I'm hoping for a tie breaker. Not sure this is the right forum but it does involve a power supply section.
There is/was a 40ish year old 10uF 25V tantalum capacitor that must be replaced. It was working at virtually it's rated voltage - 24.5V DC and is now cooked. I was advised to use something else. However, on one hand I am told an electrolytic is the way to go and on the other, a (monolithic) ceramic.
I could use a 50V Tantalum too but I wonder what opinions I'd find here.
There is/was a 40ish year old 10uF 25V tantalum capacitor that must be replaced. It was working at virtually it's rated voltage - 24.5V DC and is now cooked. I was advised to use something else. However, on one hand I am told an electrolytic is the way to go and on the other, a (monolithic) ceramic.
I could use a 50V Tantalum too but I wonder what opinions I'd find here.
Last edited:
From DIY guru Rod Elliott: (Capacitor Characteristics)
"Then of course we have tantalum electrolytics. While many sing their praises, I do not recommend their use for anything, other than tossing in the (rubbish) bin. There might be the odd occasion where you really need the properties of tantalum based caps, but such needs should be few and far between. They are unreliable, and have a nasty habit of failing short-circuit. They cannot tolerate high impulse currents and/or rapid charge/ discharge cycles, and especially don't like being shorted. Tantalum caps announce their failure by becoming short-circuited, and it can be extremely difficult to track down a (possibly intermittent) short across a supply bus that powers many ICs. I never use tantalum caps, and don't recommend them in any of the published projects. Personally, I suggest that you don't use them either. "
"Then of course we have tantalum electrolytics. While many sing their praises, I do not recommend their use for anything, other than tossing in the (rubbish) bin. There might be the odd occasion where you really need the properties of tantalum based caps, but such needs should be few and far between. They are unreliable, and have a nasty habit of failing short-circuit. They cannot tolerate high impulse currents and/or rapid charge/ discharge cycles, and especially don't like being shorted. Tantalum caps announce their failure by becoming short-circuited, and it can be extremely difficult to track down a (possibly intermittent) short across a supply bus that powers many ICs. I never use tantalum caps, and don't recommend them in any of the published projects. Personally, I suggest that you don't use them either. "
Tantalums are excellent capacitors. They are used in high end audio products and they are recommended for use in linear regulator supplies by Texas Instruments. They have high values for small bodies and small inductances. They have very good HF impedance curves and do not suffer degradation over time like many "wet tissue" electrolytics. Indeed, tantalums are somewhat self-healing when abused to a small degree.
The critical thing with tantalums is that they do not suffer reverse polarity gladly and will fail open or short if this persists. If used properly they are just fine.
We don't have details of the PSU in question but all other things being equal, it seems sensible to me to replace like for like.
The complaints I hear most about tantalums are from people who don't use them properly. Often amateur amp builders who put them in feedback paths for dc blocking and have a duff design which subjects the tantalum to reverse polarity. I recommend to experimenters to use bipolar caps, at least for initial testing, so as not to accidentally damage a tantalum.
The critical thing with tantalums is that they do not suffer reverse polarity gladly and will fail open or short if this persists. If used properly they are just fine.
We don't have details of the PSU in question but all other things being equal, it seems sensible to me to replace like for like.
The complaints I hear most about tantalums are from people who don't use them properly. Often amateur amp builders who put them in feedback paths for dc blocking and have a duff design which subjects the tantalum to reverse polarity. I recommend to experimenters to use bipolar caps, at least for initial testing, so as not to accidentally damage a tantalum.
Food for thought: Tantalums are excellent capacitors however they need to be used properly. A BIG problem is they don't like ripple currents and they fail short. Modern capacitors have low z and higher voltage ratings. Verify what type of circuit performance is required and select the new capacitor.
Duke
Duke
I agree. The question is why did they fit a Tant originally? AFAIK, they were expensive and not widely used in audio. I think some Tants had very low ESR for their day, and I'm not sure they've got too much in common with modern ones? Were they not de rigeur for big boards of fast ECL and such?
Maybe not great for 50Hz ripple, but come into their own at 100MHz or so?
I wasn't designing quite as long ago as 40 years but I remember them from military products in the late 80s. I recall a mate powering up a card with a dozen of them wrong way around, very noisy!
Also is the 10u critical or was a just a default value?
Nonsense!
