|
Mike Knudsen wrote:
Nope, you make sense. Maybe not so much demand in the Ham BA world, but over on radio+phono you'll find purists who want caps to stuff. The problem is, they already have the original old caps in the radio they're restoring, so they don't need yours. Not always...thats why I need more. Often times the original caps have already been hacked out or are in too terrible a condition for restuffing. I try to keep some made up ahead of time and where a set like a Zenith or Philco uses brand specific ones I like to replace with the same. I considered stuff-n-sell but it really is a time consuming task. I'd feel stupid trying to sell them at what they are worth dollar-wise in time...and of course anyone can do their own for free if they really care! But -- some enterprising retiree may want to stockpile pre-stuffed restored caps and sell them to other restorers, ready for insertion in the radio. If so, he'd want yours for starters to build up inventory. Ultimately, he'd take the old ones in exchange, but meanwhile he needs extras. Hopefully he'd pay enough to cover the postage :-) --Mike K. I beg for these things! And of course always pay postage! -Bill |
The capacitance and voltage rating of an electrolytic capacitor is
set by the thickness of an aluminum oxide layer deposited on the metal foil "plates" of the capacitor. Of course, the bulk amount of square inches of aluminum plate material also determines the range of capacitance that is possible with varying thicknesses of oxide. The electrolyte serves the purpose of making an intimate electrical connection to the plate. The problem is the electrolyte shows some tendancy to dissolve the oxide layer. When this happens, the capacitor's value increases, and its safe voltage rating decreases. This is why a long disused electrolytic capacitor tends to blow up when it is abruptly put back into service. Electrolytic capacitors are self adjusting for working voltage (to some degree). If they are operated for a long time at 50% of their rating, the oxide reduces in thickness, and they become higher capacitance, and lower working voltage. If you try to increase their operating voltage, they will draw too much current. They will either adapt to the new higher voltage, or they will blow up from the heat. The tolerance has nothing to do with manufacturing capabilities, or price, and everything to do with the highly variable nature of the oxide layer. -Chuck, WA3UQV Mike Knudsen wrote: In article , "Frank Dresser" writes: Just as speculation, let's say cap manufacturers have learned to make electrolytic capacitors with good precision at little extra cost. And let's imagine that setting the target capacitance to 5% - 10% low reduces the cost of the "active ingredients" by 5% -10%. Well, that would be a nice reward for knowing how to do the job! This makes very good sense. I suspect that back in the old days, manufacturers would throw in up to 100% extra foil plates area just to make sure they at least met the rated capacitance. So you would get caps well over the ratings. But yes, once they got the process down really tight, why toss in extra material. In fact, shaving it on the low side is just what the front-office bean coutners probably tell them to do nowadays! --Mike K. Oscar loves trash, but hates Spam! Delete him to reply to me. |
The capacitance and voltage rating of an electrolytic capacitor is
set by the thickness of an aluminum oxide layer deposited on the metal foil "plates" of the capacitor. Of course, the bulk amount of square inches of aluminum plate material also determines the range of capacitance that is possible with varying thicknesses of oxide. The electrolyte serves the purpose of making an intimate electrical connection to the plate. The problem is the electrolyte shows some tendancy to dissolve the oxide layer. When this happens, the capacitor's value increases, and its safe voltage rating decreases. This is why a long disused electrolytic capacitor tends to blow up when it is abruptly put back into service. Electrolytic capacitors are self adjusting for working voltage (to some degree). If they are operated for a long time at 50% of their rating, the oxide reduces in thickness, and they become higher capacitance, and lower working voltage. If you try to increase their operating voltage, they will draw too much current. They will either adapt to the new higher voltage, or they will blow up from the heat. The tolerance has nothing to do with manufacturing capabilities, or price, and everything to do with the highly variable nature of the oxide layer. -Chuck, WA3UQV Mike Knudsen wrote: In article , "Frank Dresser" writes: Just as speculation, let's say cap manufacturers have learned to make electrolytic capacitors with good precision at little extra cost. And let's imagine that setting the target capacitance to 5% - 10% low reduces the cost of the "active ingredients" by 5% -10%. Well, that would be a nice reward for knowing how to do the job! This makes very good sense. I suspect that back in the old days, manufacturers would throw in up to 100% extra foil plates area just to make sure they at least met the rated capacitance. So you would get caps well over the ratings. But yes, once they got the process down really tight, why toss in extra material. In fact, shaving it on the low side is just what the front-office bean coutners probably tell them to do nowadays! --Mike K. Oscar loves trash, but hates Spam! Delete him to reply to me. |
Chuck Harris wrote:
Electrolytic capacitors are self adjusting for working voltage (to some degree). If they are operated for a long time at 50% of their rating, the oxide reduces in thickness, and they become higher capacitance, and lower working voltage. If you try to increase their operating voltage, they will draw too much current. They will either adapt to the new higher voltage, or they will blow up from the heat. Thats very much true with older caps up to about 1970. But later model caps don't exhibit this 'memory'. -Bill |
Chuck Harris wrote:
Electrolytic capacitors are self adjusting for working voltage (to some degree). If they are operated for a long time at 50% of their rating, the oxide reduces in thickness, and they become higher capacitance, and lower working voltage. If you try to increase their operating voltage, they will draw too much current. They will either adapt to the new higher voltage, or they will blow up from the heat. Thats very much true with older caps up to about 1970. But later model caps don't exhibit this 'memory'. -Bill |
"Chuck Harris" wrote in message ... [snip] The tolerance has nothing to do with manufacturing capabilities, or price, and everything to do with the highly variable nature of the oxide layer. -Chuck, WA3UQV OK, I'll grab some of those old Ducatis and check 'em. I'm sure they haven't seen a polorizing voltage in at least 25 years, because that's when I bought them as surplus. Then I'll run 'em up to their rated voltage for 24 hours and recheck. I'll see how much extra voltage they take and hold them at that voltage for another 24 hours. Just for curiosity, I'll also check the ESR at each step with my Dick Smith meter. I know none of this is lab quality procedure, but if there's any gross changes, I think I'll catch 'em. Frank Dresser |
