Author Topic: Matching ultrasonic transducer (Read 6191 times)

watches01 · « **on:** February 21, 2018, 06:00:03 pm »

Hello,

I have a watch cleaning machine that i would like to rebuild before it goes bad. I'd like to at least change the capacitors on the ultrasonic generator board. I've read that the board has to be matched to the transducer or it will fail. On the older tube models I know you can just adjust them for maximum cavitation. This one is solid state (late 70's - early 80's). Is it enough to just listen to it and adjust it or should it be done with a scope? I don't own an oscilloscope yet but I will buy one if necessary. Can anyone explain how to do it with a scope? Any help is greatly appreciated. Thanks!

Samogon · « **Reply #1 on:** February 22, 2018, 10:12:30 pm »

I don't think that resonating frequency mismatch can do any hard except wasted money. I would pay attention to load capability. For instance transducer with higher power rating can overload output mosfets and as result burn them.

watches01 · « **Reply #2 on:** February 23, 2018, 04:04:29 am »

Thanks for the answer. I won’t be changing anything like that. I have two of these, one was rebuilt years ago and the ultrasonic is strong. I bought a second machine for a back up and it is weak so I never used it. I’m afraid it’s going to burn up like the first one did so I’d like to change the capacitors before it does. I just don’t want to cause more harm.

Samogon · « **Reply #3 on:** February 23, 2018, 05:14:06 am »

If it is weak, that could mean that final amplifier stages are damaged, may be not power mosfets but driving preamp stage. But be carefull with oscillooscope, there could be hundreds of volts amplitude on final stage. It is safe to use handheld scopemeters like fluke 97 or fluke 123,124 they can measure up to 1kV with proper probes. This scopemeters can be not very expansive 97 can be around $100.
But if you plan to do electronics i would buy new rigol or siglent, but then you will have to get expansive high voltage probe.

Armadillo · « **Reply #4 on:** February 23, 2018, 09:13:10 am »

Please don't have the wrong notion that changing the capacitors would solve your problems, furthermore most of the capacitors are film type capacitors like Polypropylene Capacitors.

The ultrasonic transducers are moving parts like speakers, don't expect them to last forever. It looks like a space capsule and can be bought even with bigger wattage, of course with matching driver board and the frequency you want.

The transducers [space capsule] have 2 connection [each] from the driver board which are "live", so don't accidentally touch on it. It needs to be loaded [not empty], otherwise it spoil easily. Measure the leads with a tong / clamp current meter than a oscilloscope would be easier.

IMO, changing the capacitors would be a waste of money for you.

I think the guy that rebuilt for you had done a fine job.

Armadillo · « **Reply #5 on:** February 23, 2018, 09:18:51 am »

Or wound 3 or 4 turns around the leads and connect to the oscilloscope so that you can see if the frequency has drifted [out of tune].

watches01 · « **Reply #6 on:** February 24, 2018, 02:43:50 am »

Thanks again for the responses. That’s what I’m wondering, has the frequency drifted and is it out of tune? I’m a beginner in electronics and I still don’t understand how to check that and I don’t know what the frequency should be. The place that fixed the first one is long gone. This is not a tank style cleaner with the horn, it’s an automatic machine with a drum style transducer that has a motor shaft running through it. The whole head goes into the jars of solution. It’s probably the same procedure anyway. I’d love to be able to maintain these machines myself.

Armadillo · « **Reply #7 on:** February 24, 2018, 04:39:16 am »

Must be some expensive watch cleaning business. Is that a custom machine? So there must be a brand and model number?
Post some pictures of the board and the transducer / tumbler drum would be easier for discussion.

You can make reference measurement on the good cleaner frequency and current for the weaker one.

You can sure maintain these machines by first getting your hands dirty, I mean you need to dismantle it by your hands to study it.

Cheers;

watches01 · « **Reply #8 on:** February 24, 2018, 07:29:34 am »

It’s a Tempo 400 made by L&R. I tried to post pictures but for some reason they won’t load from my iPhone. They are pretty expensive machines. I took one apart ten or fifteen years ago but I don’t remember much of it. I’m going to start on this one as soon as I can.

Armadillo · « **Reply #9 on:** February 24, 2018, 08:08:21 am »

I can't visualize how "ultrasonic" is it. Even the operator manual didn't use the word ultrasonic. I have not seen one before surely.

If you do teardown, maybe you can capture it in video and be the first and only one in the world.

At the hindsight, don't look difficult.

stevelup · « **Reply #10 on:** February 24, 2018, 08:35:13 am »

the word ultrasonic appears dozens of times in your attached PDF!

Armadillo · « **Reply #11 on:** February 24, 2018, 10:16:54 am »

I mean the title block

Yeah, its ultrasonic. Maybe the video didn't press the ultrasonic switch.

