another stupid polymer question (can going REALLY high ripple be bad?)

[Uranium-235]

I have an ECS with KGZ's with an ESR of 0.013. HM's are that low but mouser seems not to have any

polymers seem like a possible replacement.

they're 16V 1500uf KZG 0.013ohm, 2.55A
I'm thinking 16V 1000uf 0.012 Panasonic OS-CON with an insane 5.4A

is it really ok to go up in amperage that much? I know Higher ripple is generally better, but a ripple that high might have some effect i'm not aware of

[brethin]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

Depends on where they are used. If your talking in the VRM they will be fine.

[Uranium-235]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

Yeah cpu vrm

Thanks

[momaka]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

The ripple current rating of the caps does not impact circuit performance one single bit. Only the ESR and capacitance does.

The ripple current rating is simply the limit of... well, ripple current that the cap can handle before overheating internally to unsafe temperatures.

That said, those OS-CON caps should be more than fine. Most motherboards usually use 330-470 uF caps in those spots (the VRM high side, that is).

[Uranium-235]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

you mean input VRM. Yeah I replaced it with the 1000uf and they get a little hot under max cpu but it ran fine.

Well I think the MOS's and other parts of the VRM itself were designed with a certain input capacitance in mind, but that can be fudged to a certain point (yeah I do know most good modern all-poly VRM's inputs are designed to have 330-470uf, but I think lytic VRM's are designed different).

The circut originally had KGZ's on the input and sanyo WF's output. Pretty sure the KGZ's dried out without bloating, causing bad ripple, making most of the sanyos bloat.

replaced the sanyos with Nichicon HN's and it's been working great. Use it for my grandmothers maybe? switch the chip out with a 5200+ (instead of 5600) and bring it to 3GB of ram (she has W7 32-bit)

[stj]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

why do people use a 32bit o.s. on a 64bit system??
complete waste.

[momaka]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

you mean input VRM. Yeah I replaced it with the 1000uf and they get a little hot under max cpu but it ran fine.

VRM high side, VRM input... same thing
It's the PSU 12V rail for the CPU (i.e. 4/8-pin CPU connector)

Yeah I replaced it with the 1000uf and they get a little hot under max cpu but it ran fine.

That's normal. I've seen very few boards to have cool-running CPU caps (be it on the high side or low side). The more modern square "encapsulated coil" inductors are the worst, though - manufacturers are cutting corners everywhere these days to save on material costs, so we end up with products that have much lower MTBF.

Well I think the MOS's and other parts of the VRM itself were designed with a certain input capacitance in mind, but that can be fudged to a certain point (yeah I do know most good modern all-poly VRM's inputs are designed to have 330-470uf, but I think lytic VRM's are designed different).

Not for the input/high side. Here the ESR matters a lot more than the capacitance. On the output, though, you want to stay within a certain range, because otherwise that can mess with the PWM controller's feedback gain loop. Too little capacitance, and the gain poles can shift enough to cause oscillation and make the circuit unstable. Half the capacitance is about as far as anyone should go when doing electrolytic-->polymer mods - and that only applies to buck-type circuits. With linear regulators, keeping the same capacitance matters a lot more than the ESR. ESR just has to be low enough to not make the caps ineffective.

The circut originally had KGZ's on the input and sanyo WF's output. Pretty sure the KGZ's dried out without bloating, causing bad ripple, making most of the sanyos bloat.

Got an ESR meter? I can bet you $10 they're fine. I have seen very very few bad 16V 1500/1800 KZGs on the input. But the 6.3V, 820 uF, 1500 uF, and 3300 uF - tons! But especially the 3300 uF. The 16V 1500 uF KZGs on the input of mobos I often pull and keep as backup test caps. Used some of them more than a few times to get a temporary recap going. They are always fine.