"Chuck Harris" wrote in message ... [snip] The tolerance has nothing to do with manufacturing capabilities, or price, and everything to do with the highly variable nature of the oxide layer. -Chuck, WA3UQV OK, I'll grab some of those old Ducatis and check 'em. I'm sure they haven't seen a polorizing voltage in at least 25 years, because that's when I bought them as surplus. Then I'll run 'em up to their rated voltage for 24 hours and recheck. I'll see how much extra voltage they take and hold them at that voltage for another 24 hours. Just for curiosity, I'll also check the ESR at each step with my Dick Smith meter. I know none of this is lab quality procedure, but if there's any gross changes, I think I'll catch 'em. Frank Dresser |
--exray-- wrote:
Chuck Harris wrote: Electrolytic capacitors are self adjusting for working voltage (to some degree). If they are operated for a long time at 50% of their rating, the oxide reduces in thickness, and they become higher capacitance, and lower working voltage. If you try to increase their operating voltage, they will draw too much current. They will either adapt to the new higher voltage, or they will blow up from the heat. Thats very much true with older caps up to about 1970. But later model caps don't exhibit this 'memory'. It is all about the electrolyte. The older electrolytes tended to eat the oxide layer pretty quickly. The manufacturers rated them for a 1 to 2 year shelf life... longer if they were in continuous use. There have been alot of improvements in the electrolytes, and now the caps last virtually forever. But the oxide thickness still determines the tolerance, and as such it still changes with temperature, age and voltage. Just not as much as it used to. -Chuck -Bill |
--exray-- wrote:
Chuck Harris wrote: Electrolytic capacitors are self adjusting for working voltage (to some degree). If they are operated for a long time at 50% of their rating, the oxide reduces in thickness, and they become higher capacitance, and lower working voltage. If you try to increase their operating voltage, they will draw too much current. They will either adapt to the new higher voltage, or they will blow up from the heat. Thats very much true with older caps up to about 1970. But later model caps don't exhibit this 'memory'. It is all about the electrolyte. The older electrolytes tended to eat the oxide layer pretty quickly. The manufacturers rated them for a 1 to 2 year shelf life... longer if they were in continuous use. There have been alot of improvements in the electrolytes, and now the caps last virtually forever. But the oxide thickness still determines the tolerance, and as such it still changes with temperature, age and voltage. Just not as much as it used to. -Chuck -Bill |
Frank Dresser wrote:
OK, I'll grab some of those old Ducatis and check 'em. I'm sure they haven't seen a polorizing voltage in at least 25 years, because that's when I bought them as surplus. Then I'll run 'em up to their rated voltage for 24 hours and recheck. I'll see how much extra voltage they take and hold them at that voltage for another 24 hours. Just for curiosity, I'll also check the ESR at each step with my Dick Smith meter. I know none of this is lab quality procedure, but if there's any gross changes, I think I'll catch 'em. Frank Dresser That would be fine if you are looking to get doused with electrolyte. A better test would be to measure the capacitance as they sit. Then reform them with a 1.5K resistor in series with the supply. Then retake the measurements. If the cap isn't drawing current during the reform, it means the maker got the electrolyte formulation right, you probably won't see much change in measured characteristics. If the cap is drawing heavy current during the reform, you should see greater differences in the reformed cap vs the "NOS" cap. ESR should go down, capacitance should go down, and so should leakage current. -Chuck, WA3UQV |
Frank Dresser wrote:
OK, I'll grab some of those old Ducatis and check 'em. I'm sure they haven't seen a polorizing voltage in at least 25 years, because that's when I bought them as surplus. Then I'll run 'em up to their rated voltage for 24 hours and recheck. I'll see how much extra voltage they take and hold them at that voltage for another 24 hours. Just for curiosity, I'll also check the ESR at each step with my Dick Smith meter. I know none of this is lab quality procedure, but if there's any gross changes, I think I'll catch 'em. Frank Dresser That would be fine if you are looking to get doused with electrolyte. A better test would be to measure the capacitance as they sit. Then reform them with a 1.5K resistor in series with the supply. Then retake the measurements. If the cap isn't drawing current during the reform, it means the maker got the electrolyte formulation right, you probably won't see much change in measured characteristics. If the cap is drawing heavy current during the reform, you should see greater differences in the reformed cap vs the "NOS" cap. ESR should go down, capacitance should go down, and so should leakage current. -Chuck, WA3UQV |
Hi,
Thats very much true with older caps up to about 1970. But later model caps don't exhibit this 'memory'. I'd make that "1950." There have been alot of improvements in the electrolytes, and now the caps last virtually forever. But the oxide thickness still determines the tolerance, and as such it still changes with temperature, age and voltage. Just not as much as it used to. Funny, I've measured a dozen caps before and after reforming, new old stock from 1946 to 1997. Other than the 1946 one, which dropped from 17 to 12.1 µF, all the others *increased* their capacitance. That includes ones from 1947, 1962, and 1967. 73, Alan |
Hi,
Thats very much true with older caps up to about 1970. But later model caps don't exhibit this 'memory'. I'd make that "1950." There have been alot of improvements in the electrolytes, and now the caps last virtually forever. But the oxide thickness still determines the tolerance, and as such it still changes with temperature, age and voltage. Just not as much as it used to. Funny, I've measured a dozen caps before and after reforming, new old stock from 1946 to 1997. Other than the 1946 one, which dropped from 17 to 12.1 µF, all the others *increased* their capacitance. That includes ones from 1947, 1962, and 1967. 73, Alan |
