wouldn't consider that acceptable, table or no table. That's a clean double or more over anything I would pull thru it.


And I realize much of the table is that way, but come on, how can chassis wiring ampacity be almost 10x transmission capacity?

[DummyLoad]

"That table also shows "chassis wiring" and indeed #20 can do 11A;"


Dunno about any darned table but I would be VERY uncomfortable running 11 A thru AWG 20 wire. I flat out wouldn't consider that acceptable, table or no table. That's a clean double or more over anything I would pull thru it.


And I realize much of the table is that way, but come on, how can chassis wiring ampacity be almost 10x transmission capacity?

some perspective: a 20 mil PCB trace with a 1oz/sq in. Cu, with a 20degC rise can carry 1A on a surface trace. calculate that cross-section; that's not very much copper. you should re-think.

http://circuitcalculator.com/wordpress/2006/01/31/pcb-trace-width-calculator/
 
those tables PRR posted are VERY conservative. 20AWG can carry 11A. it will get little warm but it will and does. i wire 5-10A -48V systems with 20AWG with wire wrap termination as part of my job and so do most telcos with -48V DC plant.

i wouldn't hesitate for a second  to run 5-6A via 20AWG wire. i do, i run it for 3-6A heaters. for the preamp bottle strings i use 22AWG on a string run < 3A.

--pete

[PRR]

> uncomfortable running 11 A thru AWG 20 wire

The *copper* isn't going to melt.

http://www.powerstream.com/wire-fusing-currents.htm

#20 copper melts at 58 Amps.

The key question is the insulation.

Bad old rubber had to be derated to almost nothing at 60 deg C.

Asbestos can be run far hotter but is now out of style. Ceramic beads serve some high-temp applications. Some of the fancy-plastics hold up when hot.

LittelFuse, who is in the business of (not-)melting wires, seems to think #20 in Thermoplastic (today's basic stuff) is good for 15A at room temp, 6A in a hot-box.

In use we also want to check Voltage Drop. #20 is a bit over 10 Ohms per 1,000ft, so one foot is 0.01 Ohms. At 11 Amps we drop 0.11 Volts per foot. Assuming your 11 A (six 6550s?) is on the PT side of the chassis, you have a foot each way, 0.22V drop. This is not a large part of 6.3V, but enough to think about a larger gauge. If your six-pack-amp is broke up as an amp and a power supply with a 6-foot cord between, it is over a Volt of drop and you sure need to up-size just for voltage drop.