Use it for my grandmothers maybe? switch the chip out with a 5200+ (instead of 5600) and bring it to 3GB of ram (she has W7 32-bit)

Yeah, should be more than fine for web browsing. I have an X2 6000+ with 3 GB of RAM on W7 32-bit, and it runs anything online. YouTube 720 and 1080p are no problems, even without video acceleration. To save power, I have that X2 6000+ down-clocked to 2.4 GHz and under-volted to 1.150V. You'll be totally surprised how low those AM2 X2 chips can go on the core voltage, even at stock full frequency. I can run my 6000+ 100% stable at 1.2V core and full 3 GHz frequency - that's 0.2V lower from its stock 1.4V core. This really cuts down on the CPU temperature.

why do people use a 32bit o.s. on a 64bit system??
complete waste.

No, not a complete waste at all. There are many reasons why I still prefer 32-bit versions of Windows over their 64-bit counterparts, though it's mostly to do with compatibility.

With Windows 7, one of them is enforced driver signing: 32-bit version does NOT have it, but the 64-bit does. This can wreak havok when trying to use unsigned drivers for some old hardware (or anything home-brew for that matter). Another thing is, W7 64-bit cannot run 16-bit apps. Again, if you have an old piece of software that needs this, you'll be stuck. Either will have to run in a virtual environment or have a separate (old) PC for that.

The only reason for me to run a 64-bit OS is when I have a PC with 4 GB of RAM or more. Other than that, it's absolutely pointless for everyday use. Unfortunately, everyone seems to be pushing the 64-bit bullshit nowadays - sometimes even breaking or making things not work intentionally, just for the sake of making you upgrade to their "latest and greatest"... like MS tried with Skype, telling me I can't install Skype without IE12, or something like that. Turned out there was a super-easy way around that, by just making the installer run in "compatibility" mode for XP SP3. And then Skype installs no problems. LOL! (Not that I care anymore - Skype's been dead to me ever since MS bought it).

[stj]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

64bit is much faster for compiling.

[brethin]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

Drivers for 64 bit can be harder to find for some hardware and unless you have more than 4 gig ram installed and the board supports it, 32 works fine.

[brethin]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

64bit is much faster for compiling.

Not many do much compiling or even know what that is, But yes if the compiler is 64 bit it will compile faster usng a 64 bit OS.

[stj]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

i compile

[momaka]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

i compile

Ha! That's one of the things I DON'T do and I don't like doing either.
I have no problems spending hours upon hours working with my hands on hardware. But software (be it coding or troubleshooting)... it better work out quickly / the way it should. Otherwise I loose patience and start cussing like a mad sailor ... and software almost never works right or properly the first time these days. So I tend to stick with stuff that's been tested, tried, and true... i.e. old drivers and patches that I've collected over the years.

[kaboom]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

Ha! That's one of the things I DON'T do and I don't like doing either.
I have no problems spending hours upon hours working with my hands on hardware. But software (be it coding or troubleshooting)... it better work out quickly / the way it should. Otherwise I loose patience and start cussing like a mad sailor ... and software almost never works right or properly the first time these days. So I tend to stick with stuff that's been tested, tried, and true... i.e. old drivers and patches that I've collected over the years.

Ya mean I'm not the only one who does that?

As in the phrase "that escalated quickly," I tend to go "George Patton" when it comes to software/malware issues- to the point I can't work on others' junkputers anymore.

I feel myself "building up" and going to war with it anymore. The fact that brainwashing/trolling since vista is beyond full force doesn't help. My value system regarding tech and what should've been its potential hasn't changed; of course I take offense when the M$ (and other) $hills try to conquer everything with their idea of tech cargo-cultism/communism.

And the small thinkers always hint at trying their particular flavor-of-the week, version XYZ of something. Hint: If you're up to ver. XYZ, it's either because it sucked from the beginning, or it's being broken on purpose to force an artificial upgrade.

Well it won't work- been at this since IE-Gate.

I digress...


/profound

[momaka]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

Ya mean I'm not the only one who does that?

Yup. Sorry! You're not as old and cranky as you think you are.

