Low ESR Caps

I’m using an STM32F205 in a project. From the datasheet, it requires 2, 2.2uF low ESR caps on Vcap1 and Vcap2. It says the ESR must be less than 2ohms. But when looking at datasheets for 2.2uF caps, I can’t find the specifics on any of them with regard to ESR. All I find are graphs of other parts.

For example, this part: JMK107BJ225KA-T. The datasheet is here: http://www.yuden.co.jp/productdata/cata … cc01_e.pdf

The datasheet doesn’t have any ESR information for the 225, but it does for the 226 (which is 22uF). And it seems to be that way for almost all caps I’ve looked at. Can anyone recommend a low ESR 2.2uF cap?

Also, I’m trying to minimize board size, so all my passives are 0402. Does package size have any impact on availability of low ESR caps? One datasheet I looked at showed few options in the smaller sizes, and I’d have to go up to a 1206 to get 2.2uF (in their low ESR line–but I still couldn’t find the ESR information on the cap).

Does this device list any recommended parts? Like actual part numbers?

Well, I did find an eval board, and the BOM uses CC1206KKX7R7BB225 (datasheet: http://www.yageo.com/documents/recent/U … -50V_8.pdf). But this is a 1206. But even for this part, the datasheet doesn’t show the ESR for the 2.2uF part. It shows the ESR for a 1206 1uF and 10uF (and both are less than 2ohms). But the 0402 shows an ESR of almost 10ohms at low frequencies.

When they spec these things for low ESR, do they mean across all frequencies (even DC)?

I found a nice forum post on ESR and capacitors [here. It would clear much up on the subject of what low ESR really is.](Low ESR capacitors - Badcaps)

Another good discussion:

http://www.illinoiscapacitor.com/pdf/Pa … iction.pdf

The interesting part from it:

Observe the frequencies at which the ESRs were calculated. It is important to note that ESR will change to a lower value as the frequency increases. So, when specifying a low ESR capacitor with a specific ESR value, it is necessary to specify the frequency of the ESR or else the wrong capacitor may be selected.

This has been precisely the problem. I don’t know the frequency.

These caps look like decoupling caps. They tie to an internal voltage regulator. We may have to cheat the specs a bit. The cap will have to still be 2.2uF, but the ESR will have to be a bit higher that 2 ohms. Given your specifications, [this may be your best bet](http://www.digikey.com/product-detail/en/JMK105BJ225MV-F/587-1453-1-ND/1004084)

Ceramic caps generally have the lowest ESR. For some low-dropout regulators ceramic caps have an ESR that’s TOO low and can cause instability, but the datasheet should make this clear (ie, by specifying a minimum ESR).

In your case, any 2.2uF ceramic cap should be OK - these are just decoupling caps after all.

MichaelN:
Ceramic caps generally have the lowest ESR. For some low-dropout regulators ceramic caps have an ESR that’s TOO low and can cause instability, but the datasheet should make this clear (ie, by specifying a minimum ESR).

Sure. But wouldn't the alternate case be a problem as well? Say too high of an ESR could cause a regulator to become unstable? Since these caps are for the internal regulator, it seems a similar problem.

MichaelN:
In your case, any 2.2uF ceramic cap should be OK - these are just decoupling caps after all.

Well, they are a bit more than just the standard bypass/decoupling caps. At least I would think so. The fact that the datasheet explicitly calls out an ESR of < 2ohms, and are used for the internal regulator.

I found another document from TDK: http://www.tdk.com/pdf/esr_performance.pdf

In it, ESR is defined as: ESR = DF * Xc, where DF is the dissipation factor, and Xc is the reactance. So a small DF is one way to ensure a low ESR.

I found a part that has a DF of .075 max, which at about 2.7kHz drops to 2 ohms or less. So I think it will be fine.

I use Kemet Spice for this sort of thing. Sure I’m not always going to use a Kemet capacitor but it always gives me an idea. For critical designs though, if I can’t simulate it, I don’t use it.

http://www.kemet.com/page/kemsoft

In your case, any 2.2uF ceramic cap should be OK - these are just decoupling caps after all.

Agree with above statement. I have built a board using STM32F100 and used 4.7uF caps everywhere. Works just fine.

One package, one value of cap == low cost. :slight_smile:

If you want can share the schematics with you.

cheers