ultrasonic cleaner

[lacek]

Hi,
there is a now-famous video where Louis Rossmann compares an ultrasonic cleaner made by Crest to a Branson one. The first one shows better performance. Is there a consensus why was it so? He commented that this is due to the fact that Crest sweeps the ultrasonic frequency. I have checked brochure of the Branson one: they also sweep it.

At this point I would like to ask:
- what determines that Crest clean better than Branson? Transducer power? Tank size?
- Crest also offers 135 kHz cleaners - how would they compare to 40 kHz for electronics use?

[mcinque]

I have not much experience in PCB ultrasonic cleaning (I try to use them the less possible to avoid damaging some components), but it's since 1995 that I use them to clean small mechanic parts. In that particular sector, I can consider myself as an "experienced user".

In my opinion differences are in:

1. power/litre ratio
2. no. of transducers
3. tranducers positions

Branson is a respectable brand in ultrasonic industry, also for reliability: we had branson cleaner at work turned on for 8h 3-4 days a week for many years with heating on and had no issues at all. The tank was still perfect with no corrosion.

Since I worked with Brandon cleaners for 12 years I can't believe they don't work properly; if it was really like Louis was reporting in the video, the Branson had a fault.

In my opinion, it would had be better to compare them with the alluminium foil test to make a proper comparative test and see if the Branson had a fault.

The sweep is helpful to avoid stressing some components and to avoid dark spots. But since different brands use different sweep timings (for example Branson uses a quick 4Hz sweep while Elma 1Hz) you can have different results (quicker or slower cleaning time) but a board/component should came out properly cleaned in any good ultrasonic cleaner with a proper cleaning time and proper solution.

37KHz are a general cleaning frequency and is a general purpouse that can be used on any "general" part.
They (manufacturers) say that +125KHz are much gentle and can clean most sensitive parts without damaging them, but I don't bet on it with quartzs and mems.

Generally speaking the higher the frequency the smaller the spot the cavitation can reach (so it cleans smaller spaces) but less "violent" is the cavitation, so the cleaning action.

Comparing ultrasonic to a toothbrush, 25KHz are hard bristles, 37KHz are normal bristles and +125Khz are soft bristles.

[lacek]

Thanks for the answer. That cleaning size is something I do not get fully. Given the speed of sound in water of 1480m/s, then even 135 kHz cleaner will have a wavelength of approximately 1cm. This feels much too large than component size. In the end what matters is the bubble (cavity) size and this is much shorter than the wavelength- hence cleaning is possible. The bubbles form in the places where the wave oscillates rapidly.
Thus, I would presume that no matter whether we use 37kHz or 135 kHz the wave will not penetrate say under a BGA chip and no cavities will form under a BGA chip. Anyway the cavities are much smaller at 135kHz, so when they implode: i) they will not be able to remove the heavy stuff 2) the cleaning action is not from implosion but from "microstreaming".

So it seems that 135kHz is much more appropriate for cleaning the microelectronics, though most likely removing flux is something for which "soft brush" would be insufficient. And in no way one can clean under BGA chips.

[mcinque]

Thanks for the answer. That cleaning size is something I do not get fully. Given the speed of sound in water of 1480m/s, then even 135 kHz cleaner will have a wavelength of approximately 1cm. This feels much too large than component size.

I guess we should consider also reflections in the tank.
And in any case, that's what sweeping is designed for. It can "move" costantly the position of the wave in the solution.

And in no way one can clean under BGA chips.

I guess with sweeping you can clean under them, there is effectively some space between a bga and a pcb.

[lacek]

Thanks for the answer. That cleaning size is something I do not get fully. Given the speed of sound in water of 1480m/s, then even 135 kHz cleaner will have a wavelength of approximately 1cm. This feels much too large than component size.

I guess we should consider also reflections in the tank.
And in any case, that's what sweeping is designed for. It can "move" costantly the position of the wave in the solution.

I think that reflections do not change much: the calculated wavelength is much shorter than the size of the tank, so the eigenfunction of the tank, at least close to the middle of the tank will have almost free space wavelength.

And in no way one can clean under BGA chips.

I guess with sweeping you can clean under them, there is effectively some space between a bga and a pcb.

Yes, but this is like 0.1mm.   My understanding is that bubbles occur close to antinodes of the wave, and wave will not penetrate regions 100 times narrower (and 300 for 37kHz) than its wavelength. More precisely: I think that regions under BGA chips are always dark, no matter what frequency (they would be illuminated by the ultrasonic wave for frequency of say 10MHz).

[mcinque]

We sometimes used ultrasonic to uncloggle blood siringe needles, I guess that in someway the cleaning action works also in small spaces.