Unusual phenomonem

I clean up prototype PCBs in an ultrasonic bath, along with the glasses and anything else.
I put a board in with a couple of LED's on the board.
Bear in mind this is in water with a dash of detergent.
Amazingly, the LEDs (a blue one and a red one) started to glow. There were no other components on the board apart from headers.
I've heard of fluorescent tubes glowing in microwave ovens, but LEDs in an ultrasonic bath?
I'm guessing that the bath runs at roundabout 20 -kHz and certainly gives you a tingle if you stick your fingers in.
So somewhere, a voltage in the order of volts is being transmitted to the LEDs.
Now I know why ultrasonic descaling at the dentist is so damn painful, and I have to go in a couple of days time. Ouch.
If anyone has any thoughts, I'd like to hear.

It is more likely that the ultrasonic waves are dislocating the water into positive and negitave ions and that is being picked up by the LED and powering it. Modern LEDs require very little current to light. In fact doing a resistance measurement on an LED with a DVM is enough to light it up.

Try sticking a DVM set on volts into the water and see if it shows anything, if not try the current settings.

At a minimum, you must flush all the detergent off the board using distilled water. Otherwise your cleaning accomplishes little. The best method is use an isopropal alcohol (IPA) solution in water and an old toothbrush to clean the board.

The ultrasonic cleaning may also loosen the copper traces which are epoxied to the surface of the circuit board. With prototype boards, probably not a problem. Could also destroy the internal connections of components, which will only be found when you power the board. Not a good idea!

Paul

I'll try that with the DVM.
I thought that the ultrasonic action was mainly due to cavitation - a powerful one produces a fine mist.
I wonder if the bubble cavitation and collapse is having an electrostatic effect.
The dentist uses something the size of the normal drill and a fair amount of ultrasonic power is concentrated into the tip. All I know is that it gives me gyp round some of my molars, a bit like biting on aluminium foil.
All medical equipment has to have very low leakage currents. I'll ask the lady on Wednesday or take an LED along. Interesting

Hi Paul
When I do a board, I do exactly as you say. I give it a scrub with ethanol or propanol and a toothbrush to soften up the flux, then straight into the ulrtrasonic. The wash water is mainly distilled water with literally one drop of washing up liquid for wetting. After the clean, it gets a rinse with distilled and a warm air dry.
All the boards are prototypes and I never use it with unsealed active components.
It's by far the most effective method there is for deep cleaning and I have never had an "iffy" board as a result.

I use 99.99% isopropyl alchohol and a brush.

My "local" electronics house carries 1 gallon bottles, I fill my little pump cleaning bottle as needed.

Gallon lasts a long time.

tigger:
I clean up prototype PCBs in an ultrasonic bath, along with the glasses and anything else.
I put a board in with a couple of LED's on the board.
Bear in mind this is in water with a dash of detergent.
Amazingly, the LEDs (a blue one and a red one) started to glow. There were no other components on the board apart from headers.
I've heard of fluorescent tubes glowing in microwave ovens, but LEDs in an ultrasonic bath?
I'm guessing that the bath runs at roundabout 20 -kHz and certainly gives you a tingle if you stick your fingers in.
So somewhere, a voltage in the order of volts is being transmitted to the LEDs.
Now I know why ultrasonic descaling at the dentist is so damn painful, and I have to go in a couple of days time. Ouch.
If anyone has any thoughts, I'd like to hear.

Some ultrasonic cleaners have large diameter transducers, so I can see the mag field causing some induction on to the PCB tracks.
What power rating is the cleaner?
You might be worth putting an LED in with a small loop connecting the two leads and see what happens.
Tom.... :slight_smile:

Crossroads - I do like these dispensing bottles. I use IdealTek, expensive but well engineered. One has alcolhol (propanol or ethanol) the other white spirit. Somewhere along the career line, I "acquired" a 2.5-litre bottle of Absolute Alcohol (99.9% ethanol) which I keep for special occasions. In the UK, many of these useful chemicals are getting harder to find, thanks to misuse. At a pinch, vodka has to do.

Tom - I like the induction idea. This only a small tank, generally used for jewelry or specs, possibly in the order of 10's of Watts. I used to work with large Kerry ultrasonic baths in the kW power range. Their main use was for either getting difficult chemicals to dissolve (making up reagents), or de-gassing liquids. Not unknown for the occasional piece of engine getting a scrub. Just confirmed, bath is 70-W, put a loop of copper wire on the legs of a red LED (identical to LED on board) and not a glimmer, LED working fine. However, cut the wire loop, DMV on volts across the legs and just with one leg in the water, LED on, and anything up to 17-V over the transducer and 12 to 15-V away.

Interesting, and will be more careful in future about what gets a wash.

You actually drink from glasses and "anything else" that you've washed along with your PCB's?

Interesting - this might be the answer: Electroacoustic phenomena - Wikipedia

aarg - not the vodka glasses, the ones you wear AKA "spectacles", as in opticians.
They charge about a tenner for dipping your specs in their ultrasonic bath, but I don't think they would do the PCBs.
Joking aside, a small ultrasonic bath is a very useful bit of kit. They used to cost a fortune, but like a lot of things, the global market has brought the cost down dramatically. Mine came from CPC, a division of Farnell and cost about £30.

