Author Topic: When is a uF not a uF ?  (Read 2717 times)

0 Members and 1 Guest are viewing this topic.

Offline NY2KWTopic starter

  • Regular Contributor
  • *
  • Posts: 163
  • Country: us
When is a uF not a uF ?
« on: October 03, 2019, 10:30:34 pm »
I need to replace a tantalum in an old 555 switching supply.  I read that tantalum capacitors have a higher volumetric efficiency (CV/cc) when compared to other types of capacitors and you would need a 100uF aluminum electrolytic capacitor to replace a 10uF tantalum capacitor.  I checked both specs and a good low ESR aluminum has the same ESR about 2-3 ohms, so why do I have to go up in capacitance value?  The 555 switching supply is fed by 14vdc and outputs 33vdc (1v ripple) and uses a zener to drop to desired 24v.  The tantalum is rated at 35v right near its operating voltage so I would like to replace with a 4.7uF aluminum
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: When is a uF not a uF ?
« Reply #1 on: October 03, 2019, 10:36:57 pm »
Good grief, this is a double-bad: never expose tantalum caps to high ripple current (ie - switchers) and don't run them anywhere close to rated voltage. And note that the most common failure mode of sintered tantalum (the only kind you are likely to ever see) is to explode then catch on fire.

Am I being hyperbolic? A bit, yes, but tantalum really shouldn't be used where high AC/ripple currents are expected and it definitely benefits from generous derating of voltage.

 
The following users thanked this post: amyk, tooki

Online langwadt

  • Super Contributor
  • ***
  • Posts: 4857
  • Country: dk
Re: When is a uF not a uF ?
« Reply #2 on: October 03, 2019, 10:53:18 pm »
Good grief, this is a double-bad: never expose tantalum caps to high ripple current (ie - switchers) and don't run them anywhere close to rated voltage. And note that the most common failure mode of sintered tantalum (the only kind you are likely to ever see) is to explode then catch on fire.

Am I being hyperbolic? A bit, yes, but tantalum really shouldn't be used where high AC/ripple currents are expected and it definitely benefits from generous derating of voltage.

I have an old pair glasses with chip right in the front of one eye from an exploding tantalum capacitor
 
The following users thanked this post: MagicSmoker, tooki, newbrain

Offline tautech

  • Super Contributor
  • ***
  • Posts: 29810
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: When is a uF not a uF ?
« Reply #3 on: October 03, 2019, 11:02:09 pm »
I need to replace a tantalum in an old 555 switching supply.  I read that tantalum capacitors have a higher volumetric efficiency (CV/cc) when compared to other types of capacitors and you would need a 100uF aluminum electrolytic capacitor to replace a 10uF tantalum capacitor.  I checked both specs and a good low ESR aluminum has the same ESR about 2-3 ohms, so why do I have to go up in capacitance value?  The 555 switching supply is fed by 14vdc and outputs 33vdc (1v ripple) and uses a zener to drop to desired 24v.  The tantalum is rated at 35v right near its operating voltage so I would like to replace with a 4.7uF aluminum
Who knows unless we can examine the schematic.

If it's used for timing purposes you wouldn't want to change the value and if for local bulk capacitance replacement with just a higher voltage rating would be a satisfactory improvement.
An ali cap can be subbed in too but the designer used tant for some reason, most likely speed or size.
Avid Rabid Hobbyist.
Some stuff seen @ Siglent HQ cannot be shared.
 

Offline NY2KWTopic starter

  • Regular Contributor
  • *
  • Posts: 163
  • Country: us
Re: When is a uF not a uF ?
« Reply #4 on: October 03, 2019, 11:34:04 pm »
The schematic is attached.  The switching supply is at about 30kHz and the specs (Rockwell Collins) calls for a 4.7uF @ 35v tantalum even though the switching voltage output to the cap is 33vdc.   Can you explain what other characteristic besides ESR makes the tantalum required.  Thanks in advance
 

Offline NY2KWTopic starter

  • Regular Contributor
  • *
  • Posts: 163
  • Country: us
Re: When is a uF not a uF ?
« Reply #5 on: October 03, 2019, 11:36:54 pm »
The cap in question is C5, then(+) side feeds  two zeners, a 30v first and then a 24v for the final output - both rated at only 35vdc.
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 29810
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: When is a uF not a uF ?
« Reply #6 on: October 04, 2019, 12:12:01 am »
There should be no need to replace it unless it isn't doing the job.
If you must then stick a 50V one in.

It's the first in a traditional CLC filter and a tant would've been used for speed.
Avid Rabid Hobbyist.
Some stuff seen @ Siglent HQ cannot be shared.
 