Armadillo · « **Reply #12 on:** February 24, 2018, 11:18:41 pm »

The ultrasonic frequency is designed at 100KHz;

You can tap the feedback signal from terminal 4 [Blue Wire] of the transducer mounted below the reversible motor and adjust trimpot R5 on the circuit board until somewhere around 100KHz, while your tong meter placed on the line of the driver board read maximum peak current [around ~4A peak, my guesstimation].

I do not recommend adjust R5 by visual and feel approximation

.

If not happening, then there could be some circuit problems.

I suggest you first do what you wanted initially. Change capacitors. That's what you can do without a oscilloscope.

There are 2 electrolytics which I think you like to change C3 and C4.

OK, hope it helps. We wait for your photos and updates.

watches01 · « **Reply #13 on:** February 25, 2018, 04:50:38 am »

That’s exactly what I needed to know. Thank you so much for that information! I couldn’t find any information on this machine. Do I need high voltage probes to tap that feedback signal? I’m wondering if changing the caps will change the tuning, I’m guessing it would. This gives me a good excuse to order an oscilloscope and a clamp meter. Thanks again I really appreciate you taking the time to answer this!

Armadillo · « **Reply #14 on:** February 25, 2018, 06:14:05 am »

The transducer driver circuit board is not isolated from the "Line" hence you "CANNOT" probe the board directly even though the feedback signal [transducer terminal 4 blue wire] is low voltage signal. You do not need high voltage probe. So either you
1. power the entire driver board with isolated power supply and probe the blue wire terminal directly
OR
2. Wrap 3 or 4 turns of magnet wire around the one of the wire leading to the transducer directly [not the blue wire] and connect this magnet wires to the oscilloscope.

At the same time;
At the transducer, there is a capacitor in parallel with the transducer which resonate with the inductance of the transducer. You need to check that this capacitor is OK. Failure of this capacitor is the loss of power even though the driver frequency is correct.

The 2 electrolytic caps are DC smoothing capacitors.

You know you will be dealing with lethal voltages and take safety precautions and also order safety gears like gloves and eyes protection etc.

I have a feeling that I am offending..... some people. $:-\$

I do not know if its worth to buy instruments to maintain this, but it will be sure a lot easier to let the "Experts" do it professionally for you.

watches01 · « **Reply #15 on:** February 25, 2018, 08:05:08 pm »

I bought an Isotap isolation transformer so I should be ok there. Ill start with those capacitors and see what happens, but I’d like to have the scope anyway so I’ll probably order one. Thanks again and I’ll keep you posted.

watches01 · « **Reply #16 on:** March 03, 2018, 02:27:23 am »

I had some time today so I started on the machine. It looks like there is only one electrolytic capacitor on the generator board and it tests ok with a peak esr meter. Tomorrow I’ll take it out and test it with my capacitor tester for leakage. If it looks ok then I think the next step is to order the oscilloscope and try to see if the frequency drifted for some reason. I did find that the large power transistor wasn’t quite as tight as it should be so maybe the ground connection isn’t quite there. I’ll change all the electrolytic capacitors on all three boards while it’s apart. I did find one bad one on the motor control board. I’ll also remove the orange drop cap that is in parallel with the transducer and check it, but I don’t think those usually go bad. If you notice the one film cap is unsoldered on one side it’s because I had to do that to test it. It didn’t test good in circuit but does test good out of circuit. Hopefully I can post the picture from my iPhone.

watches01 · « **Reply #17 on:** March 03, 2018, 02:30:48 am »

Nope. Last time it was the wrong extension. This time the file is too large.

Armadillo · « **Reply #18 on:** March 03, 2018, 04:49:45 am »

watches01 · « **Reply #19 on:** March 03, 2018, 11:43:23 pm »

Here's the board

watches01 · « **Reply #20 on:** March 03, 2018, 11:44:43 pm »

Here's the schematic

watches01 · « **Reply #21 on:** March 03, 2018, 11:53:50 pm »

I checked that electrolytic and it seems ok. I'll change it anyway, but I don't expect any different result from it. I also checked the .0039 cap in parallel with the transducer and it seems fine. It looks like R5 is a variable resistor in a small blue box on the board, but there is also some type of variable component toward the center of the board. On the schematic it looks like a transformer, but looking at it I see that it is adjustable. Is it possible that R5 adjusts the frequency and the other adjustment might affect the intensity of the ultrasonic? I don't want to adjust it without knowing what it does. I ordered a scope and I'll order the caps this week.

