Through-hole MLCC under the skin

[GGBB]

I recently pulled a through-hole MLCC which in so doing fell apart. One leg came off and part of the ceramic(?) coating broke away from the body, which partly revealed what looked like an SMT capacitor. So I snipped away the rest of the cover:




It sure looks like this is just an SMD with leads tacked on and dipped in ceramic - could that be the case? Thinking about it, I suppose that makes perfect sense from a manufacturing perspective. As you can sort of see in the pics, the one lead came clean off. There's a slight concave-like ridge formation left that the lead used to lie in (which the picture doesn't show very well). How are the leads actually attached - solder or some special weld? What keeps the leads from coming off when you solder them into your circuit? Inquiring minds want to know.

[Ice-9]

Hmm

[duck_arse]

it's the other way around - surface mount is thru hole without leads fitted. seek out some datasheets/product info sheets, like from muRata or some good brand, they will have all sorts of interesting information, like how a plcc is made.

[GGBB]

it's the other way around - surface mount is thru hole without leads fitted

Tomayto tomahto?

[vigilante397]

How are the leads actually attached - solder or some special weld? What keeps the leads from coming off when you solder them into your circuit? Inquiring minds want to know.

Leads are soldered on, most likely lead-free solder. The conformal coating is what keeps them from coming off when you're soldering the cap into the circuit. It also likely adds some insulation so the solder connecting the leads to the capacitor don't get as hot as the area the iron is actually touching. If you wore off the coating around the lead and held your iron on it long enough I would expect it would come right off.

[Rob Strand]

Those through-hole MLCCs were fairly wide-spread by the early 90's.   The interleaved layers was done for sure.  However I can't remember what was inside them in that era.   So I don't know if they were produced as SMD with legs added.

While surface mount technology was known for a *long time* I don't think manufacturers started putting out parts as we know them and people using them on boards until after the mid 80's.  With technology changes there's obviously a wide crossover region between the first time these things appeared and when people started using them.

Anyway, the dates could imply through-hole MLCCs arose from SMD developments.

(This is all IMHO from what I saw in the industry.  I haven't looked up any datasheets as evidence.)

[anotherjim]

The first time I worked on SMD "chip" components was some car stereos. There was an epidemic failure of R's & C's cracking due to the flexing of the pcb bouncing when the vehicle was in motion. Of course, no legs absorbing the movement. That doesn't seem to happen anymore. Stiffer boards? Stronger parts?

[Rob Strand]

That doesn't seem to happen anymore. Stiffer boards? Stronger parts?

I still see a lot written about the problems but I don't remember the solutions.

I wonder if it's because most boards are smaller now and the deflection is less.   (I'm trying to avoid the urge to analyse a large board and a small board to see if the deflection *is* actually different over the length of a component.)

[Peter Snowberg]

I designed a product a few years ago which was a couple connectors, an SOT23 voltage regulator, an SOT23 thermal sensor, an SOT23-6 PIC, and a couple of decoupling caps all on a piece of 0.031" PCB the same size as an TO220. They were assembled in panels of 50 and at first the V scores were just fine. As time went on the PCB fab got sloppy and the V scores got shallow so the decoupling caps started to fail as the boards got bent during depanelization. Some units would show a dead short from Vdd to Vss but others would show a short of anything from a few ohms to a few hundred K. I had to redesign the test jig to contain ohm meter functionality to allow the caps to charge and then quickly test them for continuity or hopefully a complete lack of continuity across the rails. From that point forward I will never trust MLCCs anywhere they might encounter any mechanical stress. The thin board was just murder on those things.

https://www.bhphotovideo.com/c/product/1212973-REG/owc_other_world_computing_owcdidimachdd09_in_line_digital_thermal_sensor.html

[Rob Strand]

I designed a product a few years ago which was a couple connectors, an SOT23 voltage regulator, an SOT23 thermal sensor, an SOT23-6 PIC, and a couple of decoupling caps all on a piece of 0.031" PCB the same size as an TO220. They were assembled in panels of 50 and at first the V scores were just fine. As time went on the PCB fab got sloppy and the V scores got shallow so the decoupling caps started to fail as the boards got bent during depanelization.

I guess cracking the V-cuts is putting a lot more stress on the PCB than in normal use, especially if it's not done correctly.   I have done boards with V-score but I've had decent margins around the boards.  For larger boards I had a routed boarder with small connection points around the rim which are easy to crack off.

There's a whole heap of guidelines for this stuff.
eg.
https://www.eurocircuits.com/wp-content/uploads/ec2015/ecImage/pages/9-pcb-panel-guidelines/ec-panel-guidelines-english-1-2010-v3.pdf

Search:
guidelines for panelize smd boards