"Chuck Harris" wrote in message ... That would be fine if you are looking to get doused with electrolyte. A better test would be to measure the capacitance as they sit. Then reform them with a 1.5K resistor in series with the supply. Then retake the measurements. I have a Heathkit IT-28 capacitor checker. I'll use that to run up the voltage on the caps. The eye tube indication should keep me from overstressing the caps. I hope we can handle it! I'm really curious about the magnitude of the capacitance change with voltage, so I've decided to let the 25V caps cook at 15V for another data point. I've checked caps before and after running them up to their rated voltage, even some 50 year old ones, and I've never noticed a big difference in either capacitance or ESR. Well, this time I'm going to pay attention! I have blown caps with the Heathkit cap checker, but never from overheating. I have turned the voltage control up way too high, and arced them internally. If the cap isn't drawing current during the reform, it means the maker got the electrolyte formulation right, you probably won't see much change in measured characteristics. Does that mean that the manufacturers learned how to make a stable, predictable electrolytic capacitor? If they can do that much, why can't they manage to make them with precision? How long would it take a right electrolyte capacitor to "unform"? Just when did the manufacturers get the electrolyte formula right? If the cap is drawing heavy current during the reform, you should see greater differences in the reformed cap vs the "NOS" cap. ESR should go down, capacitance should go down, and so should leakage current. -Chuck, WA3UQV I've numbered up all 11 of my 25+ year-old Ducatis and checked them for capacitance and ESR. I'm sure nearly all of these things have no voltage applied since I bought them at Olson Electronics for a penny apiece, back around 1978. I've got them paralled with clip leads and I'm letting them form to 15V on the Heathkit. We'll see what we get tomorrow night. Before I powered up all the caps, I selected one of the old Ducatis and ran it up to 25V, just to see what would happen. It took about a minute to come to a low leakage point. Befo 99ufd 0.22 ohm ESR After: 100 ufd 0.20 ohm ESR It's worth mentioning that the Heathkit cap checker marks off the capacitance every 10ufd in that part of the scale. So most of these numbers are estimates. But in this case, the distinction is real. The indicator moved from the edge of the 100 ufd line to right in the middle. However, I don't know if the distinction is important. It might just be a temperature effect. The ESR is down 10% but I can't make much from that either, especially since it's the last digit. Frank Dresser |
"Chuck Harris" wrote in message ... That would be fine if you are looking to get doused with electrolyte. A better test would be to measure the capacitance as they sit. Then reform them with a 1.5K resistor in series with the supply. Then retake the measurements. I have a Heathkit IT-28 capacitor checker. I'll use that to run up the voltage on the caps. The eye tube indication should keep me from overstressing the caps. I hope we can handle it! I'm really curious about the magnitude of the capacitance change with voltage, so I've decided to let the 25V caps cook at 15V for another data point. I've checked caps before and after running them up to their rated voltage, even some 50 year old ones, and I've never noticed a big difference in either capacitance or ESR. Well, this time I'm going to pay attention! I have blown caps with the Heathkit cap checker, but never from overheating. I have turned the voltage control up way too high, and arced them internally. If the cap isn't drawing current during the reform, it means the maker got the electrolyte formulation right, you probably won't see much change in measured characteristics. Does that mean that the manufacturers learned how to make a stable, predictable electrolytic capacitor? If they can do that much, why can't they manage to make them with precision? How long would it take a right electrolyte capacitor to "unform"? Just when did the manufacturers get the electrolyte formula right? If the cap is drawing heavy current during the reform, you should see greater differences in the reformed cap vs the "NOS" cap. ESR should go down, capacitance should go down, and so should leakage current. -Chuck, WA3UQV I've numbered up all 11 of my 25+ year-old Ducatis and checked them for capacitance and ESR. I'm sure nearly all of these things have no voltage applied since I bought them at Olson Electronics for a penny apiece, back around 1978. I've got them paralled with clip leads and I'm letting them form to 15V on the Heathkit. We'll see what we get tomorrow night. Before I powered up all the caps, I selected one of the old Ducatis and ran it up to 25V, just to see what would happen. It took about a minute to come to a low leakage point. Befo 99ufd 0.22 ohm ESR After: 100 ufd 0.20 ohm ESR It's worth mentioning that the Heathkit cap checker marks off the capacitance every 10ufd in that part of the scale. So most of these numbers are estimates. But in this case, the distinction is real. The indicator moved from the edge of the 100 ufd line to right in the middle. However, I don't know if the distinction is important. It might just be a temperature effect. The ESR is down 10% but I can't make much from that either, especially since it's the last digit. Frank Dresser |
"Frank Dresser" wrote in message ... .. I'm sure nearly all of these things have no voltage applied since I bought them at Olson Electronics for a penny apiece, back around 1978. Oops! I meant polarizing voltage. I've already checked them for capacitance and ESR, but that's AC. Frank Dresser |
"Frank Dresser" wrote in message ... .. I'm sure nearly all of these things have no voltage applied since I bought them at Olson Electronics for a penny apiece, back around 1978. Oops! I meant polarizing voltage. I've already checked them for capacitance and ESR, but that's AC. Frank Dresser |
"Chuck Harris" wrote in message ... Here's the numbers from the cap tests. The format is cap number, ESR in ohms, capacitance in ufd. Column 1, out of the drawer after sitting unused for maybe 25 years. Coulmn 2, after 24 hours with 15V applied. 1 .19, 100 .18, 105 2 .22, 100 .20, 100 3 .21, 105 .20, 100 4 .20, 110 .19, 105 5 .21, 115 .20, 105 6 .21, 109 .20, 102 7 .24, 103 .22, 098 8 .23, 098 .21, 098 9 .16, 112 .16, 112 10 .21, 100 .20, 100 11 .22, 099 .21, 098 I don't notice any big changes. Capacitance and ESR may be down a bit. Or it may be a temperature effect. I've now got them all sitting at their rated voltage of 25V. Frank Dresser |
"Chuck Harris" wrote in message ... Here's the numbers from the cap tests. The format is cap number, ESR in ohms, capacitance in ufd. Column 1, out of the drawer after sitting unused for maybe 25 years. Coulmn 2, after 24 hours with 15V applied. 1 .19, 100 .18, 105 2 .22, 100 .20, 100 3 .21, 105 .20, 100 4 .20, 110 .19, 105 5 .21, 115 .20, 105 6 .21, 109 .20, 102 7 .24, 103 .22, 098 8 .23, 098 .21, 098 9 .16, 112 .16, 112 10 .21, 100 .20, 100 11 .22, 099 .21, 098 I don't notice any big changes. Capacitance and ESR may be down a bit. Or it may be a temperature effect. I've now got them all sitting at their rated voltage of 25V. Frank Dresser |
Here's the numbers from the cap tests. The format is cap number, ESR
