Bead Type NTC Thermistor - Mounting methods for mass production

[Magnethicc]

Hello everyone,
Do you guys have any idea on a good and fast way to mount these type of NTCs (https://product.tdk.com/en/search/sensor/ntc/ntc_element/info?part_no=B57861S0103F045) on to a capacitor?

I am looking for a mass production friendly method.
I am looking for a long term solution - 20 years min.
I was thinking of using 3 rounds of KAPTON tape to hold the thermistor and the capacitor together but I do not know how if the adhesive will hold for 20years.

Looking for advice,
Thank you.

[Jeroen3]

You would want a mechanical method, not a chemical method.
So either some clip or heatshrink.

Onto what type of cap do you want to attach it?

[lowimpedance]

Perhaps the Kapton tape to position the thermister then heatshrink to prevent any de-coupling of the bead from the cap. Assuming its a through hole type.

Picture of the cap to be used would be helpful.

[mawyatt]

A zip wire tie?

Best

[Jeroen3]

I'd avoid zip ties for electrolytic, risk of over-tightening.
Glues may not work on the foil around a cap, tape perhaps if you wrap all around.
But a temperature sensor is probably safety related, having the tape or glue fail is an issue.

Perhaps you can use the white electronics sealant (eg: AS1802), if the thermal properties are sufficient, to glue it between two electrolytic. Sealant also uses it's final shape to stay put.

[PCB.Wiz]

I am looking for a mass production friendly method.
I am looking for a long term solution - 20 years min.

If you want mass production and long term, why not mount a SMD sensor near the capacitors PCB leads - assuming this does PCB mount.
If you must go labour intensive, I'd avoid tape and use a light cable/zip tie, too many 'sticky' items these days simply powder over time.

[jonpaul]

Many packaged and environment protected thermisters  from Omega, Fischer Scientific.

Jon

[Magnethicc]

We though about using heatshrink but the possible issue is that since its a big capacitor (120uF, 550V) a big shrink tube is necessary and big shrink tubes needs alot of heat which we figured as a risk to damage some capacitors due to overheating by the manufacturing house - can someone comment on that?

I agree that zip-ties pose a risk of over-tightning.

SMD assembly was ruled out due to high voltage present near the capacitor we want to measure the temperature of (750V near the + Terminal of the capacitor, 500V near the - Terminal - referenced to the measurement ground).

We have made 4 samples of a capacitor with the bead sensor wraped around with 3 turns of KAPTON tape and that went into a Thermal Shock Test for accelerated lifetime simulation I will report the findings when the test will be done.
We will also do humidity tests to see if humidity will have any bad effects on the KAPTON tape.

[LaserSteve]

We like "cold shrink" at my current employer.   Once you remove the string it holds on and is not coming off.  Cold shrink the leads, not the sensor, to the cap. 

My previous employer had a Lady whose sole job was to solder and shrink leaded thermistors all day for a long-term sensor. Solder to two pieces of stranded #22 mil spec wire.  Heat shrink one leg, then heat shrink over the bead and legs was the most reliable after shock and oven testing.  Test the thermistor after soldering and shrinking, do NOT assume, test for resistance range with meter, not a quick pass fail.   That lady (grandma of 14 kids) loved her repetitive work and took pride in it.  However, failure rates during assembly final test were rather high due to heating inside the thermistor during soldering.

Do not deceive yourself that a slide on connector on thermistor leads will maintain contact for any length of  time.

Client had a metal band designed to have a pressed-in kink in it to hold it to the cylindrical target part. Band had a one-time belt- buckle of sorts to hold it to the part.  We soldered the part under contract, installing was some one else's problem.

Unless you buy a long life Kapton tape, the adhesive will only do 5-6 years at best, if facing temperature cycling. Advise against unless something is clamping / epoxied / holding it. The Kapton needs mechanical constraint if used for any length of time. Kapton works miracles , but is not a long term winner using the provided glue on the tape.

Blob of a high end non-corroding electronics grade silicone is probably your best bet. (DowSil 7091 or a two part Dow)



Steve

[wraper]

You would want a mechanical method, not a chemical method.
So either some clip or heatshrink.

Onto what type of cap do you want to attach it?