I replaced many shorted tantalums from commercial stuff, none of them was reverse polarized. Their poor surge capabilities have been admitted by manufacturers long time ago. Since I encountered a burning tantalum in my equipment in the 70ies I banned them, like Rod Elliot.
Nowadays there are no extra properties that justifies their usage.
I have recently fixed a test equipment from a major company. It has a blown tantalum capacitor that was used in a location where surge due to inductive switching is always present. The tantalum voltage rating was more than 3x the circuit 's DC voltage.
The catastrophic failure of this tantalum capacitor is ALL due to the designer.
The catastrophic failure of this tantalum capacitor is ALL due to the designer.
I found tatalum the best to get noise out of power supply rails in video circuits with digital.
Much better than any electrolytic and ceramic combos. A low impedance over a wide band I guess.
Tantalum is great when you need microvolt clean rails.
Failures I have not seen apart from overvoltage and reverse voltage. May be, all tantalum are not born equal.
Much better than any electrolytic and ceramic combos. A low impedance over a wide band I guess.
Tantalum is great when you need microvolt clean rails.
Failures I have not seen apart from overvoltage and reverse voltage. May be, all tantalum are not born equal.
Last edited:
I find the comment that a tantulum did duty for 40ish years compelling and am sorlely tempted to use a new one but with double the circuit's voltage. I am also keen to use something I can buy locally which are those, monolithic ceramic and general purpose electrolytic's. The reason is purely monitory. Living in Canada it is hellishly expensive to obtain goods from online US vendors.
I see no one touched on monolithic ceramics? I received acknowledgement from someone with good knowledge of circuits and their design, that one of those would be a good substitute in this instance. But he had benefit of a schematic to refer to.
But, someone else equally knowledgeable advised me ceramics were a bad idea.
My meager recent reading on the matter left me with the impression this "new breed" of ceramics, was more or less developed as an alternative to tantalums.
EDIT: My apprehesion regarding tantalum capacitors is based primarily on what both of the individuals consulted told me which to avoid them at all cost. Referring again to my recent read up's the concencus seems to be about 25% for and 75% against.
2ND EDIT: I just thought of something. The schematic identifies the capacitor in question as a D.T. which I now understand means Dipped Tantalum. That makes me wonder if it is some particular characteristic (other than physical size) that explains it's use. Or was it physical size?
It is starting to seem more logical to replace tantalum with tantalum.
Is it normal to agonize about a choice such as this?
I see no one touched on monolithic ceramics? I received acknowledgement from someone with good knowledge of circuits and their design, that one of those would be a good substitute in this instance. But he had benefit of a schematic to refer to.
But, someone else equally knowledgeable advised me ceramics were a bad idea.
My meager recent reading on the matter left me with the impression this "new breed" of ceramics, was more or less developed as an alternative to tantalums.
EDIT: My apprehesion regarding tantalum capacitors is based primarily on what both of the individuals consulted told me which to avoid them at all cost. Referring again to my recent read up's the concencus seems to be about 25% for and 75% against.
2ND EDIT: I just thought of something. The schematic identifies the capacitor in question as a D.T. which I now understand means Dipped Tantalum. That makes me wonder if it is some particular characteristic (other than physical size) that explains it's use. Or was it physical size?
It is starting to seem more logical to replace tantalum with tantalum.
Is it normal to agonize about a choice such as this?
Last edited:
Using a Ta cap in a low ripple current application at 2x the expected voltage will be very safe. There is no wear-out mechanism for solid Ta, only failure modes, so once it settles in successfully, it could last forever.
Many Ta caps brands are good, but I have found that the Kemet parts perform the best and seem to be the most reliable. Their impedance seems to be more consistent from unit to unit, which probably points to their high reliability - no odd 'hot spots' to worry about.
Many Ta caps brands are good, but I have found that the Kemet parts perform the best and seem to be the most reliable. Their impedance seems to be more consistent from unit to unit, which probably points to their high reliability - no odd 'hot spots' to worry about.
If you could tell us what circuit, function, temperature, etc. this tantalum capacitor has failed, you will likely get a more direct answer instead of relying on the number of yes and no you get from this thread.
I regularly buy parts from DigiKey and nothing beats their next day C$8 delivery, free shipping for >+C$100. It is way cheaper and less stressful than driving to Vancouver or Burnaby and dealing with a severely limited parts choice from the half a dozen parts stores.
I regularly buy parts from DigiKey and nothing beats their next day C$8 delivery, free shipping for >+C$100. It is way cheaper and less stressful than driving to Vancouver or Burnaby and dealing with a severely limited parts choice from the half a dozen parts stores.
Last edited:
Here is he portion of schematic containing the capacitor in question - C29.
It is of a PC board from a Teac A2300SD R2R which the SM refers to as Bias Oscillator. It also seems to contain the power supply or perhaps more appropriately AC Regulator. I have been inside two other Teac's (older than this one) and the power supply was on it's own separate board.
Regarding ordering. I am in a different situation, my employment involves driving in and around Vancouver a great deal. Consequently I am frequently nearby a local electronic parts vendor once or twice a day. Plus, my parts purchases are way less than $100, just getting what I need for the particular task at hand. I only tinker with stuff I can buy for cheap as that's all I can afford.
I do order from Digikey, Mouser, Newark Element 14, Electronic Surplus, Audio Proz and someplace in California (that gouged severely on some long obsolete oddball transistors) but only if I cannot pick up what I want during the day. It just seems something as pedestrian as a capacitor for this application should not cost huge and take a week or two to get to me.
EDIT: Thanks Monte for the endorsement of the tantalum.
It is of a PC board from a Teac A2300SD R2R which the SM refers to as Bias Oscillator. It also seems to contain the power supply or perhaps more appropriately AC Regulator. I have been inside two other Teac's (older than this one) and the power supply was on it's own separate board.
Regarding ordering. I am in a different situation, my employment involves driving in and around Vancouver a great deal. Consequently I am frequently nearby a local electronic parts vendor once or twice a day. Plus, my parts purchases are way less than $100, just getting what I need for the particular task at hand. I only tinker with stuff I can buy for cheap as that's all I can afford.
I do order from Digikey, Mouser, Newark Element 14, Electronic Surplus, Audio Proz and someplace in California (that gouged severely on some long obsolete oddball transistors) but only if I cannot pick up what I want during the day. It just seems something as pedestrian as a capacitor for this application should not cost huge and take a week or two to get to me.
EDIT: Thanks Monte for the endorsement of the tantalum.
Attachments
Last edited:
That part of the circuit seems benign but a closer look reveals L21 that can get its stored energy from C21 when it is not yet charged. The stored energy from L21 will then discharge to the one with the least impedance, your C29. I suggest replacing it with a multi-layer ceramic capacitor 10uF 35V/50V X7R/X5R 1206 or bigger.
I don't think capacitance is the main parameter for the original C29 choice but decent high frequency characteristic. An additional 1uF stacked polyester film capacitor across might also help suppress the leakage from the bias oscillator.
Replace C22 220uF/35V with a low ESR type like Panasonic FM/FR, UCC KY or Nichicon HE.
Measure the bias oscillator leakage across C29 with a high sensitivity scope and then measure at the other side of R27 56R 1W. You can add more filtering on the other side of R27 to 0V/ground/"29" if there is still some oscillator leakage that is being fed back to the 24V supply.
I don't think capacitance is the main parameter for the original C29 choice but decent high frequency characteristic. An additional 1uF stacked polyester film capacitor across might also help suppress the leakage from the bias oscillator.
Replace C22 220uF/35V with a low ESR type like Panasonic FM/FR, UCC KY or Nichicon HE.
Measure the bias oscillator leakage across C29 with a high sensitivity scope and then measure at the other side of R27 56R 1W. You can add more filtering on the other side of R27 to 0V/ground/"29" if there is still some oscillator leakage that is being fed back to the 24V supply.
Last edited:
I agree about oscillator leak suppression. Tantalum is a good design choice.
Coincidence: I own a Teac AS 2300 since 1976.
Congratulations it is a great performer.
I converted it for 220v 50 Hz, there is a pulley for that to do 50 Hz or 60 Hz available in Teac parts. The original pulley is for 60 Hz only. There are big caps at motors, they do not need a change for 60 to 50 Hz.
Coincidence: I own a Teac AS 2300 since 1976.
Congratulations it is a great performer.
I converted it for 220v 50 Hz, there is a pulley for that to do 50 Hz or 60 Hz available in Teac parts. The original pulley is for 60 Hz only. There are big caps at motors, they do not need a change for 60 to 50 Hz.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.