MarkT - I think electroacoustic gets the vote.

MarkT:
Interesting - this might be the answer: Electroacoustic phenomena - Wikipedia

Interesting effect and an interesting link there. Thought I'd try it...

Well, with a red LED and an ultrasonic cleaner I tried:

The LED on it's own, up, down, left, right, suspended, touching the bottom...
LED with legs bent out
LED with bits of wire attached
LED with (several different values of) MLCC capacitor soldered on (thinking of the acoustic susceptibility of those things)

... not a sausage, not even a glimmer :frowning:

So, if anyone can replicate this - and/or if tigger can post a picture of the PCB I'd be really interested

Yours,
TonyWilk

P.S. yes, I did test the LED, it worked fine after.

Hi,
Thinking that we are working with kHz frequencies, the LED diode junction my not be fast enough.

The circuit that @tigger had my have had some highspeed rectifying components in circuit.
Such as transistor junction.
So you may need to put a highspeed diode, Schottky even, in one of the LEDs legs.

Just a thought.. Tom... :slight_smile:
PS, No it wasn't an epiphany, they happen suddenly in your sleep at 3.00AM. (I've left myself open for that haven't I?)

Hi Tony
Here's my set-up - no pic I'm afraid.
Take one ultra bright red LED. Connect the legs with a tinned copper wire loop about 5-cm diameter. Put it in the bath, touching nothing except the water which is distilled water (old stock) with a drop (literally) of ordinarily washing up liquid. The conductivity is about 150-uSiemens.
Turn the bath on - nothing, so probably not induced.
Snip the loop of copper wire roughly in the middle.
Connect a Fluke Series 80 DMM in volts mode to the legs.
Dip one leg into the liquid, LED on, 17-V (steady) on the DMM.
I have never seen this in a many years of using ultrasonic baths.
The original PCB was a FR4 perf board with nothing on it other than some PCB headers and the dropper resistors for the LEDs - this a pre-wire wrapping clean.
The blue ultra bright LED and the ditto red LED were both clearly on, not just dim, but clearly on in daylight.
No contact with anything but the water, no vodka in the glasses.
Given that a blue LED needs about 3-V of forward voltage just to start, tells you that something is going on.
I've just repeated the LED and loop set up. One leg in the water (either works), DMM connected, anything up to 28-V on DC range showing. I touch the other leg, now it's really bright.
Now another thought starts - is this some sort of current/voltage leakage from the bath (stainless steel) to the water. The bath appears to be Class 2 (old double insulated with no earth). It's also on a 30-mA RCD.
Whatever the reason, some caution needs to be taken when dipping boards like this.
I've never done it with an MCU on the board, but plenty of CMOS logic, and FET/BJT devices have come through unscathed.

Hi,

Connect a Fluke Series 80 DMM in volts mode to the legs.
Dip one leg into the liquid, LED on, 17-V (steady) on the DMM.

AC or DC mode?

What if you dip both meter probes, no LED?

Dipping on leg in liquid, not really telling us anything cos the other legs is out of fluid.

Tom.. :slight_smile:

tigger:
Hi Tony
Here's my set-up - no pic I'm afraid.
Take one ultra bright red LED.
snip

Ah... success!
Last time I was holding the LED with insulated pliers, this time I was dipping fingers in.

What I found was leakage - the LED would light if I held one lead and dipped the other in the water. It would also light if I touched the other lead to the metal tank while the ultrasonic was on.

Holding one lead of the meter (Fluke 75 on Vac) and dipping the other probe or touching the metal read around 14Vac.
I couldn't get any voltage differential reading greater than about 0.5Vac within the water or from water to tank.

Holding the LED and wires with pliers, it wouldn't light anywhere.

This doesn't sound exactly like what you found, assuming your PCB lit up just sat there on it's own in the tank.

Yours,
TonyWilk

Hi,
With your PCB LED phenomenon was ALL of the PCB submerged?
Or part of the PCB touching the metal tank.

What @TonyWilk has found would explain it.

Tom... :o

AC or DC mode?

DC

What if you dip both meter probes, no LED?

Didn't try, but will

Dipping on leg in liquid, not really telling us anything cos the other legs is out of fluid.

But it's not making a circuit, is it?

With your PCB LED phenomenon was ALL of the PCB submerged?
Or part of the PCB touching the metal tank

Can't remember, but it's all wire-wrapped now.
I'll muddy the waters a bit.
I've just aken it apart - quite nicely made for a cheapo.
There is a substantial transducer bonded to the bottom of the stainless with epoxy, and checking with the meter, there is no electrical contact between tank and transducer and no other connections.
So there is a tank, not electrically connected to anything (all plastic enclosure) vibrating ultrasonically, all dry.
I can't see that leakage comes into it.
I'm going to hedge my bets and come down in favour of electroacoustic 51% and some other effect 49%
I suspect without a lot of effort, it's impossible to say.
It is safe to use, but I will be careful what goes in now.
Thanks for all the suggestions, something else learnt thanks to an LED

There's also triboelectric effect to consider, or simple dry friction between the container and its mountings, although that seems unlikely as the container would be a reasonable Faraday cage for the circuit and protect it,
unless the circuit was being connected electrically to outside world.