Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Electrolytic vs Tantalum at <1V

Status
Not open for further replies.

Speakerguy

Active Member
If I have a 100uF aluminum electrolytic and a 100uF solid tantalum, which will exhibit more capacitance at very low voltage levels? I need a large value capacitor which will maintain its effective capacitance at about 0.7V (it is in parallel with a 1N4148 diode). It will never ever see higher voltage in its lifetime. Thanks!
 
Last edited:
This is the ckt I'm trying to modify, it's a headphone amp:

**broken link removed**

I've changed the topology to inverting so I could eliminate any DC offset at the output, and changed C2L from a 100uF electrolytic to a 3.3uF film cap with an op amp based capacitance multiplier (which simulates the 100uF electrolytic). There will be an inverting stage in front of this so the whole thing will be non-inverting. I'm now wondering if C3L and C4L will be better off being tantalums instead of electrolytics since such low voltage will be across them. The oxide layer in tantalums is permanent, right? I think that would mean it would be much better than electrolytics but wanted to check with the you guys since you are the real ckt experts :)
 
Careful with something like a capacitance multiplier, you multiply error too.
 
The type of capacitors across the diodes doesn't matter. Use electrolytic.
 
Tend to agree with audioguru, Tantalums are good for decoupling where you want low inductance and fast response. In your application at audio frequencies the normal electrolytics would appear to be sufficient.
As a design engineer I have often used Tantalums for decoupling but they are prone to failure at times, normally going short circuit even when very conservatively rated i.e. 16V Tants working on 5V circuits.
Someone once asked me "Why are you using those horrible Tantalums?" but they do have their advantages as well as their bad reputation for failure.
They do not like being reversed in polarity and I think that is the main cause of their failure. With careful design they can behave quite well.
 
Ok. I will do some measurements in the lab when I'm up to it, but it is sounding like electrolytics are the way to go.

With regard to the capacitance multiplier - would you guys use that or would you use a 100uF bipolar electrolytic? The circuit author claims you can get away with a polarized one, and I certainly don't doubt him but I wanted to try and improve the project. But I won't make the changes if you guys say they probably won't be worth it.
 
Top Cap said:
As a design engineer I have often used Tantalums for decoupling but they are prone to failure at times, normally going short circuit even when very conservatively rated i.e. 16V Tants working on 5V circuits.

I'm glad to see someone else has the same experience of tantalums as myself, I've been criticised by some in the past for suggesting tantalums have pathetic reliability!.
 
I must be lucky thus far with my uses of tantalums. I support their use, esp. in audio modulation circuits for 2-way communications.
 
Tantums are reliable under the following conditions:

1) Use a reputable manufacturer (AVX = good, chinese anything = bad)
2) De-rate the part for both voltage and ripple current.
3) Have some way of limiting inrush startup current (tiny amount of R or better still a small amount of L)

You will have a reliable tantulum design.
 
Optikon said:
Tantums are reliable under the following conditions:

1) Use a reputable manufacturer (AVX = good, chinese anything = bad)
2) De-rate the part for both voltage and ripple current.

Tantalums have a more gradual slope of DC voltage vs lifespan. A 16V tantalum is actually only spec'ed to survive for some 1000-5000 hrs or something at 16v at room temp, 16v is a worst-case condition. Higher temps reduce the lifespan too.
There are standard formulas for derating the voltage for the lifespan you require, especially if elevated temps are present:
https://pdf.directindustry.com/pdf/avx/avx-tantalum-niobium-oxide-capacitors-catalog/33763-23877-_170.html
 
This circuit doesn't need to maximize capacitance, rather it needs to make sure that the impedance is lowest when it matters the most.

If you choose electrolytic, its Equivalent Series Resistance at high audio frequencies is the most relevant spec.

Tantalums are usually good at higher frequencies, it seems.

But I would use film capacitors, the largest reasonable value. Or ceramic caps. You might do the math on lower values; those in the circuit may have been chosen for convenience and not performance. Certainly since this is a headphone amp, you need quiet. If I used electrolytics I would have a 1uF ceramic in parallel with them.

[edit] Oh, and the feedback network is arbitrary. If the op amp has a high input impedance you can change it to 220K/10K and decrease C2 by a factor of ten.[\edit]
 
Last edited:
The 100uF seems to have been chosen so that the RC has a corner frequency at something like 1hz, which should mean the electrolytic shouldn't contribute to distortion at high audio frequencies since almost no voltage is developed across it. Or so I have been told. I'm sure 100uF was just pulled out of the air since it could be 50 or 200uF and not matter. The largest reasonably priced film cap I can get is 3.3uF, there are larger but they increase in size and cost a lot.

Would resistor noise be a problem using 220k/10k combination in this ckt? My Sennheisers are crazy efficient.

ETA: I found some 10uF mylars that might work. They are big (22.5mm lead spacing) and 4.58 a piece, but I do only need two, and I don't ever worry about cost on my projects.... With the stock resistor values it puts the ckt -3dB @ 16Hz. Hmm.
 
Last edited:
Would resistor noise be a problem using 220k/10k combination in this ckt? My Sennheisers are crazy efficient.
This amplifier can easily output 20Vp-p, like more than a watt. You could reduce the loop gain (if it would still be stable) and reduce the output noise.

I didn't notice the op amp part number?
 
It's a 5532, a 10MHz part. I was thinking the reason for all the gain was probably stability, he noted in the project article that excessive capacitive coupling between the input and output would cause oscillation.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top