Chucking a tantalum?

slowmotion

slowmotion

New member
So I'm cleaning up an old live desk to see if I can use it for anything.

I've noticed quite a few 10uf tantalum caps inside... and the desk is at least 20 years old, if not 30. I'm going to swap out all the electrolytics as first priority. I also want to ditch all the tantalums due to their age (I understand they "die" like electrolytics after a while...?).

Assuming I can fit them sizewise, can just I replace the tantalums with new low leak electrolytics? Seems to me it should be a straight swap (observing polarity obviously)? This is only for the polarized ones, btw. Everything else is films already (with ceramic decouplers on some op amps I think).

Just thought I'd ask as I don't want to order a lot of them if I'm mistaken. :rolleyes:
 
slowmotion said:
I've noticed quite a few 10uf tantalum caps inside... and the desk is at least 20 years old, if not 30. I'm going to swap out all the electrolytics as first priority. I also want to ditch all the tantalums due to their age (I understand they "die" like electrolytics after a while...?).
Click to expand...

I get the impression they last a lot longer than electrolytic caps, though I suppose if they are old enough, it still might not be a bad idea.


slowmotion said:
Assuming I can fit them sizewise, can just I replace the tantalums with new low leak electrolytics? Seems to me it should be a straight swap (observing polarity obviously)? This is only for the polarized ones, btw. Everything else is films already (with ceramic decouplers on some op amps I think).
Click to expand...

Sure. Assuming you're dealing only with applications where polarized caps are acceptable, electrolytic caps should be fine as alternatives to tantalum. If they're in the audio path (it sounds like they probably aren't), you might also consider film caps if you have room for them. There's nothing wrong with using a non-polarized cap where a polarized cap is specified. The reverse, however, is generally a bad idea. :D
 
dgatwood said:
There's nothing wrong with using a non-polarized cap where a polarized cap is specified. :D
Click to expand...

as long as the voltage rating takes into account the offset voltage plus the signal level....

iirc... tantalums can drift in value over time but usually that associated with higher voltages... ymmv..
 
Tantalum caps also Fail Catistrophicly when even slighly reverse biased and Can also Short when they Fail which can take out other components...Never liked them as they sound brittle in audio applications....

:D
 
dgatwood said:
You need to do that with a polarized cap, too, unless I'm missing something.
Click to expand...


yeah kinda looks like another brain fart but i remember trying it once on a piece in the shop and couldnt figure out why it failed... the boss just chuckled and said to double the voltage rateing... wish i could remember what exactly was the deal...
 
Minion said:
Tantalum caps also Fail Catistrophicly when even slighly reverse biased and Can also Short when they Fail which can take out other components...Never liked them as they sound brittle in audio applications....

:D
Click to expand...


how do you reverse bias a bipolar cap??:eek:
 
Minion said:
Bipolar Tantilum...Never ever heard of one or seen one....
Click to expand...

you had me worried for a second... all i have ever used were bipolar i thought... iirc they were used in the oscillators/generators of some old church organs... but then i checked wiki...


"Tantalum capacitors are a form of electrolytic capacitor. However, some forms of them are non-polar, containing two capacitors connected in series (negative to negative). "

somehow i never equated them with electrolytics at all... well there's my fact for the day...
 
Yes, generally Tantilum caps are Polarized but like Electrolytics that can be connected end to end to make a Non-polar capacitor.....

Here is a exerpt from Wikipedia:

Tantalum: compact, low-voltage devices up to several hundred µF, these have a lower energy density and are more accurate than aluminum electrolytics. Tantalum capacitors are also polarized because of their dissimilar electrodes. The cathode electrode is formed of sintered tantalum grains, with the dielectric electrochemically formed as a thin layer of oxide. The thin layer of oxide and high surface area of the porous sintered material gives this type a very high capacitance per unit volume. The cathode electrode is formed either of a liquid electrolyte connecting the outer can or of a chemically deposited semi-conductive layer of manganese dioxide, which is then connected to an external wire lead. A development of this type replaces the manganese dioxide with a conductive plastic polymer (polypyrrole) that reduces internal resistance and eliminates a self-ignition failure.[5]
Compared to aluminum electrolytics, tantalum capacitors have very stable capacitance, little DC leakage, and very low impedance at high frequencies. However, unlike aluminum electrolytics, they are intolerant of voltage spikes and are destroyed (often exploding violently) if connected in the circuit backwards or exposed to spikes above their voltage rating.


Cheers
 
Minion said:
Tantalum caps also Fail Catistrophicly when even slighly reverse biased and Can also Short when they Fail which can take out other components...Never liked them as they sound brittle in audio applications....

:D
Click to expand...

I'm refusing to take any notice of a person who completely destroys the spelling of 'catastrophically,' only to conclude with "tantalum caps sound brittle in 'audio' applications."
 
Phew thanks guys for all the mixed responses, I'm halfway through the recapping, but apart from some somewhat leaky electrolytics and some very cooked looking tantalums near the power supply (used to be blue, now a very strange green), most everything looks okay (but getting replaced nonetheless!). There is some that I just can't access unfortunately (short of rebuilding the entire thing from scratch) so fingers crossed those of you who think they last okay are right!

