48VDC preamp, gettin' a spark when I plug it in ... :O

I have a handful of preamp builds that use a 48V DC wall wart. I know this is likely caused by the filter caps not filling up fast enough. (I'm using 470uF to filter my DC power for the whole unit.) It doesn't really seem to matter which DC power brick I use, even though some of them have different resistances. The spark isn't  dangerous and I've been using it for years without any problems manifesting, but I would kiiiinnda like to fix the issue as I assume something will go wrong at some point.

What are my options for fixing this?

Put a few Ohms of resistance between the cap in the wart and the cap in the preamp.

midwayfair said:

I have a handful of preamp builds that use a 48V DC wall wart. I know this is likely caused by the filter caps not filling up fast enough. (I'm using 470uF to filter my DC power for the whole unit.) It doesn't really seem to matter which DC power brick I use, even though some of them have different resistances. The spark isn't  dangerous and I've been using it for years without any problems manifesting, but I would kiiiinnda like to fix the issue as I assume something will go wrong at some point.

What are my options for fixing this?

Click to expand...

You are connecting a charge capacitor to an uncharged one, so in order to reach equilibrium, an electric charge must be transferred, which means current, and quite sudden at that.
Follow PRR's suggestion. There is a compromise to be found between reduced current spike and loss of voltage due to Ohm's law.

I guess the "charged capacitor" on the PSU side explains why I don't get a spark if I, say, turn on a power strip while the PSU is plugged in.

I'm pulling about 20-30mA each preamp so it looks like a 10R  will drop less than a volt while the preamps are running, which is acceptable.

What I don't know is how to calculate the wattage for the resistor. What do I use as the current? The full 1A that the wart can supply?

midwayfair said:

I guess the "charged capacitor" on the PSU side explains why I don't get a spark if I, say, turn on a power strip while the PSU is plugged in.

Click to expand...

That's correct.

What I don't know is how to calculate the wattage for the resistor. What do I use as the current? The full 1A that the wart can supply?

Click to expand...

You can use the actual current drawn by whatever is connected, but for safety you should use the 1A rating. That suggests a 10W resistor.

Thanks so much for the help!

> I'm pulling about 20-30mA each preamp so it looks like a 10R  will drop less than a volt .... What I don't know is how to calculate the wattage for the resistor.

0.030A times <1V is 0.030 Watts. You can hardly buy a resistor too small.

> What do I use as the current? The full 1A that the wart can supply?

Why not the 100+ Amps that the wart's cap (if extra-good) can supply (for a very short time)?

However there IS an issue. At the moment of contact the current is still HIGH. And the voltage is full no-load supply voltage. So big power. Power falls as the preamp cap charges, and the time is short.

Here's a plot. I assume your "12V 1A" supply really means 20% sag and >1A for some time will make smoke. So it is really a 14V supply with 2 Ohms of loss. This loss is mostly in the transformer, the cap is straight to the jack.

At the instant of connection, wart cap is at 14V and preamp cap is at zero V. Say there's NO resistance in the line or jack or caps. Current is then 14V/0r or INFINITE Amps (and Watts). A pair of real caps this size may be 1r total DCR. 14V/1r is 14 Amps, 14 Volts, 196 Watts.

Putting 10 Ohms in there makes 1.4 Amps, still 14V, 19.6 Watts. (Abbey is using the assumption that your wart CAN'T dump >1A, which would be valid for some switchers, but not a basic linear wart.) Time constant is like 10mS(?). In plot below, red spike is V(R68)*I(R68) or power in the 10 Ohm resistor. If checked at 1 Second it is 0.012W, no trouble. But AT plug-in it spikes to almost 20 Watts, and does not come down to 1W until 26mS.

Can a common 1/4W resistor stand this spike? Ratings are no help. They tend to say 200% for 1 Second, not milliSecond abuse.

Put 1W in a 1/4 resistor (12V in 150r 1/4W) while you are holding it. It gets HOT but it takes a second or more. (Discount, because it takes time for skin to heat-through to the nerve; this is just for concept.) If it takes OTOO a second to hot-up, how hot can it get in a few mS? Hot enough to burn the element?

The old-old Carbon Compositions had fat coal/clay elements and would stand all sorts of spike abuse for decades. Modern Film resistors obviously have less resistor-stuff and may burn-up faster. I would have no fear of using a 1/4W film part in this application for stage use. If I were protecting a military installation and the mission required constant hot-plugging, I might spec a fatter resistor (or even ramped power).

midwayfair said:

What are my options for fixing this?

Click to expand...

The right solution is using an inrush current limiter for nominal current of 1A or less.
For example
https://hr.mouser.com/datasheet/2/600/SL0560001-972867.pdf

A 1A 60r thermistor working at 0.030A is likely to stay 59 Ohms and cause a 2V drop.

You usually pick a thermistor to run near half rated current.I dunno if they come that small.

The arc won't kill anybody. The peak power of (un-resisted) 1,400 Watts burns the plug contacts. The mS duration of the abuse limits contact damage. (The total energy is unchanged with any reasonable resistance: that cap has to fill-up.) The connector is probably OK. A little resistance (or thermistor) may take the edge off the visible arc and user-startle, and adds a useful stage of power-crap filter.

IIUC, there are more than one preamp supplying from a 48V PS.
"I'm pulling about 20-30mA each preamp so it looks like a 10R  will drop less than a volt while the preamps are running, which is acceptable."
So I assumed  the NTC in pdf  (it is very small, 5mm in dia) will be warm  at  100mA. There are NTC with smaller resistance and max. current https://www.ametherm.com/datasheetspdf/SL0530001.pdf also.
Anyway, I  just like to say that the correct engineer's choice IMO is placing there a NTC, not an ordinary resistor (non-flammable preferably).

Thanks for the thorough breakdown.

"IIUC, there are more than one preamp supplying from a 48V PS."

I have three different units that I built to run this way, with a total of 8 preamps. Two of them are baby animals -- which have the 10R resistor already, and it's actually only 1/4W with no issues.

The other ones are variations on the Hamptone FETboy, or at least they have something like the Hamptone's high-current output stage (three of them are BJT, not FETs). I have some power filtering on each preamp's board, but I had added a 470uF to the entire box of multiple preamps, hence the sparks.

I can just get a couple 10R and a couple of the NTCs to try out both. A 1-2V drop isn't too bad, I don't think any of my condensers will complain too much.