in ohms, capacitance in ufd. Column 1, out of the drawer after sitting unused for maybe 25 years. Coulmn 2, after 24 hours with 15V applied. Column 3, after 24 hours with 25V applied. 1 .19, 100 .18, 105 .19, 100 2 .22, 100 .20, 100 .22, 100 3 .21, 105 .20, 100 .22, 105 4 .20, 110 .19, 105 .21, 110 5 .21, 115 .20, 105 .22, 115 6 .21, 109 .20, 102 .22, 100 7 .24, 103 .22, 098 .25, 100 8 .23, 098 .21, 098 .24, 098 9 .16, 112 .16, 112 .18, 112 10 .21, 100 .20, 100 .22, 100 11 .22, 099 .21, 098 .23, 098 Still nothing I'd call a significant change. Also, it's worth noting that the ESR meter zero point is another source of small errors. They're at 30V right now. Frank Dresser |
Here's the numbers from the cap tests. The format is cap number, ESR
in ohms, capacitance in ufd. Column 1, out of the drawer after sitting unused for maybe 25 years. Coulmn 2, after 24 hours with 15V applied. Column 3, after 24 hours with 25V applied. 1 .19, 100 .18, 105 .19, 100 2 .22, 100 .20, 100 .22, 100 3 .21, 105 .20, 100 .22, 105 4 .20, 110 .19, 105 .21, 110 5 .21, 115 .20, 105 .22, 115 6 .21, 109 .20, 102 .22, 100 7 .24, 103 .22, 098 .25, 100 8 .23, 098 .21, 098 .24, 098 9 .16, 112 .16, 112 .18, 112 10 .21, 100 .20, 100 .22, 100 11 .22, 099 .21, 098 .23, 098 Still nothing I'd call a significant change. Also, it's worth noting that the ESR meter zero point is another source of small errors. They're at 30V right now. Frank Dresser |
"Chuck Harris" wrote in message ... That would be fine if you are looking to get doused with electrolyte. A better test would be to measure the capacitance as they sit. Then reform them with a 1.5K resistor in series with the supply. Then retake the measurements. No dousing and no noticable heat, either. I did ESR and capacitance checks with the old caps as they came out of the drawer, at 15V for 24 hours, at 25V at 24 hours, and 30V at 24 hours. I had to use a different power supply to get 30V. The Heathkit checker takes a large step from 25V to 50V. I didn't want to deal with limiting the checker's voltage. If the cap isn't drawing current during the reform, it means the maker got the electrolyte formulation right, you probably won't see much change in measured characteristics. If the cap is drawing heavy current during the reform, you should see greater differences in the reformed cap vs the "NOS" cap. ESR should go down, capacitance should go down, and so should leakage current. -Chuck, WA3UQV I think we can both say with confidence that Ducati got the electrolyte formulation right, way back two or three decades ago. I didn't measure any significant change in capacitance or ESR (except ESR went up on a couple on the last test). If they had a +/- 20% tolerance, they were in spec at the start and stayed in spec at every step of the test. But I never noticed a big change when I've checked electrolytics before. Here's the numbers from the cap tests. The format is cap number, ESR in ohms, capacitance in ufd. Column 1, out of the drawer after sitting unused for maybe 25 years. Coulmn 2, after 24 hours with 15V applied. Column 3, after 24 hours with 25V applied. Column 4, after 24 hours with 30V applied. 1 .19, 100 .18, 105 .19, 100 .20, 105 2 .22, 100 .20, 100 .22, 100 .22, 105 3 .21, 105 .20, 100 .22, 105 .23, 103 4 .20, 110 .19, 105 .21, 110 .30, 108 5 .21, 115 .20, 105 .22, 115 .33, 110 6 .21, 109 .20, 102 .22, 100 .24, 110 7 .24, 103 .22, 098 .25, 100 .25, 110 8 .23, 098 .21, 098 .24, 098 .23, 100 9 .16, 112 .16, 112 .18, 112 .20, 115 10 .21, 100 .20, 100 .22, 100 .20, 115 11 .22, 099 .21, 098 .23, 098 .23, 105 There might be a couple of trends. Both capacitance and ESR went up on a few. I can't make much of that because of possible measurement errors. There are no gross changes off the shelf and with voltage applied. These old caps tested with pretty good precision. The newer ones test in even tighter groups. Frank Dresser |
"Chuck Harris" wrote in message ... That would be fine if you are looking to get doused with electrolyte. A better test would be to measure the capacitance as they sit. Then reform them with a 1.5K resistor in series with the supply. Then retake the measurements. No dousing and no noticable heat, either. I did ESR and capacitance checks with the old caps as they came out of the drawer, at 15V for 24 hours, at 25V at 24 hours, and 30V at 24 hours. I had to use a different power supply to get 30V. The Heathkit checker takes a large step from 25V to 50V. I didn't want to deal with limiting the checker's voltage. If the cap isn't drawing current during the reform, it means the maker got the electrolyte formulation right, you probably won't see much change in measured characteristics. If the cap is drawing heavy current during the reform, you should see greater differences in the reformed cap vs the "NOS" cap. ESR should go down, capacitance should go down, and so should leakage current. -Chuck, WA3UQV I think we can both say with confidence that Ducati got the electrolyte formulation right, way back two or three decades ago. I didn't measure any significant change in capacitance or ESR (except ESR went up on a couple on the last test). If they had a +/- 20% tolerance, they were in spec at the start and stayed in spec at every step of the test. But I never noticed a big change when I've checked electrolytics before. Here's the numbers from the cap tests. The format is cap number, ESR in ohms, capacitance in ufd. Column 1, out of the drawer after sitting unused for maybe 25 years. Coulmn 2, after 24 hours with 15V applied. Column 3, after 24 hours with 25V applied. Column 4, after 24 hours with 30V applied. 1 .19, 100 .18, 105 .19, 100 .20, 105 2 .22, 100 .20, 100 .22, 100 .22, 105 3 .21, 105 .20, 100 .22, 105 .23, 103 4 .20, 110 .19, 105 .21, 110 .30, 108 5 .21, 115 .20, 105 .22, 115 .33, 110 6 .21, 109 .20, 102 .22, 100 .24, 110 7 .24, 103 .22, 098 .25, 100 .25, 110 8 .23, 098 .21, 098 .24, 098 .23, 100 9 .16, 112 .16, 112 .18, 112 .20, 115 10 .21, 100 .20, 100 .22, 100 .20, 115 11 .22, 099 .21, 098 .23, 098 .23, 105 There might be a couple of trends. Both capacitance and ESR went up on a few. I can't make much of that because of possible measurement errors. There are no gross changes off the shelf and with voltage applied. These old caps tested with pretty good precision. The newer ones test in even tighter groups. Frank Dresser |
Hi,
Frank wrote: I think we can both say with confidence that Ducati got the electrolyte formulation right, way back two or three decades ago. I didn't measure any significant change in capacitance or ESR (except ESR went up on a couple on the last test). If they had a +/- 20% tolerance, they were in spec at the start and stayed in spec at every step of the test. But I never noticed a big change when I've checked electrolytics before. I didn't find any big changes either, measuring high-voltage electrolytics from as far back as 1947. Looking at your data, within the limits of experimental error I don't see any changes at all. That pretty much kills the myth about capacitors reforming to a new working voltage and changing the oxide-layer thickness. However I'm sure we haven't seen the last of it. I believe there's still a website with quotes from Deeley's book on electrolytic capacitors. What it doesn't say is that this book was published in 1939. 73, Alan |
