| Electronics > Projects, Designs, and Technical Stuff |
| Normal electrolytic capacitor in place of low and super low ESR electrolytic cap |
| << < (3/6) > >> |
| David Hess:
It can be done but to attempt to get the same performance, lots of physically smaller capacitors will need to be used in parallel which may not be possible due to available space. You could also get good results using lots of solid tantalum capacitors in parallel because they are better at higher frequencies but this is more expensive. |
| T3sl4co1l:
Had to do that a while ago, I had built another variant of my customizable flyback power supply, this one: but found I only had -- argh, whatever they are, I think 2200uF 35V 85C Samwha (generic type too I think?) in my junk bin. Tried it with two in parallel, initially (the ones by the black vertical choke). Got way too hot for comfort. I put two more in parallel, wired up on stilts, behind them (you can see how they stand taller than the others -- they aren't soldered flush into the PCB!), and that got the temperature more reasonable, though I wouldn't be surprised if the internal (peak core) temperature is still a bit uncomfortable. So in the end, it's way more capacitance than I expected to need, because I used what I had on hand (would've been fine if I felt like ordering parts of course, caps with adequate ratings are easy to find online over here). Consequently, the error amp compensation was way off, which took some adjustment, and I had to add extra aux supply capacitance to ensure startup (because, if there isn't enough hold-up on the primary side, the controller runs out of juice before the output has stabilized, and it sits there ticking away in hiccup mode). It also doesn't really help that the current limit seems to be more like 8A with the exact components in there. So the whole thing got much hotter than I expected. Mainly in the output caps, and the primary snubber. Which I initially had as UF5406, which desoldered itself. That was a fun "clink" sound when I noticed what it happened to... (the supply survived -- extra switching and probably avalanche loss too, didn't instantly explode the transistor, and I noticed the smell. The fallen diode also didn't short anything out, kind of miraculous.) Replaced that with a SiC schottky (the TO-220FP floating in front of the main switch -- yes, it really is tack-soldered onto pigtails there.. :P ), which runs ice cold all the time. That was also around the time I added the fan, to help ensure no more runaway recovery losses, at least until such time as the fan chokes up with dust... (By the way, this is basically a stock UC3842 offline isolated supply circuit; refer to the appnote for a schematic. :) ) Tim |
| The Electrician:
--- Quote from: soldar on December 24, 2018, 07:11:26 pm --- High ESR causes heat which causes higher ESR which causes more heat which causes ... you get the picture. Once the chain reaction gets started it is self-sustaining. --- End quote --- ESR of aluminum electrolytics does not increase with increasing temperature--it decreases. See: https://www.avnet.com/wps/portal/abacus/resources/engineers-insight/article/understanding-esr-in-electrolytic-capacitors under the heading "Constructing the capacitor" is found this: "In aluminum electrolytics, ESR falls as temperature increases - its effects reduce as assemblies warm up." Here: https://en.wikipedia.org/wiki/Electrolytic_capacitor under the heading "ESR and dissipation factor tan δ" is found: "For electrolytic capacitors generally the ESR decreases with increasing frequency and temperature." Here: http://www.cde.com/resources/catalogs/AEappGUIDE.pdf Figure 7 on page 6 shows how ESR decreases with increasing temperature. This paper: http://lipo.ece.wisc.edu/2002pubs/2002_35.pdf in figure 4 gives an example of ESR decreasing with increasing temperature. |
| wraper:
--- Quote from: Miyuki on December 08, 2018, 08:21:23 pm ---Can be replaced with more general purpose ones or help with something like film cap --- End quote --- Nope, it cannot. Film cap will barely take any ripple current on itself and general purpose electrolytic will die within a few weeks or months. |
| wilfred:
I got nearly 6 weeks before failure when I tried normal ESR electrolytics in my PVR's switching power supply. That was pretty much on all the time. The low ESRs I replaced them with have been going over two years now. |
| Navigation |
| Message Index |
| Next page |
| Previous page |