When I set up my PCs, I take my long sweet time with that. But that's because I have to change almost every default setting to something that is more reasonable. Then there's the part of installing 3rd party software that adds functionality that doesn't exist otherwise. On that note, I'm very picky on what I put on a computer. Just because a piece of software is popular or free (or both) doesn't mean it's good. I have a small but carefully-picked library of software that I put on every new Windows OS install - nothing to hog the system down to do updates or start up with every boot when it's not needed.

When I am done with all of that, the idea is to have a system that will serve me well in the long term with minimum maintenance and annoyances (which nowadays translates for me as "no updates, thank you very much"). For example, on this laptop that I'm typing from now (Dell Latitude with Pentium 3)... it's a slow and completely outdated device, in terms of specs. However, that doesn't stop it from booting faster and feeling more responsive than *some* modern 8+core Windows 10 rigs today (the ones with all kinds of bloatware). It's because when I set it up 8 years ago, I put only the software and patches it *needed*. Since then, I haven't had to touch almost anything (save for the browser, because that's the only thing that forces me to update anything). It still boots just as fast as the day I installed the OS. And it will continue to do so, because I don't put any bloatware on it. Same goes for my "main" (work) desktop: I set it up in 2013 and it's been running the same way since then. I've added quite a bit of software on it over the years (mostly related to work I needed to do and university - as that was the PC for my university dorm), but again nothing that would hog system resources. And both of these are very reliable machines too in terms of hardware. On that note, I always keep separate PCs for gaming systems, as the video cards in those always compromise reliability (at least any mid and high-end video card made in the last 15 years).

As in the phrase "that escalated quickly," I tend to go "George Patton" when it comes to software/malware issues- to the point I can't work on others' junkputers anymore.

Oh, I can, but I just give a clear disclaimer that if I have to do anything on the OS-side of things, I will "nuke" and delete things at will. Any program that slows down the OS is removed from startup at best (or uninstalled otherwise.), even if the original issue I was called to work on has nothing to do with this.

And all in all, I don't mind working on other peoples' computers, if that means fixing something. But when it comes to using their computer to complete something (like work or simple activity), this can tick me off badly, especially on a computer that runs like crap.

And the small thinkers always hint at trying their particular flavor-of-the week, version XYZ of something. Hint: If you're up to ver. XYZ, it's either because it sucked from the beginning, or it's being broken on purpose to force an artificial upgrade.

Yup, so true. You might even say the same about Ubuntu and a few other "distros" these days.

I feel myself "building up" and going to war with it anymore. The fact that brainwashing/trolling since vista is beyond full force doesn't help. My value system regarding tech and what should've been its potential hasn't changed; of course I take offense when the M$ (and other) $hills try to conquer everything with their idea of tech cargo-cultism/communism.

Same here... but I think I've went far enough off-topic as it is with my last several posts here, so I'll save it for some other thread when the time comes.

[Uranium-235]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

64-bit uses more ram for the windows sxs. Windows will perform better in low (< 4GB) if it's 32-bit. Past that, it won't see it

64-bit will perform as good if you have >4GB

if you have limited slots (2xddr2 in this case). I would rather try to get 2-3GB just for firefox/thunderbird

[ratdude747]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

Let's see:

Java: Faster on 64 (or so I'm told and have seen)
Some programs (such as chrome on linux): No longer supporting x32. CAD software at work (Autodesk): x64 only (although you'll need way more than 4GB of RAM for it).

For old x86 hardware and old 16 bit software, 32 bit is required... but anwhere else, I x64 it. Faster and more compatible with newer software, which for me, is what works best.

[stj]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

well 64bit is better in every way, but obviously is uses double the ram - it's double the buss width in hardware.

having said that, if your using Linux you can just get away with 3gig of ram on 64bit - i'm doing it.

[Uranium-235]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

Let's see:

Java: Faster on 64 (or so I'm told and have seen)
Some programs (such as chrome on linux): No longer supporting x32. CAD software at work (Autodesk): x64 only (although you'll need way more than 4GB of RAM for it).