Hi,
Frank wrote: I think we can both say with confidence that Ducati got the electrolyte formulation right, way back two or three decades ago. I didn't measure any significant change in capacitance or ESR (except ESR went up on a couple on the last test). If they had a +/- 20% tolerance, they were in spec at the start and stayed in spec at every step of the test. But I never noticed a big change when I've checked electrolytics before. I didn't find any big changes either, measuring high-voltage electrolytics from as far back as 1947. Looking at your data, within the limits of experimental error I don't see any changes at all. That pretty much kills the myth about capacitors reforming to a new working voltage and changing the oxide-layer thickness. However I'm sure we haven't seen the last of it. I believe there's still a website with quotes from Deeley's book on electrolytic capacitors. What it doesn't say is that this book was published in 1939. 73, Alan |
Ok Alan,
What do you imagine is happening when you reform an old electrolytic? Let's go back once again to an old industry reference on electrolytics. The 1968, "Reference Data For Radio Engineers", published by ITT. The articles in it are authored by experts from academia, and industry. First, a useful definition that provides a clue: ----------------------------------------------- Forming Voltage (electrolytics): The voltage at which the anodic oxide has been formed. The thickness of the oxide layer is proportional to this voltage. Next a description: ------------------ Aluminum Electrolytics This is the most widly know electrolytic capacitor and is used extensively in radio and television equipment. It has a space factor about 6 times better than the equivalent paper capacitor. Types of improved reliability are now available using high-purity (better than 99.99%) aluminum. Conventional aluminum electrolytic capacitors which have gone 6 months or more without voltage applied may need to be reformed. Rated voltage is applied from a dc source with an internal resistance of 1500 ohms for capacitors with a rated voltage exceeding 100 volts, or 150 ohms for capacitors with a rated voltage equal to or less than 100 volts. The voltage must be applied for one hour after reaching rated value with a tolerance of +/- 3 percent. The capacitor is then discharged through a resistor of 1 ohm/volt. [by rated value, they mean, of course, rated voltage - cfh] Now a description of the construction of electrolytic capacitors: ------------------------------------------------------------------------- ELECTROLYTIC CAPACITORS Electrolytic capacitors (Fig 11) employ for at least one of their electrodes a "valve metal". This metal when operated in an electrolytic cell as the [My attempt at an ascii figure 11 - cfh] - + O-----|C|. . . . . . . .|O|A|-----O |A| . . . . . .|X|N| |T| Conducting .. .. .. .|I|O| |H| Electrolyte . . .|D|D| |O|. . . . . . . .|E|E| |D| . . . . . .| | | |E|. . . . . . .| | | O---------SERIES-R---------+---)|-------+---O | | +-Leakage R--+ Note the series R is caused by the leads and electrolyte, etc. the Leakage R is caused by the oxide layer. FIG. 11 anode, forms a layer of dielectric oxide. The most commonly used metals are aluminum or tantalum. The valve metal behavior of these metals was know about 1850. Tantalum electrolytic capacitors were introduced in the 1950's because of the need for highly reliable miniature capacitors in transistor circuits over a wide temperature range. These capacitors were made possible by improved refining and powder metallurgy techniques. The term "electrolytic capacitor" is applied to any capacitor in which the dielectric layer is formed by an electrolytic method. The capacitor does not necessarily contain an electrolyte. The oxide layer is formed by placing the metal in a bath containing a suitable forming electrolyte, and applying voltage between the metal as anode and another electrode as cathode. The oxide grows at a rate determined by the current flowing, but this rate of grwoth decreases until the oxide has reached a limiting thickness determined by the voltage. For most practical purposes it may be assumed that the thickness of the oxide is proportional to the forming voltage. [ a figure showing some characteristics of oxides...] The structure of these oxide layers plays an important part in determining their performance. Ideally they are amorphous but aluminum tends to form two distinct layers, the outer one being porous. Tantalum normally forms an amorphous oxide which, under conditions of a high field strength of the oxide layer, may become crystaline. Depending on the forming electrolyte and the surface condition of the metal, there is an upper limit of voltage beyond which the oxide breaks down. The working voltage is between 25 and 90 percent (according to type) of the forming voltage at which stable operation of the oxide layer can be obtained. To produce a capacitor it is necessary to make contact to the oxide layer on the anode, and there are two distinct methods of doing this. The first is to use a working electrolyte that has sufficient conductivity over the temperature range to give a good power factor. There are many considerations in choosing the working electrolyte, and the choice is usually a compromise between high and low temperature performance. The working electrolyte also provides a rehealing feature in that any faults in the oxide layer will be repaired by further anodization. In aluminum electrolytic capacitors the working electrolyte must be restricted to those materials in which aluminum and its oxide are inert. Corrosion can be minimized by using the highest possible purity of aluminum. This also reduces the tendency of the oxide layer to dissolve in the electrolyte, giving a better shelf life. ------------------------------------------------------------------------ Notice three things about the above dissertation: 1) The oxide layer is formed before assembly, and its thickness determines both the working voltage, and the capacitance. 2) The working electrolyte makes contact to the oxide layer, and also performs a rehealing feature that repairs any faults in the oxide layer by reanodizing the layer. 3) The working electrolyte will dissolve the oxide layer over time if no voltage is applied to the capacitor. This is the cause of shelf life. I believe this discription. I have personally witnessed it many many times with old aluminum electrolytics (with wet working electrolytes). You can continue to think it a myth, but it isn't. -Chuck Harris, WA3UQV Alan Douglas wrote: Hi, Frank wrote: I think we can both say with confidence that Ducati got the electrolyte formulation right, way back two or three decades ago. I didn't measure any significant change in capacitance or ESR (except ESR went up on a couple on the last test). If they had a +/- 20% tolerance, they were in spec at the start and stayed in spec at every step of the test. But I never noticed a big change when I've checked electrolytics before. I didn't find any big changes either, measuring high-voltage electrolytics from as far back as 1947. Looking at your data, within the limits of experimental error I don't see any changes at all. That pretty much kills the myth about capacitors reforming to a new working voltage and changing the oxide-layer thickness. However I'm sure we haven't seen the last of it. I believe there's still a website with quotes from Deeley's book on electrolytic capacitors. What it doesn't say is that this book was published in 1939. 73, Alan |