All of the mechanical methods will suck ass compared to a simple blob of RTV or 2 component silicone. Both mechanically and heat transfer wise.

[Jeroen3]

@wraper, yeah I mentioned the sealant. Even if it fails chemically it mostly stays in place due to blob shape. But application consistency is tricky!
Plus, I recently opened a bike charger there some white blobs and conformal coating were loose in the box.

A few rounds of kapton tape can work. But I don't think humidity will be your problem. It will be dry heat + vibration that will turn the tape to dust.

[Gyro]

After bending the thermistor leads, GW Instek used a blob of ordinary heatsink compound to couple the temperature sensing thermistors to their linear PSU heatsinks.

I don't know the voltages involved in this case but I might be tempted to interpose a layer of Kapton (in a non-mechanical role) rather than relying just on the thermistor lacquer. Of course if the capacitor is already sleeved...

[jheissjr]

its a big capacitor (120uF, 550V) a big shrink tube is necessary and big shrink tubes needs alot of heat which we figured as a risk to damage some capacitors due to overheating

The sleeve on electrolytics with label marks is a heat shrink material. The question is, is the sleeve shrunk before or after assembling the capacitor.

[Phoenix]

SMD assembly was ruled out due to high voltage present near the capacitor we want to measure the temperature of (750V near the + Terminal of the capacitor, 500V near the - Terminal - referenced to the measurement ground).

TI do a high voltage isolated SMD temperature sensor, might work for you?
https://www.ti.com/product/ISOTMP35

[jbb]

A concerning thought: many electronics have a cross shaped notch in the top of the can. This is an emergency vent, which you probably don’t want to block.

Some of outside-the-box ideas:
- can you build a thermal estimator into the product which can estimate the capacitor temperature (eg using the temperature of a microcontroller and an estimate of the capacitor current) to acceptable accuracy?

- Texas Instruments makes an interesting isolated temperature sensor https://www.ti.com/product/ISOTMP35R-Q1
You could solder this onto the board next to the capacitor legs (someone already suggested, but I wanted to contribute the isolated part)

- I guess this will go into a housing of some kind. Can you fit a spring onto the housing that will gently push the thermistor on?

- could you mount a contactless IR temperature sensor somewhere which points at the capacitor? Dust could be a problem…

[lowimpedance]

We though about using heatshrink but the possible issue is that since its a big capacitor (120uF, 550V) a big shrink tube is necessary and big shrink tubes needs alot of heat which we figured as a risk to damage some capacitors due to overheating by the manufacturing house - can someone comment on that?

I agree that zip-ties pose a risk of over-tightning.

SMD assembly was ruled out due to high voltage present near the capacitor we want to measure the temperature of (750V near the + Terminal of the capacitor, 500V near the - Terminal - referenced to the measurement ground).

We have made 4 samples of a capacitor with the bead sensor wraped around with 3 turns of KAPTON tape and that went into a Thermal Shock Test for accelerated lifetime simulation I will report the findings when the test will be done.
We will also do humidity tests to see if humidity will have any bad effects on the KAPTON tape.

Well that's more relevant information, but a picture of the actual capacitor next to the thermisitor would narrow down a more useful suggestion !
Given the proximity of high voltage I'm leaning more to a combo of RTV and a carefully applied zip tie in case the RTV lets go.
Can you show a picture of the assembly you made?

[PCB.Wiz]

TI do a high voltage isolated SMD temperature sensor, might work for you?
https://www.ti.com/product/ISOTMP35

That's an impressive part, would save a lot of labour and messing about.

[CatalinaWOW]

If this is really mass production (100s of thousands or more) a custom metal clip might make sense.  You likely are looking for a physical restraint for the capacitor anyway.  Biggest risk is dimensional changes in the components over the production life.  But you can design the clip to accept a fairly large changes.  If you are concerned about a conductor you could design the clip in an appropriate plastic.  ABS or Nylon.  You could probably get away with 3D printed prototypes for proof of concept, but would definitely want to go to injection molding for the longer run.

[gspeed]

What about non-contact temperature sensors like MLX90614

[Jeroen3]

There are also thermal jumpers. Basically SMD resistors without metal. With proper calibration these can be used to measure through the capacitor pins if these are high voltage.
This is an option if volumes are very high.