wet tantalum higher power audio coupling capacitor?

[coppercone2]

For coupling audio+ signals (possibly from small power amp) to lower Z loads, could a 2x series 'bipolar' capacitor made with wet tantalum capacitors work ?

Not much info about it. I read capacitor current leakage is the main problem with bipolar caps made from electrolytics. the wet tantalum capacitor seems to be listed as having much lower leakage then solid tantalum.

Like a few hundred uF value. This would be in a pomona project box with BNC connectors on it.

[CosteC]

Test it
Issue is unequal leakage between two caps causing reverse polarisation of one of them leading to failure.

When testing, measurements of leakage shall be done in room and in elevated temperature as minimum - leakage usually rise with temperature.

[jonpaul]

schematic? power level? V C?

j

[Gyro]

Yes, wet Tantalum leakage currents are much lower than the equivalent Aluminium electrolytics. This means that you could use much higher value balancing networks. You should at least put a reverse biased diode across each capacitor - this would cause distortion if the caps got so out of balance that one of the diodes became forward biased but would at least prevent capacitor reverse bias damage.

The alternative would be a high value resistor linking the centre point to bias source so that each capacitor would be provided with a defined positive bias voltage. Edit: This method is sometimes used in speaker crossovers where people want to avoid using bipolar electrolytics and include a bias battery and polar electrolytics.


P.S. This appears to be a double post of https://www.eevblog.com/forum/testgear/low-frequency-ac-coupling-box-bnc/

[T3sl4co1l]

If you want to throw away a couple hundred bucks, I have a bridge that recently came on the market...

Tim

[coppercone2]

ebay. i agree digikey tantalum is not acceptable.

i might even have a bandolier of them in the parts box for 7.99 a long time ago

its not a duplicate of the thread because I don't expect this to exist the other thread is meant for someone to type in something like 'try tektronix 0192" which may bring a ready made box that may or may not work. i don't expect anyone to even list finished couplers here because its specifically about capacitor type  for the other thread I only expect a low power solution that might handle 50 ohm function gen to >600 ohm input impedance audio circuit (i doubt they even make it for 50 to 50). it would still be handy to have an 'approved' test equipment coupler box made with whatever company thought was a good idea for lab work. If someone made a 'power' capacitor coupler commercially I would be impressed, they would want you to buy a transformer most likely I think.

and the power voltage levels for this one probobly under 50 volts to however much power seems reasonable, probobly no more then a few amps. really depends on what tantalum I can get. its really hard to say what frequency and power load might come by for a generic lower frequency amp coupler.. that's why I did not put a spec

[tridac]

Why the need for a capacitor, is it to block dc, or what ?. Solid Tantalum caps usually have lower esr than ali caps, but need to take account of the ripple rating if carrying a ac power load. If it gets warm or hot, it's probably overloaded. Tend to be expensive, but bargains on Ebay. Here, picked up 78 x 100uf 35v Kemet J old stock, for 17 ukp. If cheap enough, would always use for one off projects in preference to ali electrolytics...

[coppercone2]

yeah a DC block I noticed some of these things have a small DC signal thats messing with my op amp stuff. But that is low power. For higher power I did not come across a use but when I was shopping for boxes I thought why not

I made like a composite op amp that was driving me insane all day with bizzare signals, I finally realized the function gen 'AC' mode had a 30mV offset. When I manually put it into offset mode and zero it out, the op amps were fine. I thought it was some kind of bull shit with ground planes (like weird slopes in negative cycle). This thing would work, to a certain power level, then it would 'latch' into a waveform that the negative portion of cycle was turning into a linear ramp line between like 65-80% of its duration.

Compensation, input problems, idk.. turned out that the small offset was making it crazy. Not clipping. like if you erased a portion of the bottom sine wave after the lowest voltage point, and drew a strait line, then turned it sinusoidal again. and you need to reset signal to basically 0 before it went away. partial AC latchup failure or something

because I had it wired on strip board with a bunch of wires to a star ground. I thought wow do i really need a ground plane for this low frequency? (lots of decoupling capacitors were there already). Guess not, it was just the offset voltage on the input signal.

probobly should take a look inside the FG to see if maybe there is a bad cap or what (tektronix cfg280 or something, cheap unit.. but when you put a noctua fan in it, its very quiet and a pleasure to use for prototypes).

I got the strip breaker tool (4mm drill bit in a starrett collet) and the keystone rivet-in threaded standoffs, and it turns out strip board is a pleasure to work on. Also bought an assortment of teflon solid wire in alot of different bright colors (22 awg?). With the big ones there is no problem to mount a heat sink on it. Highly recommend getting the large strip board segments (like 6x4 inches at least). Just put standoffs in it offset from your PCB vise and its really handy. way more fun then simulation. I just need to get the smallest hand drill they have for breaking strips with the drill bit, bit of a pain manually.

[tridac]

A cap sounds like a good idea. Iirc, you said you were driving a 600 ohm load ?. If so work out the cap value by setting the reactance at lowest frequency of interest to be  small fraction of the input 600 ohms, say a tenth. Haven't checked, but my guess would be about 10uF would be about right. At a few tens of mV offset, just use two tant caps back to back, as the leakage at that sort of mV level should be almost zero anyway...