Ok Alan,
What do you imagine is happening when you reform an old electrolytic? Let's go back once again to an old industry reference on electrolytics. The 1968, "Reference Data For Radio Engineers", published by ITT. The articles in it are authored by experts from academia, and industry. First, a useful definition that provides a clue: ----------------------------------------------- Forming Voltage (electrolytics): The voltage at which the anodic oxide has been formed. The thickness of the oxide layer is proportional to this voltage. Next a description: ------------------ Aluminum Electrolytics This is the most widly know electrolytic capacitor and is used extensively in radio and television equipment. It has a space factor about 6 times better than the equivalent paper capacitor. Types of improved reliability are now available using high-purity (better than 99.99%) aluminum. Conventional aluminum electrolytic capacitors which have gone 6 months or more without voltage applied may need to be reformed. Rated voltage is applied from a dc source with an internal resistance of 1500 ohms for capacitors with a rated voltage exceeding 100 volts, or 150 ohms for capacitors with a rated voltage equal to or less than 100 volts. The voltage must be applied for one hour after reaching rated value with a tolerance of +/- 3 percent. The capacitor is then discharged through a resistor of 1 ohm/volt. [by rated value, they mean, of course, rated voltage - cfh] Now a description of the construction of electrolytic capacitors: ------------------------------------------------------------------------- ELECTROLYTIC CAPACITORS Electrolytic capacitors (Fig 11) employ for at least one of their electrodes a "valve metal". This metal when operated in an electrolytic cell as the [My attempt at an ascii figure 11 - cfh] - + O-----|C|. . . . . . . .|O|A|-----O |A| . . . . . .|X|N| |T| Conducting .. .. .. .|I|O| |H| Electrolyte . . .|D|D| |O|. . . . . . . .|E|E| |D| . . . . . .| | | |E|. . . . . . .| | | O---------SERIES-R---------+---)|-------+---O | | +-Leakage R--+ Note the series R is caused by the leads and electrolyte, etc. the Leakage R is caused by the oxide layer. FIG. 11 anode, forms a layer of dielectric oxide. The most commonly used metals are aluminum or tantalum. The valve metal behavior of these metals was know about 1850. Tantalum electrolytic capacitors were introduced in the 1950's because of the need for highly reliable miniature capacitors in transistor circuits over a wide temperature range. These capacitors were made possible by improved refining and powder metallurgy techniques. The term "electrolytic capacitor" is applied to any capacitor in which the dielectric layer is formed by an electrolytic method. The capacitor does not necessarily contain an electrolyte. The oxide layer is formed by placing the metal in a bath containing a suitable forming electrolyte, and applying voltage between the metal as anode and another electrode as cathode. The oxide grows at a rate determined by the current flowing, but this rate of grwoth decreases until the oxide has reached a limiting thickness determined by the voltage. For most practical purposes it may be assumed that the thickness of the oxide is proportional to the forming voltage. [ a figure showing some characteristics of oxides...] The structure of these oxide layers plays an important part in determining their performance. Ideally they are amorphous but aluminum tends to form two distinct layers, the outer one being porous. Tantalum normally forms an amorphous oxide which, under conditions of a high field strength of the oxide layer, may become crystaline. Depending on the forming electrolyte and the surface condition of the metal, there is an upper limit of voltage beyond which the oxide breaks down. The working voltage is between 25 and 90 percent (according to type) of the forming voltage at which stable operation of the oxide layer can be obtained. To produce a capacitor it is necessary to make contact to the oxide layer on the anode, and there are two distinct methods of doing this. The first is to use a working electrolyte that has sufficient conductivity over the temperature range to give a good power factor. There are many considerations in choosing the working electrolyte, and the choice is usually a compromise between high and low temperature performance. The working electrolyte also provides a rehealing feature in that any faults in the oxide layer will be repaired by further anodization. In aluminum electrolytic capacitors the working electrolyte must be restricted to those materials in which aluminum and its oxide are inert. Corrosion can be minimized by using the highest possible purity of aluminum. This also reduces the tendency of the oxide layer to dissolve in the electrolyte, giving a better shelf life. ------------------------------------------------------------------------ Notice three things about the above dissertation: 1) The oxide layer is formed before assembly, and its thickness determines both the working voltage, and the capacitance. 2) The working electrolyte makes contact to the oxide layer, and also performs a rehealing feature that repairs any faults in the oxide layer by reanodizing the layer. 3) The working electrolyte will dissolve the oxide layer over time if no voltage is applied to the capacitor. This is the cause of shelf life. I believe this discription. I have personally witnessed it many many times with old aluminum electrolytics (with wet working electrolytes). You can continue to think it a myth, but it isn't. -Chuck Harris, WA3UQV Alan Douglas wrote: Hi, Frank wrote: I think we can both say with confidence that Ducati got the electrolyte formulation right, way back two or three decades ago. I didn't measure any significant change in capacitance or ESR (except ESR went up on a couple on the last test). If they had a +/- 20% tolerance, they were in spec at the start and stayed in spec at every step of the test. But I never noticed a big change when I've checked electrolytics before. I didn't find any big changes either, measuring high-voltage electrolytics from as far back as 1947. Looking at your data, within the limits of experimental error I don't see any changes at all. That pretty much kills the myth about capacitors reforming to a new working voltage and changing the oxide-layer thickness. However I'm sure we haven't seen the last of it. I believe there's still a website with quotes from Deeley's book on electrolytic capacitors. What it doesn't say is that this book was published in 1939. 73, Alan |
