High Pitch Squeal from Inductor/Capacitor

Thread Starter

bwilliams60

Joined Nov 18, 2012
1,442
I have a question about an ECM I am working on. The unit took on some water and as you can see from the picture, it corroded a terminal at the end of a very large capacitor. I have this board repaired but now when I turn it on, there is a constant high pitched squeal coming from the large inductor/capacitor area and I can't put my finger on it. The connection point is also the hottest point on the board at around 155 degrees while the rest of the board idles around at about 70 degrees. Does this sound normal or do I need to go do some more digging. I have no schematics to go by. I am sure this is part of the power supply for this board so would it generate a lot of heat? Does this type of cap have reason to squeal?
The repair has been done properly. I laid a large copper contact over the area, soldered it to the capacitor and both legs coming into it and epoxied it to the board. There are no more codes being set by this unit so that part is cured. The code it had before was a power issue.
The large green cap is 14uF and the large black unit is the inductor.
Cheers
 

Attachments

Alec_t

Joined Sep 17, 2013
14,280
14uF seems a very strange value and also very low if it's a reservoir cap. Are you sure of the value?
155 F or C?
Is the cap a low ESR type? It's probably gone high ESR so overheats. Replace with a new low ESR 105°C rated one.
I can't put my finger on it
I'm not surprised if the temperature is 155 :D.
 

ebp

Joined Feb 8, 2018
2,332
ECM meaning ??

The long thin tabs visible on the back of the board are the capacitor terminals, yes?

At that capacitance it is going to be a film type, and it is probably handling a very large amount of current, possibly tens of amps. Are there heavy foils to the capacitor on the top of the PCB, or if not, is it a multi-layer PCB? The tracks on the back make no sense for a capacitor of that type.

Is the capacitor in parallel with the inductor?

If the high temperature is right at the terminal I suspect there is likely still some connection issue.

Are there other visible markings on the cap? I'm assuming you're saying 14 µF based on markings, not measurement.
 

Thread Starter

bwilliams60

Joined Nov 18, 2012
1,442
ECM meaning engine control module or electronic control module, depending on who you talk to.
The long thin terminals are the capacitor terminals. The capacitor is marked MKT1822-614/515-1 14uF +/- 10% 150V
https://www.vishay.com/docs/26012/mkt1822.pdf

The inductor is marked 2541559M01 and is also very hot at the same time. The hottest spot seems to be the terminal that I repaired and yet, I know that the solder joint is 100%. The only question I might have is if there is a mid layer in the board which I am not connecting to. Otherwise, I am sure this is a good connection.
Any thoughts on the squealing noise?
The temperature is in Fahrenheit.
 

spinnaker

Joined Oct 29, 2009
7,830
See if you can pick up some freeze spray. There are videos on line that show how to use canned air its place. Try spraying the component that you thing is causing the squealing. If it quiets, you found it. Try to target the component itself to make sure that is the issue.

You can then move on to your joint that you think is 100%.
 

ebp

Joined Feb 8, 2018
2,332
What does this module, or the system it belongs to, actually do?
Is the system working OK except for the noise?

MKT
is a standard prefix for metallized polyester film. It isn't the type that is normally used for high current at high frequency. MKT are sort of the most general purpose and lowest cost of film capacitors.
The data sheet says "vacuum deposited aluminum" and "extended metallized film" which means that the metallization on the polyester goes right to the edge and aluminum is deposited on each end of the winding to make the connections for the leads. Sometimes there can be problems with those connections, which can lead to heating which can cause further damage. I doubt if that is the case here, but it isn't impossible. Roederstein is an old company that was making high quality capacitors long before Vishay acquired the company. I don't see the wide lead type in the data sheet you found. It may be an older type. I have Roederstein data going back quite a long time - I'll see if I can find it there, but I may not get to it til tomorrow. It won't really help us much, anyway.
It does seem odd to me that it would make noise - at least anything easily audible. Usually that type of cap is made by winding long strips of the metallized film, which makes the physical structure pretty tight.

The inductor part number looks like just a house number to me. It is very common for manufacturers to have inductors custom made to specification. I don't see anything in the number that suggests value, though the "254" might mean 25 x 10^4 nanohenries or 250 µH, which is not be beyond the realm of possibility.

There is a possibility that something is still wrong and the circuit is operating at the wrong frequency or is unstable. This can cause something that normally works at ultrasonic frequency to have sonic-range frequencies sort of superimposed on the higher frequency. I've certainly heard lots of switch mode power supplies "sing" in the audible range when the intended frequency is up in the 50-100 kHz range. If you have an oscilloscope, you might be able to pick up enough to get some idea of what is going on by just holding a probe close to the inductor or capacitor. Sounds weird, but it does actually work quite well sometimes to get a general idea of what's up.