[coppercone2]

no, I mean a commercial audio coupler expects a 600 ohm load. I figure it can be made to support much lower with the right capacitor. 10uF is about as big as you can get with foil before it gets stupid

[floobydust]

For a telephone bridge, somebody used back-back electrolytic capacitors with reverse diodes. I vaguely recall 680uF 100V parts maybe, for low end freq. response testing in the lab.
It didn't work very well. It either added distortion (rectification) or mysteriously would go low value, all intermittent of course which drove me nuts.
Using two capacitors, they're never identical for capacitance or leakage, so charge distribution is never symmetrical and one cap gets the short end of the stick.
What is that exactly? I found it gets a -ve voltage across it (to the diode clamp limit say -0.7V), its leakage current then goes up, which in time that cap needed reforming to then work right.
So the whole affair needed reforming every year or so. I don't think tantalums can take any -ve voltage at all, whereas electrolytics are supposedly OK with say -1V damage-wise.
But leakage current still needs to be taken into consideration with clamp-diodes present.

[coppercone2]

alright tantalum may not solve anything.


how about just non polarized electrolytics? I mean they exist. No one is raving about them but is it the best thing for the job?

like this one
https://www.digikey.com/en/products/detail/nte-electronics-inc/NPA470M25/11649403

put like 4 of em in parallel in a hammond box

[JeremyC]

For coupling audio+ signals (possibly from small power amp) to lower Z loads, could a 2x series 'bipolar' capacitor made with wet tantalum capacitors work ?

Not much info about it. I read capacitor current leakage is the main problem with bipolar caps made from electrolytics. the wet tantalum capacitor seems to be listed as having much lower leakage then solid tantalum.

Like a few hundred uF value. This would be in a pomona project box with BNC connectors on it.

I believe bi polar electrolytic capacitors made by Rybycon or Nichicon will be fine. Purchase them from trusted vendor and not from Aliexpress, Ebay, etc.

[BrianHG]

They are a little pricey, but have you considered film capacitors?
TDK makes AC film capacitors at 100uf, 220uf and 470uf.
They are a little big, but they should have very little leakage current.
For price, and size at >100uf you wont beat AC electrolytic caps.

[coppercone2]

Hmm the 10uF film I have is the size of a film canister already.

price is not that much of an object because I see leakage can wreck havok and waste mad time on sensitive circuits that expect AC.

I thought it would be the size of a coffee can.

I think I found my best bet here
https://www.digikey.com/en/products/detail/vishay-beyschlag-draloric-bc-components/MKT1820756065/5380169

[BrianHG]

1 thing I used to do is guarantee the input offset voltage of my amp allowing me to use a DC cap.
For example, I had a 50v offset to GND, so I knew I could get away with a 150v DC decoupling cap for my input as no audio equipment should drive +51v DC into an input driving that cap into reverse polarity.

[BrianHG]

Take a look at all the blue plastic box caps in this link....

10uf to 100uf AC film caps

[BrianHG]

Example 10uf for under 5$
https://www.mouser.ca/ProductDetail/EPCOS-TDK/B32522C0106K189?qs=35GPJvxY%252Bl7IpAJmsxu9Lg%3D%3D
https://www.mouser.ca/ProductDetail/EPCOS-TDK/B32523Q1106K?qs=p%2F12KT2rJdXy%252B2Sovvj%2F8g%3D%3D
second one is 63vac and costs less...

[coppercone2]

I bought the 560uF + box

that should be pretty darn universal for testing stuff without the extra bs of soldering on a AC coupling network on a prototype of breadboard.

somewhat expensive but good accessories seem to be anyway lol

I think a different circuit I was testing years ago got written off as not viable by myself because of this DC offset shit. god knows how different things could have been lol

[coppercone2]

i swear I triple checked this thing and I still managed to solder a ground to signal.

does anyone feel like there is a invisible man making mistakes for you when you prototype?

[BrianHG]

i swear I triple checked this thing and I still managed to solder a ground to signal.

does anyone feel like there is a invisible man making mistakes for you when you prototype?

You have no fu__en idea....

[coppercone2]

lol and two incorrect resistors for a inverter

flashlight in m mouth holding the pcb upside down poking it with a probe while jiggling capacitors with my thumb.


some fucking how when I was putting my hand on it, it worked better, when it was haywire, and there was also a broken trace on the strip board, this fixture involved a coffee mug too


you can use a flashlgiht pocket clip to clip a flash light on your lip when you get tired of holding it with ur teeth

i some how made a triangle wave generator

i wish I had whatever JC denton had in DX. i know he did prototyping. flashlight glow eyes. or just rudophs nose. a electrician came up with that one

and the compensation sculpture needs to be managed before you tilt it over into a fucking power rail. never get anywhere if you do the compensation neatly. needs to be a low value cap with long leads that you solder other caps to.

[coppercone2]

it fits with the largest box and that coupling capacitor. Bit big though. I used flurosilicone rubber to sandwich the big cap in there

I will make a small version using a 22uF polystyrene capacitor I have

[Gyro]

...
I will make a small version using a 22uF polystyrene capacitor I have

A small version with a 22uF Polystyrene? I doubt it!

[Vovk_Z]

i swear I triple checked this thing and I still managed to solder a ground to signal.

does anyone feel like there is a invisible man making mistakes for you when you prototype?

Yes, he is going around and waits until you get tired to the end of day or until you do something repeatable and you start to think "I've done it 5 times in a row several hours ago and this is the sixth time and it will be the same, what can go wrong".