"Chuck Harris" wrote in message ... [snip] Notice three things about the above dissertation: 1) The oxide layer is formed before assembly, and its thickness determines both the working voltage, and the capacitance. 2) The working electrolyte makes contact to the oxide layer, and also performs a rehealing feature that repairs any faults in the oxide layer by reanodizing the layer. 3) The working electrolyte will dissolve the oxide layer over time if no voltage is applied to the capacitor. This is the cause of shelf life. I believe this discription. I have personally witnessed it many many times with old aluminum electrolytics (with wet working electrolytes). You can continue to think it a myth, but it isn't. -Chuck Harris, WA3UQV Which "it" is the myth? If we are discussing the idea that the capacitance of electrolytic capacitors changes in a significant way with operating voltage and long term storage, shouldn't it be very evident? Have there been any graphs published showing how capacitance changes with operating voltage or storage time? Frank Dresser |
"Chuck Harris" wrote in message ... [snip] Notice three things about the above dissertation: 1) The oxide layer is formed before assembly, and its thickness determines both the working voltage, and the capacitance. 2) The working electrolyte makes contact to the oxide layer, and also performs a rehealing feature that repairs any faults in the oxide layer by reanodizing the layer. 3) The working electrolyte will dissolve the oxide layer over time if no voltage is applied to the capacitor. This is the cause of shelf life. I believe this discription. I have personally witnessed it many many times with old aluminum electrolytics (with wet working electrolytes). You can continue to think it a myth, but it isn't. -Chuck Harris, WA3UQV Which "it" is the myth? If we are discussing the idea that the capacitance of electrolytic capacitors changes in a significant way with operating voltage and long term storage, shouldn't it be very evident? Have there been any graphs published showing how capacitance changes with operating voltage or storage time? Frank Dresser |
"Alan Douglas" adouglasatgis.net wrote in message ... I didn't find any big changes either, measuring high-voltage electrolytics from as far back as 1947. Looking at your data, within the limits of experimental error I don't see any changes at all. That pretty much kills the myth about capacitors reforming to a new working voltage and changing the oxide-layer thickness. However I'm sure we haven't seen the last of it. I believe there's still a website with quotes from Deeley's book on electrolytic capacitors. What it doesn't say is that this book was published in 1939. 73, Alan One site has much of the Electrolytic Capacitors book online, but it does carry a disclaimer: "Great strides have been made in all aspects of capacitor technology since Electrolytic Capacitors was published in 1938. Most of the material presented here at this time is directly from the original edition. The reader should therefore be cautious in regards to the technical veracity of any claims herein. " This is from: http://www.faradnet.com/deeley/book_toc.htm Frank Dresser |
"Alan Douglas" adouglasatgis.net wrote in message ... I didn't find any big changes either, measuring high-voltage electrolytics from as far back as 1947. Looking at your data, within the limits of experimental error I don't see any changes at all. That pretty much kills the myth about capacitors reforming to a new working voltage and changing the oxide-layer thickness. However I'm sure we haven't seen the last of it. I believe there's still a website with quotes from Deeley's book on electrolytic capacitors. What it doesn't say is that this book was published in 1939. 73, Alan One site has much of the Electrolytic Capacitors book online, but it does carry a disclaimer: "Great strides have been made in all aspects of capacitor technology since Electrolytic Capacitors was published in 1938. Most of the material presented here at this time is directly from the original edition. The reader should therefore be cautious in regards to the technical veracity of any claims herein. " This is from: http://www.faradnet.com/deeley/book_toc.htm Frank Dresser |
Hi,
I would guess that "re-forming" an electrolytic is actually reforming only the defects in the oxide film. At least that would explain why it works, but doesn't affect overall capacitance. I've only found one capacitor that decreased its value significantly after reforming, an Aerovox 8µF 450V unit dated 1946. That started out at 17µF and decreased to 12.1µF after 12 hours, but still showed 0.46mA leakage. All others, including a pair of NOS *wet* Sprague 16µF 450V caps, showed little if any change. The Spragues by the way ended up at 4mA leakage, so I don't think I'll be using them in a radio. Cheers, Alan |
Hi,
I would guess that "re-forming" an electrolytic is actually reforming only the defects in the oxide film. At least that would explain why it works, but doesn't affect overall capacitance. I've only found one capacitor that decreased its value significantly after reforming, an Aerovox 8µF 450V unit dated 1946. That started out at 17µF and decreased to 12.1µF after 12 hours, but still showed 0.46mA leakage. All others, including a pair of NOS *wet* Sprague 16µF 450V caps, showed little if any change. The Spragues by the way ended up at 4mA leakage, so I don't think I'll be using them in a radio. Cheers, Alan |
Hi,
Frank wrote: One site has much of the Electrolytic Capacitors book online, but it does carry a disclaimer: http://www.faradnet.com/deeley/book_toc.htm Yes, you're right, I just checked that site in 1997 (www.archive.org) and that does have the same preface. Nothing seems to have been added, however, and the present site says it was last updated in 2000. Cheers, Alan |
Hi,
Frank wrote: One site has much of the Electrolytic Capacitors book online, but it does carry a disclaimer: http://www.faradnet.com/deeley/book_toc.htm Yes, you're right, I just checked that site in 1997 (www.archive.org) and that does have the same preface. Nothing seems to have been added, however, and the present site says it was last updated in 2000. Cheers, Alan |
Hi Alan,
Because of the physical construction of an electrolytic cap, it MUST change capacitance if the oxide grows thinner in storage, or thickens thru reforming... But, I too notice that sometimes the change is large, and othertimes it is not. I suspect that what is happening is the oxide layer thins out only in spots (probably around impurities) in some caps. These spots are large enough to readily affect the leakage current, but are small with respect to the total surface area of the plates. Because they are a small percentage of the total surface area, they only minimally affect the total capacitance. -Chuck Harris Alan Douglas wrote: Hi, I would guess that "re-forming" an electrolytic is actually reforming only the defects in the oxide film. At least that would explain why it works, but doesn't affect overall capacitance. I've only found one capacitor that decreased its value significantly after reforming, an Aerovox 8µF 450V unit dated 1946. That started out at 17µF and decreased to 12.1µF after 12 hours, but still showed 0.46mA leakage. All others, including a pair of NOS *wet* Sprague 16µF 450V caps, showed little if any change. The Spragues by the way ended up at 4mA leakage, so I don't think I'll be using them in a radio. Cheers, Alan |