Can you tell if the body of the capacitor is hot at the end with the hot lead? The plastic case makes the finger method somewhat dubious & thermocouple probes are inclined to pick up noise. Any chance you have some Tempilstiks that might help? I'm not sure we'd learn much, since if the terminal on the cap is copper foil it'll be hard to tell which end sources the heat.

It's an interesting board. I can't recall seeing another where the power semi's were just glued to the heatsinks.

I'm sorry, I know none of this is of any help. Without knowing more about the what the circuit is supposed to do, I doubt if I'll be able to help in any useful way.
 

Thread Starter

bwilliams60

Joined Nov 18, 2012
1,442
This ECM is from a Cummins Diesel engine. It is a CELECT PLUS ECM and came in because it had a code indicating that the power supply was not functioning properly. I took it apart and found water intrusion on this corner of the board. I removed the cap and the inductor, repaired the board and put it back together. When it heated up, I replaced those two components with parts from a donor. The same result. So now I am back to this connection/capacitor/inductor and know something isn't right and can't put my finger on it. I am wondering if there is a middle layer in this board and it is not connecting to the cap. I may remove it tomorrow and see if I can find another trace that I may have missed. I can only see the two and I know I have a good connection to both as well as the cap. Scratching my head on this one.
 

Thread Starter

bwilliams60

Joined Nov 18, 2012
1,442
I got the temperature down to about 120-125 degrees F. Not sure if that is acceptable or not. The high pitch squeal seems to almost match my tinitus so hard to pick up on but it is still there slightly. Rejigged the connection and used a better material. Apparently I grabbed a piece of brass by mistake. Replacing it with copper seems to be conducting better.
 

debe

Joined Sep 21, 2010
1,389
Whilst im not familiar with Cummins, it looks like a high voltage inverter. They make squealing noises when there is a short or overload on the output. Ford diesels use an inverter to step 12V dc up to 110V dc for activating the fuel injectors. This may be a similar setup.
 

Thread Starter

bwilliams60

Joined Nov 18, 2012
1,442
Good point Debe. You may be on to something. What you say makes sense. I have an older unit here and that voltage is stepped up to 100 volts while this one sits at 13 VDC. Not sure though as the other one has a transformer and a Power transistor right near this setup.
I took the unit apart again and went back to work on the area surrounding the connection as it seemed to be the hottest point. The board is really well masked but under a bright light I thought I saw another trace so I started digging in to the board and lo and behold, I found another trace. Now I am going to excavate off to the side and see if there is anything else I might be missing.
 

ebp

Joined Feb 8, 2018
2,332
Knowing there is a need for high voltage for injectors is helpful. I was trying to imagine what a single circuit like that would be doing on an engine control module.

A good design won't squeal in pain when shorted or overloaded, it will limit current or shut down gracefully. Sometimes you'll get a tick-tick-tick as the circuit tries to start, goes into overload, shuts down & tries again. However, sometimes the overload circuit isn't all that well designed and squealing is a possibility.

If the circuit is supposed to produce some voltage higher than the battery voltage and the inductor just has a single winding, it will almost certainly be a simple boost converter. In a boost converter, if the boosting is not working the output will be "a diode drop" lower than the input voltage. Depending on the diode type and the load, this voltage will typically be in the range of about half a volt to a bit over a volt. If the boost is working a little but keeps shutting down, the voltage may be a little higher - this really requires an oscilloscope.

In a boost circuit, if the output voltage is nearly exactly equal to the input voltage, it means the diode is probably shorted, however this almost always means a transistor will also have died because of the way a boost converter works.

A single shorted turn on the inductor will cause all sorts of problems.

===
The change with using copper instead of brass is surprising. Brass is a much poorer conductor than copper, but it isn't like the current had a long way to go. The copper on the board is unlikely to be more than 0.003" thick (this would be a little more than "2 ounce" copper, meaning that it weighs two ounces per square foot), so unless you used really thin brass shim stock or the like, your repair probably had lower resistance than the original connection.
 

Thread Starter

bwilliams60

Joined Nov 18, 2012
1,442
Thanks EBP for shedding some light on this and making sense of the voltages I seen. I can understand a lot of things but still a large learning curve away from understanding everything.
I am going to dig into this board and see just how damaged it really is. Good lesson for me.

This may be a dumb question but any idea what an inverter circuit might look like in this unit?
 
Last edited:

PAFMC01

Joined Mar 20, 2013
12
This ECM is from a Cummins Diesel engine. It is a CELECT PLUS ECM and came in because it had a code indicating that the power supply was not functioning properly. I took it apart and found water intrusion on this corner of the board. I removed the cap and the inductor, repaired the board and put it back together. When it heated up, I replaced those two components with parts from a donor. The same result. So now I am back to this connection/capacitor/inductor and know something isn't right and can't put my finger on it. I am wondering if there is a middle layer in this board and it is not connecting to the cap. I may remove it tomorrow and see if I can find another trace that I may have missed. I can only see the two and I know I have a good connection to both as well as the cap. Scratching my head on this one.
Not likely to have a middle layer...such boards cost more and auto manufacturers are dead keen on cost minimisation. If it has a processor on board then it may well have a multilayer board but not otherwise. That's not definite - just probable.

Noise is likely coming from the inductor. Try putting a hollow plastic tube onto it and the other end near (NOT in) your ear - like a stethoscope.

The board looks nasty in many spots - lots of gunk or corrosion on it. Checked all suspect looking tracks for continuity?

I'm afraid what with the soaking and corrosion, and even the gunk (if its conductive), the fault may not be near the L or C. Overheating implies too much current. Measure the input current, then try to trace where it is going. Could be a transistor blown by the original fault (for example the capacitor which you replaced). Inductor suggests switchmode power supply, which implies power transistor and schottky diode for damping. Look around those components.

What is it supposed to do exactly? Seems like it could be injector control amongst other things. Injectors = solenoids. Solenoids= high current switching and kickback voltages. Kickback voltages= diodes.

edit: oops - looks like others have covered this possibility already...

No error codes implies the rest is fine but something is awry in the output stages.

These are all just guesses and things to try. Sadly this is not a fault that can be readily diagnosed by remote.
 

Thread Starter

bwilliams60

Joined Nov 18, 2012
1,442
There definitely is a middle layer to this board. The dark area seems to be connected directly to +12VDC. If I am looking at this correctly, only the top and bottom layer are connected to the capacitor and the 12V layer skirts past the capacitor opening but does not connect to it. I guess this would make sense since another board I have, has 100VDC on it when it is running. They are not identical boards but from the same family. Here are some pics outlining the damage from oxidation.(water) I excavated down to uncover the middle layer.
 

Attachments

shortbus

Joined Sep 30, 2009
10,045
Really have no clue and just a WAG. Is it possible when you resoldered the components solder flowed into the middle layer of the board? Causing a short? I'm assuming you used extra flux(liquid flux) to get a good joint due to the corrosion present.
 

takao21203

Joined Apr 28, 2012
3,702
There definitely is a middle layer to this board. The dark area seems to be connected directly to +12VDC. If I am looking at this correctly, only the top and bottom layer are connected to the capacitor and the 12V layer skirts past the capacitor opening but does not connect to it. I guess this would make sense since another board I have, has 100VDC on it when it is running. They are not identical boards but from the same family. Here are some pics outlining the damage from oxidation.(water) I excavated down to uncover the middle layer.
When a board looks like this, definitely a write off, especially when it is a multilayer.
Repairing 4 layer is not a good thing all too easily the vias are destroyed or become unreliable.

If the board had the moisture for weeks or months there might be all kinds of corrosion.
 

takao21203

Joined Apr 28, 2012
3,702
I have a question about an ECM I am working on. The unit took on some water and as you can see from the picture, it corroded a terminal at the end of a very large capacitor. I have this board repaired but now when I turn it on, there is a constant high pitched squeal coming from the large inductor/capacitor area and I can't put my finger on it. The connection point is also the hottest point on the board at around 155 degrees while the rest of the board idles around at about 70 degrees. Does this sound normal or do I need to go do some more digging. I have no schematics to go by. I am sure this is part of the power supply for this board so would it generate a lot of heat? Does this type of cap have reason to squeal?
The repair has been done properly. I laid a large copper contact over the area, soldered it to the capacitor and both legs coming into it and epoxied it to the board. There are no more codes being set by this unit so that part is cured. The code it had before was a power issue.
The large green cap is 14uF and the large black unit is the inductor.
Cheers
Looking at 1) 80% chance its a multilayer
Looking at 2) too much corrosion damage, write off
Looking at 3) there might be more issues like this one, and they may manifest after some additional weeks months and require a repair again.
 

Thread Starter

bwilliams60

Joined Nov 18, 2012
1,442
Shortbus I guess the possibility exists in your train of thought but I would think it would have to travel far in the layer to actually create the short.
takao21203, I agree this is going to be a difficult repair but I have nothing to lose. Like you say, it is a write-off but I am having a hard time finding a replacement for this customer so I am going to see what I can do while I look for another one. Thanks everyone for your feeedback. all good., keep it coming.
 
Top