Online T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: When is a uF not a uF ?
« Reply #7 on: October 04, 2019, 12:47:29 am »
Not so big a deal, as awful as this is (yikes, a 555 in a Tektronix?  An open loop switcher at that?), it's not huge, and an adequately sized tant will do.

The caps after the resistor don't see much ripple so it doesn't matter much; C4 is the critical one, seeing full switching ripple, and also being a smaller value it could stand to be a lower ESR type.  I don't recall that there's much selection in dipped leaded tantalum, so you might have to go with another type if you do need to replace that one.

Agreed, 50V is better than 35V rating.

The modern substitute would be ceramic or aluminum polymer.  And the ESR doesn't seem to matter here, as there's no compensation loop to be weary of...

Electrolytics are even available in good ESRs at that value, but I wouldn't recommend them overall.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline NY2KWTopic starter

  • Regular Contributor
  • *
  • Posts: 163
  • Country: us
Re: When is a uF not a uF ?
« Reply #8 on: October 04, 2019, 01:27:34 am »
This is all on a small pcb 24v supply for a oven XO in a 40 year old transceiver...  The cap has shorted (finally) and I have a 4.7uF 50v tantalum on order but if I were to replace it - the adage I have heard is that the replacement should be multiples higher capacitance.  That was the basis for the subject of this thread... with ESR being equal why is a higher capacitance recommended when replacing tantalum caps?
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 29810
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: When is a uF not a uF ?
« Reply #9 on: October 04, 2019, 02:13:40 am »
This is all on a small pcb 24v supply for a oven XO in a 40 year old transceiver...  The cap has shorted (finally) and I have a 4.7uF 50v tantalum on order but if I were to replace it - the adage I have heard is that the replacement should be multiples higher capacitance.  That was the basis for the subject of this thread... with ESR being equal why is a higher capacitance recommended when replacing tantalum caps?
Simply, because tants are fast you need a much larger capacitance to absorb the same spikes that a tant can.
However the modern electrolytic is a more advanced device that designers of yesteryear had access to so when tants became mainstream their advantages were seized upon for better performance and better physical density. What wasn't fully understood back then was their susceptibility to failure when used close to the max voltage rating.
Typically if one was to insist not to use tants, a low ESR bypassed with a ceramic will effectively offer the same benefits as a tant despite growing the BOM count.
Avid Rabid Hobbyist.
Some stuff seen @ Siglent HQ cannot be shared.
 

Offline mzzj

  • Super Contributor
  • ***
  • Posts: 1298
  • Country: fi
Re: When is a uF not a uF ?
« Reply #10 on: October 04, 2019, 05:32:14 am »


Simply, because tants are fast you need a much larger capacitance to absorb the same spikes that a tant can.

Define "speed"

I see tantalum drop-ins more like fashion during certain period..

compared to electrolytics:
+more stable ESR vs aging vs. temp?. Not necessarily any lower ESR than al-electrolytic and the ones that I managed to find from my parts drawer actually had higher ESR than bog standard al-electrolytic!
+lower leakage? (haven't actually measured and historic specs of tantalum drop-in ignitors are scarce so this is more of hear-say)
+"longer life" (but NASTY failure mode)
+possibly higher capacitance in same size... again very much debatable what benefit you have left after you derate the tantalums below volcano-mode. 
 

Offline tautech

  • Super Contributor
  • ***
  • Posts: 29810
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: When is a uF not a uF ?
« Reply #11 on: October 04, 2019, 06:47:55 am »


Simply, because tants are fast you need a much larger capacitance to absorb the same spikes that a tant can.

Define "speed"
Comparable to ceramics.

Quote
I see tantalum drop-ins more like fashion during certain period..
You might but they still have a place in designs like todays SMD high value tantalums.


Quote
compared to electrolytics:
+more stable ESR vs aging vs. temp?. Not necessarily any lower ESR than al-electrolytic and the ones that I managed to find from my parts drawer actually had higher ESR than bog standard al-electrolytic!
+lower leakage? (haven't actually measured and historic specs of tantalum drop-in ignitors are scarce so this is more of hear-say)
+"longer life" (but NASTY failure mode)
+possibly higher capacitance in same size... again very much debatable what benefit you have left after you derate the tantalums below volcano-mode.
It might appear you are happy with using part parameters on the edge of reliability whereas once designers were much more conservative.
Tants are a good example where it is wise to overate them and a lesson the industry has learned.

They still have the advantage of offering the designer a single part that can provide both bulk capacitance and bypassing capability. They'll be with us until better technologies are developed.
Avid Rabid Hobbyist.
Some stuff seen @ Siglent HQ cannot be shared.
 

Offline mzzj

  • Super Contributor
  • ***
  • Posts: 1298
  • Country: fi
Re: When is a uF not a uF ?
« Reply #12 on: October 04, 2019, 09:01:55 am »


Simply, because tants are fast you need a much larger capacitance to absorb the same spikes that a tant can.