Cheers
 
robin watson said:
I'm refusing to take any notice of a person who completely destroys the spelling of 'catastrophically,' only to conclude with "tantalum caps sound brittle in 'audio' applications."
Click to expand...

That's wonderful, but tant caps measure an order of magnitude higher in distortion than alumimum electros . . .
 
robin watson said:
I'm refusing to take any notice of a person who completely destroys the spelling of 'catastrophically,' only to conclude with "tantalum caps sound brittle in 'audio' applications."
Click to expand...

Seing as my posts weren"t meant for you I really couldn"t care less what you think of my "Catastophic" Spelling.....Maybe if you had something usefull to add to this thread your Opinion would mean something...Freaking Troll!!!:mad:
 
Hehe I love forum bitchouts. So I replaced pretty much the rest of them, a few more to go though, powered up the desk and it runs (hooray for me actually getting the polarity correct about 60 times over!)....

Thing is, it's still noisy even with fresh caps (well nearly) all round. Hmm. Might need to look at the op amps next I guess, or the feedback loop resistors as suggested in another thread.

As a side note, my purpose for doing this was to reduce noise. I didn't really achieve that - but I have noticed a big jump in general sound quality. It's still got the hiss (maybe a little less?) but the lows are tighter and it seems generally more pleasant to listen to. A good result, if not the result I intended.
 
Do you have hiss or buzz? The former isn't likely a problem with caps. Anyway, hiss. Check the grounding scheme, but it's probably a function of the ICs selected, the gain staging, and lastly the series resistance. To some extent, you might be able to lower noise by increasing current here and there, but noise tends to be dominated by the first active stage, so I'd look towards replacing that with a quieter IC, and then having a look at the summing amps.
 
mshilarious said:
Do you have hiss or buzz? The former isn't likely a problem with caps. Anyway, hiss. Check the grounding scheme, but it's probably a function of the ICs selected, the gain staging, and lastly the series resistance. To some extent, you might be able to lower noise by increasing current here and there, but noise tends to be dominated by the first active stage, so I'd look towards replacing that with a quieter IC, and then having a look at the summing amps.
Click to expand...
Hmm cool. I guess it's more hiss, my noise floor is higher than it should be. I'm certain it was never intended to be as high as it is... but maybe I'm expecting too much from it (it IS a live desk). If I have all the channel faders down but the mains up full, it does hiss a bit...

I will swap out a couple of the opamps and see if I can notice a difference, but reading up on them is hurting my brain... so many opinions and I'm not quite "getting" it all...

The sends from the boards to the output jacks are not shielded either - would addressing this make a difference, or is it pretty much pointless considering it's all inside a big grounded metal box?
 
slowmotion said:
The sends from the boards to the output jacks are not shielded either - would addressing this make a difference, or is it pretty much pointless considering it's all inside a big grounded metal box?
Click to expand...

Unless you're getting hum, my guess is your time would be better spent swapping op amps or FETs.
 
slowmotion said:
Hmm cool. I guess it's more hiss, my noise floor is higher than it should be. I'm certain it was never intended to be as high as it is... but maybe I'm expecting too much from it (it IS a live desk). If I have all the channel faders down but the mains up full, it does hiss a bit...
Click to expand...

That would be because that's not proper gain staging! With the channel faders all down, you still have the channel buffer amp feeding the summing amp. With the master fader full up, you are amplifying the noise of the channel amps, summing network, and summing amps. But you wouldn't do that normally, you would keep the master fader down, the channel faders at unity, and the channel trims up. You would also mute any unused channels.

When you are reading opamp specs, you are interesting in voltage and current noise; primarily voltage noise on the input opamps. The summing amp I would have to think about, perhaps current noise is a bigger problem there . . . hmmmm. Anyway, get the input opamps quieter and go from there.

What does opamps does the board have now?
 
mshilarious said:
That would be because that's not proper gain staging! With the channel faders all down, you still have the channel buffer amp feeding the summing amp. With the master fader full up, you are amplifying the noise of the channel amps, summing network, and summing amps. But you wouldn't do that normally, you would keep the master fader down, the channel faders at unity, and the channel trims up. You would also mute any unused channels.
Click to expand...
Indeed, but I figured it would show up the "most" noise. I certainly don't actually use it like that (hehe)... perhaps it's an unfair measure.
When you are reading opamp specs, you are interesting in voltage and current noise; primarily voltage noise on the input opamps. The summing amp I would have to think about, perhaps current noise is a bigger problem there . . . hmmmm. Anyway, get the input opamps quieter and go from there.

What does opamps does the board have now?
Click to expand...
So each channel, in the mic preamps there is a 748 up front, and 3 x 4558s for the eq and other gain stages. The line ins omit the 748 but are otherwise similar.

In the main section there is another 2 x 4558s (for each side), one being used as a virtual ground (which I do not fully understand how that works) and one as the actual output.

Sadly all the 4558s (which is the bulk of them by far) aren't socketed, so I've gotta put some sockets in to try different ones. Ugh.

My initial plan is to socket the main output 4558s and one channel and try some TL072s in there to see what's what.

Does that sound like a decent starting point? Should I change the "virtual earth" op amps too?

Oh and thanks for all the info so far too!
 
You must log in or register to reply here.
Back
Top