Tantalum axial cap questions.

[xwarp]

Given the fact that it seems that most of the equipment I have worked on, some issues were the result of bad tantalum caps.

Anyone have any opinions on the reliability of Kemet high reliability axial tantalums?

I ask because I have several GM electronic control modules from the late 80's and am concerned about their filtering capability being this old.

I am not sure of the specific brand that these are in the ECM's, but they appear to be very similar to the Kemet high reliability types.

Any thoughts would be appreciated.

[SeanB]

If glass frit sealed they last forever except if reverse voltage applied, when they go bang. Other types with a rubber seal or with a solid dry electrolyte they might give problems.

[xwarp]

A friend of mine that retired from where I currently work had previously worked for GM as an engineer knows a bit about the ECM's and what not.

Waiting on a reply from him to see if he might know a little more about the technical info of the caps that GM used in them.

Here is a picture of a few of the caps.

[SeanB]

C7 is a solid electrolyte one, and C2 is a wet slug type. They do not have any real failure mode other than going short circuit if overvoltage is applied, so if they are working in there leave them. If they fail then replace the solid ones with a 125C electrolytic, preferably a Nichicon or Panasonic one, of a higher voltage rating. Wet slugs you replace with another of the same type. The ECU manufacturer would not place them in there unless nothing else would work, as they are very expensive even in OEM quantities. They likely cost more than the rest of the control semiconductors combined.

Glue the replacements down with an acetoxy free silicone, and reinstate the conformal coat on the board and component when it has cured. As you most likely will be using a radial lead component you will need some thin sleeving to insulate the one lead that you will have to extend to get to the other axial lead hole.

[David Hess]

The wet slug types sometimes leak do to corrosion by the sulfuric acid but that is the only failure I have observed with either of the hermetically sealed types.

[T3sl4co1l]

I would be surprised if wet slug anything ever entered an automotive product.  They're massively expensive and overqualified for mere automotive applications.

For that matter, anything hermetic, but perhaps the choice was made over an equivalent aluminum electrolytic.

FWIW, M39003 is the M-number for dry (MnO2 cathode) tantalums, and M39009 I think is the wet slug version.

Tim

[xwarp]

Appreciate the responses.

As much as I have learned about the GM TBI system due to the fact that about a year and a half ago, I converted my 1985 carbureted truck to TBI.

During that process I learned quite a bit about the TBI system that the ECM's seem to be one of the most reliable components of that system, but had to ask about the caps in it.

[David Hess]

I would be surprised if wet slug anything ever entered an automotive product.  They're massively expensive and overqualified for mere automotive applications.

For that matter, anything hermetic, but perhaps the choice was made over an equivalent aluminum electrolytic.

FWIW, M39003 is the M-number for dry (MnO2 cathode) tantalums, and M39009 I think is the wet slug version.

They might use them for reliability at high temperatures.  Aluminum electrolytics strike me as a particularly poor choice for anything not intended to fail just out of warranty but it would not surprise me if the car companies considered that a feature these days.

I have been replacing dry tantalum capacitors with good aluminum electrolytics of from 2 to 4 times the capacitance to get an equivalent ESR.  Does that seem about right?

What would be an appropriate replacement for a wet tantalum capacitor?  In the circuits where I have run across them, it is not clear why they were selected.

[T3sl4co1l]

I've seen plenty of auto stuff with radials (THT or SMT).  I've heard that they can be preferred for the difference in "flammability".

These days, ceramics are big enough to use in most places, of course.

Tim

[c4757p]

I have been replacing dry tantalum capacitors with good aluminum electrolytics of from 2 to 4 times the capacitance to get an equivalent ESR.  Does that seem about right?

I frequently just use a low-ESR aluminum of the same capacitance. I've been using Nichicon PW series for that lately, they're decent and affordable.

I don't usually use ceramics in arbitrary circuits because of the frequent lack of data on bias coefficient for the through-hole types. I'll use them if I'm sure they're just decoupling a low-voltage rail though.

[David Hess]

I have been replacing dry tantalum capacitors with good aluminum electrolytics of from 2 to 4 times the capacitance to get an equivalent ESR.  Does that seem about right?

I frequently just use a low-ESR aluminum of the same capacitance. I've been using Nichicon PW series for that lately, they're decent and affordable.

The Nichicon PW series is also my first choice.  My reason for using 2 to 4 times the capacitance is that even low impedance aluminum electrolytics do not match solid tantalum capacitors.  Ripple in switching regulators is almost always proportional to ESR rather than capacitance.

I have a local surplus electronics store with a good selection of inexpensive solid tantalum capacitors so I usually find something suitable there for a reasonable cost.  Since all I know is their voltage rating and capacitance, I will check them on my impedance bridge and burn them in for a couple days while monitoring leakage but I have never found a bad one.

I don't usually use ceramics in arbitrary circuits because of the frequent lack of data on bias coefficient for the through-hole types. I'll use them if I'm sure they're just decoupling a low-voltage rail though.

For bulk decoupling, the low ESR of ceramic (and film) capacitors can make frequency compensation difficult.  Some newer integrated regulators are designed to handle low ESR filter capacitors.

[c4757p]

My reason for using 2 to 4 times the capacitance is that even low impedance aluminum electrolytics do not match solid tantalum capacitors.  Ripple in switching regulators is almost always proportional to ESR rather than capacitance.

Yup, I would do this as well in the specific case of tantalum caps in a switching regulator. Luckily I don't see that all too much

For bulk decoupling, the low ESR of ceramic (and film) capacitors can make frequency compensation difficult.  Some newer integrated regulators are designed to handle low ESR filter capacitors.

Yeah, I wouldn't use one directly at the output of a regulator. In my experience, though, the decoupling capacitors spread over a board are typically "far enough" (in terms of impedance) from the regulator not to cause a problem.