What happened to ~1F, ~50mR ESR supercaps? I can't find any

[dicethrow]

I'm trying to build a LM3550 (https://www.ti.com/product/LM3550) based LED flash circuit. Note the datasheet contains dates from 2009...
The datasheet requires a small size supercapacitor, of size around 0.5-1 farad (for reasonable flash duration of about 30ms, and reasonable charge time of about 3 seconds), and ~50 mR ESR (to minimise voltage headroom required for ohmic losses at currents around 5A).

However, the referenced supercap from the datasheet (TDK EDLC272020-501-2F-50) is obsolete according to both TDK and digikey, however I can find no non-obsolete part with similar specs.
It seems that supercaps with the desired ESR are now 10 F+. This means that both charging time, and therefore battery charge wastage, would increase in order to achieve desired flash performance.

So where did parts like this go?
Are they all being bought up by renewable energy projects before they hit distributors like digikey?
Did the manufacturers all pursue high-capacity supercaps, as there's no market for low-capacity, low-ESR supercaps?
Are we meant to use other technologies, like lipo batteries or supersized electrolytic capacitors?
Or am I not looking hard enough?

I did manage to find a ~50mF electrolytic capacitor that might just work for my purposes but it's huge and doesn't provide an ESR figure, but I'd like to see if there's a better option.

Thanks for the help!

[bdunham7]

That's a very rare prismatic packaged specialty supercap, quite different from most others.  Just like flash-rated electrolytics, it's a small specialty market and probably their only customers are cell phone manufacturers--and they probably aren't putting that big of a part in a phone.

Here is your part:

https://www.cap-xx.com/wp-content/uploads/2023/11/CAP-XX-DMF1F-Datasheet-V1_9.pdf

But they want $23/ea with a min qty of 3, the price goes down rapidly if you are an actual commercial customer with any volume. 

[Someone]

Noticed the same, perhaps:

There was very poor demand and they never ran the lines again?
or
The reliability didn't meet customers expectations? (I'm seeing very short quoted lifetimes on newer parts)

[coppercone2]

I have some prismatic super caps. They don't last. Like at all.

That part would literally need to be spring loaded with a chassis port to be 'good'.

[dicethrow]

Great thanks!

Regarding their short lifespan, what failure mechanism are you thinking of? Increase in ESR, loss of capacity, or loss of charge?

[coppercone2]

I dunno they don't do shit so I guess ESR. Saddest parts find ever

[Vovk_Z]

IDK, but I guess you may use 'usual' supercap in parallel with e.g. an appropriate polymer capacitor (>= 10 mm diameter). So you'll have ultra low ESR combined with a large capaitance from a supercap. And you may place polymer capacitof closer to the circuit so you'll get even better ESR and ESL.

[dicethrow]

Thanks for the idea, I’ll have a think about how that could work

Alternatively, does anyone see an issue with this: instead of using a supercapacitor, connect the vout pin with a negligible/minimum sized capacitor (4.7uF the datasheet mentions I think), and connect the high side of the LEDs to a high capacity single cell lipo battery instead (ie the high side not connected to the LM3550), and just use the LM3550 as a current source controller for the NMOS. I guess one challenge would be that as the lipo’s voltage reduces with use, ESR voltage drops would mean that flash brightness would reduce too. Although that could be compensated with longer flash duration.

Is there a better chip that can be used for current-pulse control of a NMOS? Honestly I do like that aspect of the LM3550. Perhaps it could be done with an opamp, but the current regulation waveforms in the datasheet look really nice and flat.

edit: yes there are a whole category of chips like this: things like ‘smart switches’ by TI. I’ll have a look for non-supercap approaches now as I’ve been convinced that they’re not a good design approach for this application.
edit: or something like this set for 5A http://www.pcbheaven.com/userpages/
LED_driving_and_controlling_methods/?topic=worklog&p=3, but with a pulse enable circuit.
edit: now I'm leaning toward using an efuse (TPS259470L), set at a 5A current limit, which will effectively act as a 25mR resistor if lipo ESR is dominant (~80mR for example), or as a larger effective resistance if battery ESR is smaller. And no need for an external NMOS and gate driver, this should be able to handle it all in one chip.

anyway I'll keep plodding along, but if you have ideas on a good way to do this let me know.

[strawberry]

flash-rated electrolytics,

arent that they are any better than regular capacitors? not much data about them

[bdunham7]

arent that they are any better than regular capacitors? not much data about them

They're not as common as they used to be, but photoflash applications require very low ESR and ESL.  They were often very long and skinny as a shorter, wider foil has less resistance.

[mariush]

That cap-xx (or a similar one) is available on tme.eu at around 7$ a piece, 6$ if you get 10 (plus vat): https://www.tme.eu/en/details/dmf3z5r5h474m3dta0/supercapacitors/cap-xx/

They also have eaton pvh series , the 1.5F model has 120mOhm esr : https://www.tme.eu/en/details/phv-5r4v155-r/supercapacitors/eaton-electronics/  or  https://www.tme.eu/en/details/phv-5r4h155-r/supercapacitors/eaton-electronics/

the series goes to 5F .. down to 75 and 65 mOhm esr but doesn't seem to be stocked at tme.eu