For old x86 hardware and old 16 bit software, 32 bit is required... but anwhere else, I x64 it. Faster and more compatible with newer software, which for me, is what works best.

yes well my grandmother uses firefox and thunderbird. Both have 32-bit versions which work fine.

she dosen't use cad, java (no java programs, and I won't make the mistake of having the plugin in the browser-thats a virus waiting to happen)

I barely care if they stop supporting 32-bit firefox, she uses just facebook and yes I know that is full of hacked ads and targets everybody for vulnerabilities, and dropping support might mean they stop fixing vulnerabilities. But I don't care. A recent shift in the desktops I have means I might be able to get her a 5200+ with 4gb (actually the board I just capped as per this thread).

[stj]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

switch from firefox to waterfox.
faster and no(less?) spyware.

[momaka]

Re: another stupid polymer question (can going REALLY high ripple be bad?)

(no java programs, and I won't make the mistake of having the plugin in the browser-thats a virus waiting to happen)

I'm with you on that one.

Unless I have a very specific software that requires Java, I don't install it on any of my computers anymore.

I barely care if they stop supporting 32-bit firefox, she uses just facebook and yes I know that is full of hacked ads and targets everybody for vulnerabilities, and dropping support might mean they stop fixing vulnerabilities. But I don't care.

The majority of infected systems I've seen didn't come from people who didn't update their browser. It came from user stupidity... or rather the user not being well educated in basic cyber security, I should say. So updating the browser isn't that important anyways. I'm still using various versions of FireFox 40 up to 52, alongside my Portable 24. No problems whatsoever, even on the occasional shady torrent website.

The first step is to always install AdBlock Plus with FireFox - that way all those fake "Upgrade Flash Player" and "Upgrade Java" ads get removed from displaying (but not blocked), along with all other clickbait crap like "Click here to start your download". Adding a good HOSTs file to block stuff before they get to AdBlock is even more important - that way, you get system-wide protection and not just protection at the browser-level.

A recent shift in the desktops I have means I might be able to get her a 5200+ with 4gb (actually the board I just capped as per this thread).

Should be more than enough. Facebook is a pig with the RAM, but 4 GB should definitely be enough. The 5200+ (which is IIRC a 2.6 or 2.7 GHz AM2 CPU) should chug along just fine. I downclock my X2 6000+ to 2.4 GHz, and even that is more than enough for YouTube 720p videos fullscreen and other not-so-optimized websites.

Let's see:

Java: Faster on 64 (or so I'm told and have seen)

LOL, who uses Java anymore these days?

Java became dead to me many years before Adobe Flash, and Flash has been dead for quite a while too. I was probably one of the last holdouts (up until last year I think) trying to trick websites like YouTube to use it instead of HTML5-based players. That said, some websites still do use Flash. But it's not widespread anymore.

Some programs (such as chrome on linux): No longer supporting x32.

Heh, and some Linux distros used to pride themselves with being faster and more backwards-compatible for older PCs.... So much for that.

CAD software at work (Autodesk): x64 only (although you'll need way more than 4GB of RAM for it).

Well, if you must absolutely have the latest CAD version, then yes.
Nothing says you can't use an older version to do the same work.
... unless of course you're rendering some major gaylord detail that will absolutely require more than 4 GB of memory.

For old x86 hardware and old 16 bit software, 32 bit is required... but anwhere else, I x64 it. Faster and more compatible with newer software, which for me, is what works best.

For the majority of software, there's nothing faster about a 64-bit program over its 32-bit counterpart. That's just more daily nonsense hype to get people to buy newer hardware. I'm not condoning 64-bit hardware or anything like that, but it's just simply wrong to claim that 64-bit software is always faster than 32-bit. It's not.

As for more compatible with newer software - yes.

On that note, I really hate the way Windows Vista/7 (and I guess everything newer) handles 64-bit vs. 32-bit programs. Once you go with a 64-bit Windows, even 32-bit doesn't always work right, especially programs that don't actually install but run as stand-alone executables and you try to set some shell options with them. In 64-bit Windows, that almost never works. With 32-bit Vista and 7 - sometimes. With XP - all day every day fine.