Armadillo · « **Reply #22 on:** March 04, 2018, 12:31:51 am »

You can see the 2 spraque capacitors there. I like spraque, pretty good reliable capacitor.
You should unsolder them from the board and verify the ESR or DF.
BTW, What are you replacing them with?

R5 adjust the frequency.

Armadillo · « **Reply #23 on:** March 04, 2018, 01:07:19 am »

1 of the spraque is the film type capacitor; you should also check it.

watches01 · « **Reply #24 on:** March 04, 2018, 01:52:58 am »

I’m going to see what’s available from Mouser and try to use quality components. That sprague film capacitor tested like it was open while in circuit. I desoldered one side and tested and it seems ok. I have an old Heathkit IT-28 and I checked it for leakage. Seems ok but I’ll change it anyway. Esr was good on both.

watches01 · « **Reply #25 on:** March 04, 2018, 02:18:07 am »

I also found this video and I’m going to see if I can use this method to find the actual resonant frequency. It says in the original manual that the board has to match the transducer and both should be replaced at the same time, so it may not be enough just to adjust to 100khz. I’d like to verify in case it’s off.
https://youtu.be/-E7zlQEk5MA

watches01 · « **Reply #26 on:** March 04, 2018, 02:35:55 am »

By the way on the right side of the board as pictured there is a coil with the core sticking out. That core was rubbing on the steel machine cover. I’ll have to see if I can move the board slightly, I don’t know if I’ll have room.

Armadillo · « **Reply #27 on:** March 04, 2018, 03:11:59 am »

Quote from: watches01 on March 04, 2018, 02:35:55 am

By the way on the right side of the board as pictured there is a coil with the core sticking out. That core was rubbing on the steel machine cover. I’ll have to see if I can move the board slightly, I don’t know if I’ll have room.

Can you post a picture of the underside of the board, so that I can trace and identify the coil.

Of course, the driver board is kind of ridiculous compared to the modern board, but it was matched to the old transducer.

Armadillo · « **Reply #28 on:** March 04, 2018, 03:24:13 am »

Quote from: watches01 on March 04, 2018, 02:18:07 am

, so it may not be enough just to adjust to 100khz.

Refresh memory and the tong meter; Essentially you want energy to be transferred efficiently per manufacturer designed horns/transducer fit up.

Quote from: Armadillo on February 24, 2018, 11:18:41 pm

... adjust trimpot R5 on the circuit board until somewhere around 100KHz, while your tong meter placed on the line of the driver board read maximum peak current [around ~4A peak, my guesstimation].

watches01 · « **Reply #29 on:** March 04, 2018, 04:38:29 am »

Okay, that makes sense. If the current draw is correct it won’t burn up the board. I’ll post a picture of the underside tomorrow from my computer. I have to resize it.

Armadillo · « **Reply #30 on:** March 04, 2018, 07:00:51 am »

Correction, your circuit board shows that the fuse is 2A. Ehm..... lower capacity. [What part number is the transistor used. A closed up photo of the transistor would be helpful].

So, find the peak current at somewhere around 100KHz. Essentially is tuning for best power transfer at around the resonant frequency.

watches01 · « **Reply #31 on:** March 04, 2018, 05:17:21 pm »

The large power transistor is MJ12005. The other transistor is bent over and glued to the backside of the board so I can't read it, but the schematic says it's UMT3584.

Armadillo · « **Reply #32 on:** March 04, 2018, 10:41:49 pm »

Thank you for those nice photos.

MJ12005 is rated at 8A continuous current, much higher than the 2A fuse I would say. Maybe this is a universal board.? There is a series ~~diode~~ resistor just after the fuse to limit the current which is a relief. The isolation transformer should be able to supply at least 300 watts but I suppose the transducer is only less than 100W type.

The ferrite rod you can use a diamond tile cutter to cut it shorter if you want. I suppose that ferrite coil and the 1uf film capacitor form a series crude filter shaper circuit. Careful ferrite is brittle.

R5 has a definite control on the resonant frequency. I would suggest tune R5 for Peak RMS current at around 100KHz and then fine tune the ferrite coil again for improved Peak RMS current. Note that the tong meter cannot response faster than your tuning. So I suggest tune from 80 KHz to 120 Khz in step of 2 KHz and tabulate the results on a piece of paper to graph it.

Those were just my suggestion.

Edit: The isolation transformer refers to an external power isolation transformer to be use to power the driver board up for the test.

watches01 · « **Reply #33 on:** March 05, 2018, 12:56:29 am »

I won’t try to cut the ferrite rod because I would be afraid of breaking it. I’ll try to move the board slightly. I can probably drill the mounting bracket holes a little bigger to get some movement. That coil is not one that is adjustable. The adjustable one is towards the center of the board. It has 4 colored wires coming off of it. That’s a good idea to tabulate the results, I’ll do that. Is there any danger of burning out the board by running too far out of frequency? These boards are known to burn out and I’m wondering what happens to them. Is it possibly too much current draw from bad adjustment? Maybe it’s just a component failure. I’m trying to prevent it, but maybe I can’t.

