Aren't the ones on the board regular electrolytics? Is there a reason not to recap with electrolytics (other than maybe leaking in another 20 years?)
feipoa wrote on 2024-08-15, 00:12:I remember recapping this board in 2012 with aluminium polymer caps. I don't know if this type of capacitor is ideal for this motherboard
appiah4 wrote on 2024-08-19, 09:08:Aren't the ones on the board regular electrolytics? Is there a reason not to recap with electrolytics (other than maybe leaking in another 20 years?)
Generally, replacing electrolytics with other electolytics is OK. The aluminum polymer caps mentioned by feipoa are a kind of electrolytics caps that are intended to be very durable (even more than 20 years), but also have a very low ESR. The range of ESR values is also quite broad for different types of wet (standard) electrolytic caps, ranging from general purpose caps to ultra-low-ESR caps. Some voltage regulators rely on the resistance in the output cap to dampen oscillation, so lowering the ESR when replacing caps might be a bad idea. 486 boards like the MB-8433UUD-A have just one voltage regulator, which is used for low-voltage 486 processors. This regulator is a low-dropout regulator, and low-dropout regulators are known to be more prone to oscillation when used with unexpected low ESR output capacitors than old linear regulators. For some low-dropout regulators, the datasheet specifies which capacitance/ESR combinations are guaranteed to result in stable operation. Using aluminum polymer caps might be out of spec. On the other hand, there are not a lot of high-capacitance capacitors on that board at all, and the couple of caps for the core voltage is distributed around the 486 socket and not placed directly next to the regulator, so the trace resistance and inductance is counteracting the low impedance of polymer caps.
After a quick check of Vcore stability, even with load transients (like switching in/out of HLT mode), I wouldn't worry about the polymer caps.
mkarcher wrote on 2024-08-19, 11:26:Generally, replacing electrolytics with other electolytics is OK. The aluminum polymer caps mentioned by feipoa are a kind of el […]
feipoa wrote on 2024-08-15, 00:12:I remember recapping this board in 2012 with aluminium polymer caps. I don't know if this type of capacitor is ideal for this motherboard
appiah4 wrote on 2024-08-19, 09:08:Aren't the ones on the board regular electrolytics? Is there a reason not to recap with electrolytics (other than maybe leaking in another 20 years?)
Generally, replacing electrolytics with other electolytics is OK. The aluminum polymer caps mentioned by feipoa are a kind of electrolytics caps that are intended to be very durable (even more than 20 years), but also have a very low ESR. The range of ESR values is also quite broad for different types of wet (standard) electrolytic caps, ranging from general purpose caps to ultra-low-ESR caps. Some voltage regulators rely on the resistance in the output cap to dampen oscillation, so lowering the ESR when replacing caps might be a bad idea. 486 boards like the MB-8433UUD-A have just one voltage regulator, which is used for low-voltage 486 processors. This regulator is a low-dropout regulator, and low-dropout regulators are known to be more prone to oscillation when used with unexpected low ESR output capacitors than old linear regulators. For some low-dropout regulators, the datasheet specifies which capacitance/ESR combinations are guaranteed to result in stable operation. Using aluminum polymer caps might be out of spec. On the other hand, there are not a lot of high-capacitance capacitors on that board at all, and the couple of caps for the core voltage is distributed around the 486 socket and not placed directly next to the regulator, so the trace resistance and inductance is counteracting the low impedance of polymer caps.
After a quick check of Vcore stability, even with load transients (like switching in/out of HLT mode), I wouldn't worry about the polymer caps.
Nice analysis.
I figured the aluminium polymer caps may not be the most optimal if taking a blind look at things, but that's what I had on hand at the time. I think I was aimlessly troubleshooting the IBM 5x86-133/2x setup in attempts to lower the core voltage, or for some BSOD issue; I just don't recall any longer. What I do recall was the caps didn't change whatever I was testing for, so I left them installed. It is good to know that they should be OKto leave be. I often wondered if I should swap the caps back to electrolytics.
Plan your life wisely, you'll be dead before you know it.
So far capacitors have never been a major problem for me, excluding specific cases with bulged, exploded, missing elements.
This may be a bigger deal for early 2k electronic components, but not for products from the 80ies-90ies time period.
pshipkov wrote on 2024-08-21, 16:28:This may be a bigger deal for early 2k electronic components, but not for products from the 80ies-90ies time period.
The importance of capacitors increased significantly with switching voltage regulators appearing everywhere, which is a late 90s thing. The rising demand for capacitors that are suitable for switch-mode regulators created a big supply, some of which was crappy.
This does not mean everything is fine with capacitors of the 80s, though. If you try to power up an IBM 5150 mainboard, exploding tantalums on the 12V rails are very likely. A lot of Ensoniq Soundscape sound cards have very silent output due to broken capacitors. Most Mac classic Mainboards need all capacitors replaced. I had multiple AT supplies fail due to broken capacitors. But you are right: I don't remember any 386/486 mainboard that needed a recap.
