Following this question, I have fullfilled JP's wishes: https://www.diyaudio.com/community/...ot-your-typical-amplifier.390272/post-7122671
I have measured a number of vintage 10µ/25V capacitors. One exception is the "control", modern cap, radial of course, ROHS, miniature, 105°C and much smaller.
I have also measured a very old (1960's) cap made by Philips, but since it is 6.3V type, I measured its leakage current at 5V instead of 10V for the rest.
They have all been measured in identical conditions: I plugged them in parallel on a breadboard, with removable links to measure the current, and I fired them all at the same time, leaving them a few minutes to stabilize.
None had ever been used, and they have been rudely awakened from a sleep ranging from 2~3years for the control to 60 years for the eldest.
The ITT and Philips axials are both from the late seventies.
On this pic, the bottom one is the oldest and the top one the control:
Here are the results:
Unsurprisingly, the oldest is the worst: it has had a long time to lose its water, and has a poor esr. The ITT and Philips are very much in the same league. The more modern, radial and extremely light Philips is crap; probably the worst caps made by Philips.
The modern control has a tighter tolerance, a lower leakage (normal anyway, it didn't sit for decades on a shelf), but a higher esr, probably due to the higher temperature rating and the higher resistivity electrolyte
I have measured a number of vintage 10µ/25V capacitors. One exception is the "control", modern cap, radial of course, ROHS, miniature, 105°C and much smaller.
I have also measured a very old (1960's) cap made by Philips, but since it is 6.3V type, I measured its leakage current at 5V instead of 10V for the rest.
They have all been measured in identical conditions: I plugged them in parallel on a breadboard, with removable links to measure the current, and I fired them all at the same time, leaving them a few minutes to stabilize.
None had ever been used, and they have been rudely awakened from a sleep ranging from 2~3years for the control to 60 years for the eldest.
The ITT and Philips axials are both from the late seventies.
On this pic, the bottom one is the oldest and the top one the control:
Here are the results:
Unsurprisingly, the oldest is the worst: it has had a long time to lose its water, and has a poor esr. The ITT and Philips are very much in the same league. The more modern, radial and extremely light Philips is crap; probably the worst caps made by Philips.
The modern control has a tighter tolerance, a lower leakage (normal anyway, it didn't sit for decades on a shelf), but a higher esr, probably due to the higher temperature rating and the higher resistivity electrolyte
Thank you. Results probably will change after a few hours under voltage, especially the leakage.
Yes, a few minutes compared to decades of inactivity is ridiculously small, but it shows the resilience of good, reputable caps: 1 or 2µA of leakage is a bit large for such small caps, but it would be no problem in many applications, and it probably remains within the datasheet limits.
The main goal was to compare ITT and Philips, but I can test better, more modern caps. I am not holding my breath though: they will certainly outperform their ancestors, but not by huge amounts.
By the seventies, manufacturers like Siemens had achieved a remarkable level of excellence. Size was huge by today's standards though
The main goal was to compare ITT and Philips, but I can test better, more modern caps. I am not holding my breath though: they will certainly outperform their ancestors, but not by huge amounts.
By the seventies, manufacturers like Siemens had achieved a remarkable level of excellence. Size was huge by today's standards though
There is no harm in using recently produced stuff and then pick the better ones. Radial caps are the de facto standard as well.
I find this information very interesting, and I have an anecdote to share concerning my experience those blue-sleeved Philips axial caps from the 1970s.
Two 22µF/40V Philips axials were included as final coupling caps in a DAC I bought a couple of years ago (one of those with eight-paralleled TDA1387 chips in it). The Philips caps can be seen in this photo, top-left.
One of the quirks of this dac in use was that I would get noticeable pops when it locked onto a signal, or when the lock was disrupted, such as when opening the CD tray on a player that I was using to feed a SPDIF signal to the dac.
Last month, I decided to pull those caps and replace them. First I checked what level of VDC they were holding back from the outputs, and it was 2.70V, so a cap was definitely needed. I replaced them with 2.2µF/450V Panasonic ECW PP film caps (the brown "chicklet" ones), which just fit in the space. I measured the blue Philips caps with a cheap LCR meter (the only such meter I have) so they are not great measurements, but nonetheless these caps tested horribly: 30µF, 10%Vloss, 6Ω ESR. For comparison, the Panasonic PP caps I replaced them with measured 2.14µF, 0%Vloss, 0.35Ω ESR on the same meter.
But however questionable these measurements might be, my experience after replacing them with the PP caps (with only one tenth the capacitance) was telling: no more pops. Not one.
Two 22µF/40V Philips axials were included as final coupling caps in a DAC I bought a couple of years ago (one of those with eight-paralleled TDA1387 chips in it). The Philips caps can be seen in this photo, top-left.
One of the quirks of this dac in use was that I would get noticeable pops when it locked onto a signal, or when the lock was disrupted, such as when opening the CD tray on a player that I was using to feed a SPDIF signal to the dac.
Last month, I decided to pull those caps and replace them. First I checked what level of VDC they were holding back from the outputs, and it was 2.70V, so a cap was definitely needed. I replaced them with 2.2µF/450V Panasonic ECW PP film caps (the brown "chicklet" ones), which just fit in the space. I measured the blue Philips caps with a cheap LCR meter (the only such meter I have) so they are not great measurements, but nonetheless these caps tested horribly: 30µF, 10%Vloss, 6Ω ESR. For comparison, the Panasonic PP caps I replaced them with measured 2.14µF, 0%Vloss, 0.35Ω ESR on the same meter.
But however questionable these measurements might be, my experience after replacing them with the PP caps (with only one tenth the capacitance) was telling: no more pops. Not one.
I've seen two reports on assessing shelf life of e-caps - link below to one, but I can't find the other just now (I recall it was related to defence stocks). Any assessment is prone to a wide variation in sample-to-sample outcome, and just testing a few samples needs a lot of caution when interpreting what that may mean generically.
https://digital.library.unt.edu/ark:/67531/metadc3104/m2/1/high_res_d/thesis.pdf
I was able to locate a leakage spec on a 1960's e-cap range, which supports the general view that consumer-grade technology has improved over the decades, And as Elvee indicates, well made e-caps from as early as the 70's can still exhibit excellent parameters nowadays if not stressed over the decades, with no good reason to replace them if they are in-circuit. Of course stress can take its toll, so parts sitting in a known hotter local environment, or liable to external stress (first filter cap in a power supply) are definite red-flags.
https://digital.library.unt.edu/ark:/67531/metadc3104/m2/1/high_res_d/thesis.pdf
I was able to locate a leakage spec on a 1960's e-cap range, which supports the general view that consumer-grade technology has improved over the decades, And as Elvee indicates, well made e-caps from as early as the 70's can still exhibit excellent parameters nowadays if not stressed over the decades, with no good reason to replace them if they are in-circuit. Of course stress can take its toll, so parts sitting in a known hotter local environment, or liable to external stress (first filter cap in a power supply) are definite red-flags.
I have added some other types to the line-up, and I have increased the reforming time to a few hours (not 24 hours though).
The first of the list is a solid aluminum made by Philips, but now discontinued.
I have added two modern types: Nichicon, KME series.
I have also included a professional, high-rel type, with an additional resin seal over the rubber: Marcap, axial 10µ/50V.
Finally, I also added a classic Siemens tantalum 10µ/35V, in a metal case and a glass seal. Unsurprisingly, it beats all the rest on all counts.
The additional caps have a higher voltage rating than the original 25V ones: 35V to 63V, which helps regarding the leakage current.
The deeper reforming generally had the effect of slightly decreasing the capacitance, and slightly increasing the ESR, both of which are normal. High temp caps have a higher ESR, which is normal because of the higher resistivity of the electrolyte at higher temperature
The first of the list is a solid aluminum made by Philips, but now discontinued.
I have added two modern types: Nichicon, KME series.
I have also included a professional, high-rel type, with an additional resin seal over the rubber: Marcap, axial 10µ/50V.
Finally, I also added a classic Siemens tantalum 10µ/35V, in a metal case and a glass seal. Unsurprisingly, it beats all the rest on all counts.
The additional caps have a higher voltage rating than the original 25V ones: 35V to 63V, which helps regarding the leakage current.
The deeper reforming generally had the effect of slightly decreasing the capacitance, and slightly increasing the ESR, both of which are normal. High temp caps have a higher ESR, which is normal because of the higher resistivity of the electrolyte at higher temperature
Elvee, are you sure the Siemens is not a wet tantalum cap?
BTW it would help to mention series name and data code of all tested caps.
BTW it would help to mention series name and data code of all tested caps.
Absolutely not: wet tantalum have a silver body, PTFE/elastomer seal, etc.
This one is a regular solid type, but I have the same model from many origins: Kemet, LCC etc. and I can test them too. I suspect that the Siemens will remain the leader though
I'll try to find additional info, but these caps are very small size, and generally there little else than the value, voltage, temperature and manufacturer logo printed on them
This is the Siemens cap, but unfortunately I couldn't find a complete datasheet on the net, just a summary.
https://www.tedss.com/CSR13E106KP
If someone has kept a databook from the seventies, it is probably there.
https://www.tedss.com/CSR13E106KP
If someone has kept a databook from the seventies, it is probably there.
Bonjour! The Philip solid aluminum is that one of series SAL-RPM 122 or 128? Just like BG these really need some time to perform optimally. SAL-RPM 128 are capable of reverse polarity 0.3 x Ur (DC) and 0.8 x Ur (AC). Also 175 degrees Celsius maximum temp and 20.000 hours lifetime @ 125 degrees Celsius. I like them a lot as they don't suffer from aging as much as wet electrolyte caps.
Last edited:
Probably 128: it is printed above the value:
Below is another code, L5M. Probably the date or place of production.
Other details: the Siemens tantalum has 8624 as date code, meaning it is somewhat more recent than what I thought.
I have tested similar 150D's from other brands, and they measured well too, but Siemens remained on top, except for Kemet which was marginally better. All the tantalum were better than any of the others, (unfortunately): tantalum is bad for the planet, bad for the people and bad for the wallet, but performance-wise it is difficult to outperform.
The Nichicon KME's have additional figures: the radial has 09 T3 stamped in the aluminum, and the axial has 31 H(5) printed on the sleeve. Barely legible, and I have no idea about the meaning
Below is another code, L5M. Probably the date or place of production.
Other details: the Siemens tantalum has 8624 as date code, meaning it is somewhat more recent than what I thought.
I have tested similar 150D's from other brands, and they measured well too, but Siemens remained on top, except for Kemet which was marginally better. All the tantalum were better than any of the others, (unfortunately): tantalum is bad for the planet, bad for the people and bad for the wallet, but performance-wise it is difficult to outperform.
The Nichicon KME's have additional figures: the radial has 09 T3 stamped in the aluminum, and the axial has 31 H(5) printed on the sleeve. Barely legible, and I have no idea about the meaning
Hi, I used to think the same of coltan (tantalum) but the situation meanwhile has changed. The company that buys 50% of the worldwide mining does not buy anything from conflict states anymore and some of the manufacturers that bought coltan from conflict states don't do that anymore. IMHO checking if the materials are from conflict free mining is not a task of the end user but it should be regulated by international laws.
Tantalum SMD caps are simply very good. Manganese dioxide/niobium capacitors can be quite good too. These look the same as tantalum capacitors. Only good experiences with modern SMD versions, the drop type through hole types were a disaster.
BTW do you know these? It was said these were used for space applications. I normally throw away old stuff but I kept a number of NOS/NIB tantalum caps I once bought in a lot of different parts. I am pretty sure these wet tantalums were made with coltan from conflict areas, probably being the reason why I never have used them. The text says mallory CL17DM121MP3, 120 µF 90V DC, 60V rated at 175 degrees Celsius (!!!), date code 1968. Price then was 20 US $ a piece so they were something special.
Tantalum SMD caps are simply very good. Manganese dioxide/niobium capacitors can be quite good too. These look the same as tantalum capacitors. Only good experiences with modern SMD versions, the drop type through hole types were a disaster.
BTW do you know these? It was said these were used for space applications. I normally throw away old stuff but I kept a number of NOS/NIB tantalum caps I once bought in a lot of different parts. I am pretty sure these wet tantalums were made with coltan from conflict areas, probably being the reason why I never have used them. The text says mallory CL17DM121MP3, 120 µF 90V DC, 60V rated at 175 degrees Celsius (!!!), date code 1968. Price then was 20 US $ a piece so they were something special.
Attachments
Last edited:
No, I have never seen tantalums in such a case; the specs are really impressive. I have seen paper caps having a very similar look, but not tantalum. I also have an odd-looking wet tantalum somewhere; I'll try to find it and post a pic.
Your caps have a glass-metal seal, which is odd for wet tantalum: normally, the electrolyte is sulphuric acid, which precludes many materials for the case and connections
Your caps have a glass-metal seal, which is odd for wet tantalum: normally, the electrolyte is sulphuric acid, which precludes many materials for the case and connections
Here is the odd-looking cap; it is a 15µ/100V IIRC. It is tarnished because it is made of solid silver.
Ok I dug them up for archeological purposes 🙂 Please note the 125 degrees rating on the Sprague 600D high reliability series 10/50. I also found a stash of NOS silver mica, potted polystyrene caps and a few calibration capacitors.
As you can see large value tantalum caps did exist in 1981 (if you were prepared to pay the price). ITT TAA series 330/16.
The 600D are still made:
https://www.vishay.com/docs/42047/600d.pdf
As you can see large value tantalum caps did exist in 1981 (if you were prepared to pay the price). ITT TAA series 330/16.
The 600D are still made:
https://www.vishay.com/docs/42047/600d.pdf
Attachments
Last edited:
Well, replacing them with 1/10 the capacitance invalidates the comarison. The RC time constant is now 1/10 meaning that a pop is 1/10th duration and 1/10th amplitude.
This doesn't tell you anything about the caps.
Jan
Thanks for this comment. I thought the pop was probably a DC impulse that the cap was not successfully blocking, but I freely admit that I don't know that.
My electronics knowledge is quite basic, but growing (albeit slowly and unevenly) as I read more here, and consult electronics texts from time to time to shore up some basics. For example, I did consider what the effect on freq response would be by reducing the capacitance of that cap by a factor of ten, but for the high input impedance of the preamp the dac was feeding, the higher corner frequency was still plenty low enough for me.
Thanks again.
My electronics knowledge is quite basic, but growing (albeit slowly and unevenly) as I read more here, and consult electronics texts from time to time to shore up some basics. For example, I did consider what the effect on freq response would be by reducing the capacitance of that cap by a factor of ten, but for the high input impedance of the preamp the dac was feeding, the higher corner frequency was still plenty low enough for me.
Thanks again.