Armadillo · « **Reply #34 on:** March 05, 2018, 05:14:23 am »

When its out of frequency, the impedance is higher and thus the current is lower. When is at resonant, the impedance is at the lowest and the current is highest. It is actually here that I am a bit concern whether the design was originally catered for at resonant. But the designer did put a series resistor there to limit the current and protect the transistor.
IMO, I do not think it will burn out the board because as you have seen, the transistor is rated at 8A. At most the fuse will blow first, but since this is a "monitored tuning" by the tong meter, it won't be like catastrophic uncontrolled event. The important thing is to fill the tank with water at a level you normally will use to wash the parts and not empty.
Alternatively you can use a variac to power the isolation transformer and at every frequency steps, you slowly adjust the variac voltage up from 50V to 110vac while monitoring the tong meter not to overshoot 2A. For me, I am happy with the fuse protection and the board is so simple it didn't fear me at all. The first time I did micro smd, I was all shaky, but after that I realized that I put too much fear into it. It was all so simple and forgiving. I laugh at it while I crossed the learning hurdle.
That Heathkit IT-28 you mentioned "Green eyes" is interesting. You should also check the capacitance while you are at it. Interesting it can charge and check the leakage though I don't know how low is low leakage it can measure. I have no experience with the Heathkit. I think the bidding price is increasing after you mentioned it here.
So, we look forward to after your components have been replaced, hopefully is the component failure like you said.
Regards;

Edit: The components are aged and may not be able to be subjected to off-frequency potential. But don't you think is a good test to eliminate failing component also? Do cater for it nevertheless.

watches01 · « **Reply #35 on:** March 05, 2018, 04:05:53 pm »

Everything has been ordered, the caps and test equipment should be here this week. When my other machine went bad I remember it smoked and burnt components but it was 10 or 15 years ago so I don't remember exactly what. I'm wondering if I should change all the resistors now just for dependability. I just repaired a timing machine that had a burnt resistor and some bad caps. Luckily it didn't damage anything.

watches01 · « **Reply #36 on:** March 06, 2018, 12:52:10 am »

I think I just found a problem. I was checking the resistors and there is one that I believe is R10 on the schematic. It's kind of hard to tell the third color, but if it is R10 it should be brown, that would be orange, black, brown, gold which would be 300 ohms +/- 5% like the schematic says. It measures 2.9 ohms in circuit. It's in series with R9 which does measure 2.7 ohms as it should. If the third band was gold it would be a 3 ohm resistor, but it's definitely not the same color metallic gold as the tolerance band. I also found a brown, green, orange, gold resistor, which should be 15kohms, that measures 11.5kohms. I don't see a 15k resistor on the schematic. maybe this is a revised board, I don't know. I'll have to try to trace these componenets. Either way there is something wrong with the resistor that I think mis R10. I'm attaching a picture of it.

3db · « **Reply #37 on:** March 06, 2018, 01:27:30 am »

I think you should lift one end of the resistor and measure its value.

3DB

Armadillo · « **Reply #38 on:** March 06, 2018, 01:41:22 am »

Agree.
That's because the resistor could be in parallel to other components.
In case you are measuring high ohm resistor, make sure your hands are not in contact with the probe, otherwise it will measure your body resistance as well.

watches01 · « **Reply #39 on:** March 06, 2018, 01:42:31 am »

You’re right. 278ohms.

watches01 · « **Reply #40 on:** March 20, 2018, 01:24:24 am »

I changed all the capacitors and I did change that one resistor because it was slightly out of tolerance. I put the board back in and ran the machine to check it. It was still weak. I used a multimeter to check the frequency and it said 41 kHz. I disconnected the transducer and using the method in the video above I determined the resonant frequency to be 84.4 kHz. I connected the oscilloscope and tried turning R5 but it didn’t seem to make any difference, so I disconnected terminal four and I connected the oscilloscope to terminal four without the transducer connected. The board made a buzzing sound like a transducer and I was afraid it might burn up but I found that turning R5 did adjust the frequency now. I adjusted it to 84 kHz and then reconnected the transducer. It seems to work fine now and I think it’s adjusted correctly but that second adjustment in the center of the board doesn’t seem to make any difference. I thought maybe it would vary the duty cycle but it doesn’t seem to. It looks to be running that about a 50% duty cycle and the board is drawling .82A. I’m very happy with the outcome and I very much appreciate the help!


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Matching ultrasonic transducer (Read 6191 times)

Share me