Hi Alan,
Because of the physical construction of an electrolytic cap, it MUST change capacitance if the oxide grows thinner in storage, or thickens thru reforming... But, I too notice that sometimes the change is large, and othertimes it is not. I suspect that what is happening is the oxide layer thins out only in spots (probably around impurities) in some caps. These spots are large enough to readily affect the leakage current, but are small with respect to the total surface area of the plates. Because they are a small percentage of the total surface area, they only minimally affect the total capacitance. -Chuck Harris Alan Douglas wrote: Hi, I would guess that "re-forming" an electrolytic is actually reforming only the defects in the oxide film. At least that would explain why it works, but doesn't affect overall capacitance. I've only found one capacitor that decreased its value significantly after reforming, an Aerovox 8µF 450V unit dated 1946. That started out at 17µF and decreased to 12.1µF after 12 hours, but still showed 0.46mA leakage. All others, including a pair of NOS *wet* Sprague 16µF 450V caps, showed little if any change. The Spragues by the way ended up at 4mA leakage, so I don't think I'll be using them in a radio. Cheers, Alan |
Chuck Harris wrote:
Hi Alan, Because of the physical construction of an electrolytic cap, it MUST change capacitance if the oxide grows thinner in storage, or thickens thru reforming... But, I too notice that sometimes the change is large, and othertimes it is not. I suspect that what is happening is the oxide layer thins out only in spots (probably around impurities) in some caps. These spots are large enough to readily affect the leakage current, but are small with respect to the total surface area of the plates. Because they are a small percentage of the total surface area, they only minimally affect the total capacitance. -Chuck Harris Let me pose a question...not knowing inimately how electrolytics were or are made. Seems to me that the 'extra' oxide, ie thicker plates, are taking up some of the physical space that was formerly the electrolyte (part of the dielectric, so to speak) thereby leaving the plates closer together. That would indicate more oxide=more capacitance. In the case of thin oxide (not holes) I don't see how thick or thin would relate to leakage as long as there was something there. Running with the same thought, if the oxide is totaaly absent what makes it redeposit on the wax paper/mylar? Maybe the leakage is a result of the metallic compounds being absorbed by the electrolyte and the reforming process sends them back to the original metal, albeit somewhat randomly. Does this make any sense? -Bill |
Chuck Harris wrote:
Hi Alan, Because of the physical construction of an electrolytic cap, it MUST change capacitance if the oxide grows thinner in storage, or thickens thru reforming... But, I too notice that sometimes the change is large, and othertimes it is not. I suspect that what is happening is the oxide layer thins out only in spots (probably around impurities) in some caps. These spots are large enough to readily affect the leakage current, but are small with respect to the total surface area of the plates. Because they are a small percentage of the total surface area, they only minimally affect the total capacitance. -Chuck Harris Let me pose a question...not knowing inimately how electrolytics were or are made. Seems to me that the 'extra' oxide, ie thicker plates, are taking up some of the physical space that was formerly the electrolyte (part of the dielectric, so to speak) thereby leaving the plates closer together. That would indicate more oxide=more capacitance. In the case of thin oxide (not holes) I don't see how thick or thin would relate to leakage as long as there was something there. Running with the same thought, if the oxide is totaaly absent what makes it redeposit on the wax paper/mylar? Maybe the leakage is a result of the metallic compounds being absorbed by the electrolyte and the reforming process sends them back to the original metal, albeit somewhat randomly. Does this make any sense? -Bill |
Hi Bill,
The electrolyte is kind of like a salt water soaked paper, if you will, it is very low resistance. Its purpose is twofold: first, it creates an intimate contact with the aluminum oxide, which is the only dielectric in the cap, and second, it provides a method of repairing flaws in the aluminum oxide layer. The electrolyte IS the second plate of the capacitor. If the electrolyte wasn't there, (eg. it was paper instead), the dielectric constant of the paper would dominate and the capacitance would be similar in magnitude to that of a paper capacitor. It is very important that there not be any nonconductive material (other than the aluminum oxide layer, that is) between the plates. Another way an aluminum electrolytic capacitor could be made would be to build up the oxide layer on one of the plates, and then coat the oxide with a liquid metallic layer, such as silver epoxy paint, or "nickel print" to form the other plate. That should give you the high capacitance of a wet aluminum electrolytic cap, but the oxide wouldn't ever degrade. -Chuck --exray-- wrote: Chuck Harris wrote: Hi Alan, Because of the physical construction of an electrolytic cap, it MUST change capacitance if the oxide grows thinner in storage, or thickens thru reforming... But, I too notice that sometimes the change is large, and othertimes it is not. I suspect that what is happening is the oxide layer thins out only in spots (probably around impurities) in some caps. These spots are large enough to readily affect the leakage current, but are small with respect to the total surface area of the plates. Because they are a small percentage of the total surface area, they only minimally affect the total capacitance. -Chuck Harris Let me pose a question...not knowing inimately how electrolytics were or are made. Seems to me that the 'extra' oxide, ie thicker plates, are taking up some of the physical space that was formerly the electrolyte (part of the dielectric, so to speak) thereby leaving the plates closer together. That would indicate more oxide=more capacitance. In the case of thin oxide (not holes) I don't see how thick or thin would relate to leakage as long as there was something there. Running with the same thought, if the oxide is totaaly absent what makes it redeposit on the wax paper/mylar? Maybe the leakage is a result of the metallic compounds being absorbed by the electrolyte and the reforming process sends them back to the original metal, albeit somewhat randomly. Does this make any sense? -Bill |
| All times are GMT +1. The time now is 10:58 AM. |
Powered by vBulletin® Copyright ©2000 - 2025, Jelsoft Enterprises Ltd.
RadioBanter.com