Define "speed"
Comparable to ceramics.


Bullshit.
I asked you to define "speed" because it doesn't really specify anything.

Tantalums and especially typical "classic" tantalum "drops" are much closer to al-electrolytic than ceramics for "speed"  >:D (high frequency bypass) .
Similar inductance and ballpark same ESR vs. ceramics that have something like 1000x lower ESR. 

Low ESR SMD tantalums are bit different animal but average tant drops are not that special.

http://datasheets.avx.com/tap.pdf  16v 10uF ESR 2.5ohm 1,2uA 5x10mm

vs. al-electrolytic
https://www.nichicon.co.jp/english/products/pdfs/e-upw.pdf 16v 10uF ESR 0.6ohm  4.8uA 5x11mm
or smaller version 16V 22uF ESR 0.95ohm 4x7mm

So al-electrolytic is both smaller size and lower ESR..
Lower leakage tho.
« Last Edit: October 04, 2019, 09:05:42 am by mzzj »
 

Online T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: When is a uF not a uF ?
« Reply #13 on: October 05, 2019, 06:58:34 am »
Heh no, ordinary (MnO2 electrolyte) dry tants have modest ESR.

Speed is:
τ = ESR*C

It's around, ohh, what's a typical figure for tant, 1uF and 2Ω, so 2µs?

Ceramics are easily 1/10th-1/100th of that.

Polymer tants improve this figure a bit, though you might opt for polymer aluminum because of price and conflict minerals.

It's noteworthy that tants of the same value as electrolytics, do tend to outperform them -- the electrolytic's impedance is complex, and rises at low temperature, while the tant is simple and stable.  This made them an excellent, premium-quality alternative back in the day, but we have many more options available today. :-+

Tim
« Last Edit: October 05, 2019, 07:00:33 am by T3sl4co1l »
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: tautech

Offline mzzj

  • Super Contributor
  • ***
  • Posts: 1298
  • Country: fi
Re: When is a uF not a uF ?
« Reply #14 on: October 05, 2019, 08:28:29 am »
Heh no, ordinary (MnO2 electrolyte) dry tants have modest ESR.

Speed is:
τ = ESR*C

It's around, ohh, what's a typical figure for tant, 1uF and 2Ω, so 2µs?

Ceramics are easily 1/10th-1/100th of that.

Polymer tants improve this figure a bit, though you might opt for polymer aluminum because of price and conflict minerals.

It's noteworthy that tants of the same value as electrolytics, do tend to outperform them -- the electrolytic's impedance is complex, and rises at low temperature, while the tant is simple and stable.  This made them an excellent, premium-quality alternative back in the day, but we have many more options available today. :-+

Tim

Typical esr numbers for 1uf tants look much worse than that, ie 10ohms for 1uf 25v drop.
http://datasheets.avx.com/tap.pdf
Old ones I had in junk boxes were about same range or even bit worse.
 

Offline dmills

  • Super Contributor
  • ***
  • Posts: 2093
  • Country: gb
Re: When is a uF not a uF ?
« Reply #15 on: October 06, 2019, 07:04:57 pm »
Tants in power supply applications were very much a thing back in the day before SMPSUs got everywhere (And forced elco manufacturers to start doing some research), and had a bit of a resurgence with the early PNP output LDOs (That elco research gave us parts with ESR that was too LOW!), but I generally feel that there are usually better options today.

Remember, back in the 80s there were not many good options if you wanted more then a ceramic or film could give you (maybe 1uF so so in a reasonably sized package) and needed decent HF behaviour, the electrolytic caps of the time were generally inductive by the time you hit the AM Radio bands, the tant filled this gap. 

One place I still see them is in PLL loop filters and such where the impedance is high and you want a reasonably stable cap that is low leakage and cannot be microphonic, same for reference supply generators and such, but those are high impedance applications.

 

Offline aiq25

  • Regular Contributor
  • *
  • Posts: 243
  • Country: us
Re: When is a uF not a uF ?
« Reply #16 on: October 08, 2019, 04:56:42 pm »
I never used tantalum capacitors in designs but I have worked on automotive projects where during Worst Case Circuit Analysis I had to de-rate electrolytic capacitors by 35-45 % (customer requirement) during lifetime calculations. These were for power supply circuits.
 

Online splin

  • Frequent Contributor
  • **
  • Posts: 999
  • Country: gb
Re: When is a uF not a uF ?
« Reply #17 on: October 10, 2019, 11:46:11 pm »

Tantalums and especially typical "classic" tantalum "drops" are much closer to al-electrolytic than ceramics for "speed"

Nah.  Exploding Tants are much faster than exploding Al electrolytics    >:D
 
The following users thanked this post: mzzj


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf