Assembly of ferrite Chokes (to reduce RF)

[stephenweinstein]

We need to put ferrite chokes onto the wires (which are loose, not a cable) connecting a circuit board to a motor and to a battery. We buy the board with the wires already soldered to it and the connectors already crimped to the other end of the wires. The IDs of the one-piece solid chokes are too small to fit the connectors through. The two-piece split chokes (snap-on type, etc.) all are either too large on the outside to fit in our limited space or are too small on the inside for the wires. Aside from unsoldering the wires from the board, putting the one-piece chokes onto the wires, and resoldering the wires to the board, is there a way to make this work?

 

[BrianG]

Presumably you are trying to stop interference from the motor getting into the circuit board? You don't say how this solution was determined, but it sounds like you are trying to thread the wire through a ferrite "bead". As ferrite beads need reasonably close contact with the wire, it would seem that you have no choice but to unsolder the wire, thread the ferrite down it then resolder the wire.

Alternatively, you may get sufficient inductance (if the wire is long enough) by passing the wire around and through a small ferrite toroid. Cores for toroids would hava a larger hole than a bead, so if you pass the connector through the core and back a few times so that you have a few turns as a "coil" this may give the desired inductance.

 

[stephenweinstein]

We are trying to reduce RF emissions to comply with a legal requirement. I think this originate in the board and are radiated by the wires going to the motor and the wires going to the battery. The interference is not affecting the board to our knowledge.

To unsolder and resolder, we would need to train someone and we would need to buy lead-free solder to comply with ROHS.

How large is a toroid? The bead just barely fits in the space under the board, so if a toroid is much larger, we will not have room.

 

[BrianG]

You don't say what kind / size of connector you have to contend with on the lead out wires (my guess is individual blade connectors). "Toroids" come in a whole range of sizes as they are used for winding toroidal transformers and chokes. If you have to fit the choke under the board then it sounds like even a smallish one (with say a quarter inch hole through the middle) will be too large to fit.

Are you sure this is a conducted EMC problem - in which case why doesn't the board manufacturer fix it at source? Presumambly your product has failed an EMC test. You should have been told what kind of emissions the board failed on, i.e. range of frequencies and whether it is a conducted problem or radiated problem or both.

If you can give more spcific information it may be possible to help.

 

[stephenweinstein]

Two are right-angle blade connectors which (because of the right angle) are more difficult to fit through an opening than straight blade connectors.

We considered changing the board but decided against it.

I've seen the test results, including graphs of dB versus frequency, not just the numbers. It is a radiated test, but, since the chokes help, I assume that the problem is emissions (although generated by the board) are radiated by the wires, not directly by the board.

 

[BrianG]

Identifying the source of EMC problems can be difficult to identify: conducted problems in on area can turn in to radiated problems in another part of the equipment. In my experience chokes tend to be better at stopping conducted problems, unless perhaps the wires are of the right length to resonate as an antenna. Unless the board is screened how can you be sure it's the wires causing it to fail a radiated test?

Are you able to say what the product is? Could still do with lots more information:

What is the main frequency radiated?
How long are the wires?
Is the product portable, i.e. hand-held?
EMC test measurement distance?

Possible solutions:
Is the board screened? Could it be placed in a screened box?
Can the wires be bundled together, if so they could be screened with braid?

 

[stephenweinstein]

You have some good ideas, but I am only on this project until the end of this week, and most of them would take longer than that to implement. Also, I have no authorization to pay for more emissions tests, so the solution needs to be an implementation of something that we have already tested. The answers to your questions are:

With chokes on the wires, it passes. With no chokes on the wires, it fails.

Cordless power tool.

Proprietary.
A few inches
Yes.
10 meters.

No. We do not have room to add an additional box; we could add screening to an existing component, but decided not to do so.
Maybe, but we would have to retest it, and I do not have authorization (or time) for that.

 

[BrianG]

OK, so we have a power tool. You say it can pass EMC tests with chokes fitted, so you must have made at least one prototype with suitable chokes. Silly question: why can't you use the same chokes now on the production units? Is it because these were identified too late for the board manufacturer's production run to be included on the lead out wires ? If you want to ensure the tool is EMC compliant then you may have to bite the bullet of post manufacture fix-up costs regardless.

You have not said what sort of frequency or spectrum of EMI is giving the problem other than "proprietary". What does that mean? Is your product remote controlled or something?
EM interference is NOT proprietary!

Can you be sure that it's not the motor producing the problem rather than the board? If your wires from the board (i.e. to the motor?) are only a few inches long, a choke midway would tend to damp out problems from either source.

If the EMI is only present when the motor is operating, one suggestion to try - unless it already has one - is to place a ceramic capacitor of say 1nF to 10nF directly across the motor terminals using the shortest connecting leads possible.

 

[stephenweinstein]

1. The board manufacturer wants too much money to include the chokes in the production units. Also, not all the boards we buy from the manufacturer need the chokes, and for reasons that are not important to this discussion, we are not going to get some boards from the manufacturer with them and some without them.
2. It is not remote controlled. Someone other than me decided that all test data is proprietary. I think it has something to do with the ability of another company to determine what motor we are using and clone our product.
3. Both the motor and the board are producing emissions problems. We are using capacitors to solve the motor problems and chokes to solve the board problems. Neither chokes without capacitors nor capacitors without chokes reduced the emissions enough.

 

[biff44]

Sounds like you are band-aiding the problem with chokes at the ouput power cord. It might be useful to figure out what is actually causing the emissions. If your motor controller is a pulse type, maybe you have too fast of a rise and fall time, and a capacitor to slow the edges placed at the right point in the controller will help.

 

[stephenweinstein]

Multiple persons tried to locate the source and fix the problem and the source. None were able to do so.

You may be right about using a capacitor to slow the edges, but I am only on this project until the end of this week, and it would take longer than that to evalaute your idea. Also, I have no authorization to pay for more emissions tests, so the solution needs to be an implementation of something that we have already tested.

 

[biff44]

What I do in these cases is I have a calibrated antenna and a spectrum analyzer for available lab equipment. Then when I come back from a certified compliance testing facility I can make my own in-house measurements of the problem area, try various fixes, and see if I can make the problem go away. For example, if the compliance testing lab tells me that I have a spurious emission at 29 MHz that exceeds the requirement by 7 dB, I can play around in my own lab until I have that spurious signal reduced by perhaps 12 dB. Then I can take it back to the compliance lab for a final test that I usually pass.

In this manner, I usually only make 1 or 2 trips to the compliance lab, and save bocu $.

 

[IRstuff]

Why can't you get a snap-together toroid?

TTFN

 

[stephenweinstein]

IRstuff:

I tried, but all the ones I found with enough impedance where either bigger on the outside, smaller on the inside, or both. We have very limited space in the assembly.

biff44:

We only sent the product for EMC compliance testing twice; it passed the second time. The problem is now that I need a solution that can be manufactured, without making a third trip (1: failure, 2: pass, but cannot manufacture, 3: can manufacture). We do not have a spectrum analyzer and I did not have authorization to buy one.

 

[BrianG]

I think you must realise now after all our suggestions that this problem should have been sorted out before your boss / company decided to press the "go" button.

If costs have been cut to the last cent so that no time can be allowed to analyse the problem properly, then how does your company do the maths on risk, i.e. possible legal consequences from selling a non-compliant product compared to finding a reliable solution?

Even the few cents more that the board manufacturer wants to charge for including chokes (on all boards) is probably cheaper than the consequences of legal action.