Power supplies in-cable capacitors, why people dislike them ?

[Solid State Soul ( SSS )]

When reading power supplies recommendations threads i often find people dislike PSUs with bulky cables that has capacitors in them but then they are there to further help with the ripple suppression for the power supply so clearly they have a legit function yet i often fine people dislike them, sometimes going as far to avoid PSU's with in-cable caps You see many manufactures highlighting in-cable caps as a helpful features that slightly increase your PSU's performance in ripple suppression yet the wise people recommend i dont get one ?

Can someone please elaborate why that is ?

 

[robot zombie]

They make routing a pain in some situations. Especially with most cases now being set-up to have the main cables bend from around the back of the board to meet the connector.

If they're hanging it's fine... but sometimes that's not the best way to go... in which case the cable and especially the connector on board may face undue strain. Or maybe not fit at all, forcing you to run it across the main compartment instead of the back of the case.

 

[delshay]

Has anyone opened one up to see what type of capacitor?

EDIT: I have dual Cornell Dubilier 38000uf capacitor (total 76000uf) with custom built 8 pin that plugs direct into the back of my Seasonic 1200w PSU. It does nothing for the GPU, but CPU gain 1MHz overclock.

With a Scopemeter you can see a change, but it's so small it's hard to spot, but it is there.

 

[Bill_Bright]

I have mixed feelings about them. In theory they would help suppress any AC anomalies, including ripple and that's a good thing as ripple affects stability.

But IMO, the PSU should already be doing a good job at suppressing them before the voltage leaves the PSU itself. So when I see those type of cables, it makes me wonder why they are needed. Were they included after-the-fact, because some samples of the PSU failed in-house quality control testing? ??? We don't know.

That said, as an electronics technician and student of electronics for 45+ years, I understand parasitic capacitance (and inductance) is definitely unwanted, is unavoidable, but can be mitigated by using those in-line capacitors. And it should be noted that they also suppress such capacitance and inductance from being introduced into the cables from nearby devices and other cables.

I note one of the best methods of preventing these unwanted anomalies is to increase the insulation wall thickness. But then users would complain their cables were too fat, less flexible, harder to hide and tuck out of the way. So the trade off is a very small bump in the cable.

With off-brand generic PSUs, I might think in-cable capacitors are there to compensate for poor PSU ripple suppression. But for PSU makers like Seasonic or EVGA, companies known for making top quality PSUs, I have to believe in-cable capacitors are there to ensure users of their supplies get the best possible performance from those supplies. And I'm okay with that.

So I say we just must learn to live with them, and learn to appreciate their function.

 

[P4-630]

Or maybe not fit at all, forcing you to run it across the main compartment instead of the back of the case.

I had exactly this from the CPU power connector to my new Seasonic PSU, I couldn't bend it properly, not sure but the cable might have been a bit shorter than the one from my previous PSU but next year I may buy a new case, with a psu cover which my current case doesn't have now.

 

[Bill_Bright]

the cable might have been a bit shorter than the one from my previous PSU

I should note shorter cables are less susceptible to parasitic capacitance issues too.

 

[dont whant to set it"']

, but CPU gain 1MHz overclock.

Because more often than not it's that extra 1 MHz bclk increase that yeld's a record?

 

[eidairaman1]

I have mixed feelings about them. In theory they would help suppress any AC anomalies, including ripple and that's a good thing as ripple affects stability.

But IMO, the PSU should already be doing a good job at suppressing them before the voltage leaves the PSU itself. So when I see those type of cables, it makes me wonder why they are needed. Were they included after-the-fact, because some samples of the PSU failed in-house quality control testing? ??? We don't know.

That said, as an electronics technician and student of electronics for 45+ years, I understand parasitic capacitance (and inductance) is definitely unwanted, is unavoidable, but can be mitigated by using those in-line capacitors. And it should be noted that they also suppress such capacitance and inductance from being introduced into the cables from nearby devices and other cables.

I note one of the best methods of preventing these unwanted anomalies is to increase the insulation wall thickness. But then users would complain their cables were too fat, less flexible, harder to hide and tuck out of the way. So the trade off is a very small bump in the cable.

With off-brand generic PSUs, I might think in-cable capacitors are there to compensate for poor PSU ripple suppression. But for PSU makers like Seasonic or EVGA, companies known for making top quality PSUs, I have to believe in-cable capacitors are there to ensure users of their supplies get the best possible performance from those supplies. And I'm okay with that.

So I say we just must learn to live with them, and learn to appreciate their function.

Why not put the caps inside the psu right before entering the cable?

 

[Bill_Bright]

Why not put the caps inside the psu right before entering the cable?

There should already be filter caps inside the PSU. I think the idea is to suppress any capacitance caused by the cables or introduced into the cables by surrounding high speed circuits - which can happen.

 

[Assimilator]

When reading power supplies recommendations threads i often find people dislike PSUs with bulky cables that has capacitors in them but then they are there to further help with the ripple suppression for the power supply so clearly they have a legit function yet i often fine people dislike them, sometimes going as far to avoid PSU's with in-cable caps You see many manufactures highlighting in-cable caps as a helpful features that slightly increase your PSU's performance in ripple suppression yet the wise people recommend i dont get one ?

Can someone please elaborate why that is ?

1. In-cable caps have to be secured to their cables, which means you get the standard bundle-of-wires-with-cheap-sleeving instead of the flat Flexforce cables.
The latter are better because they are easier to route - not just because they're thinner, but because the smooth plastic-sheathed Flexforce isn't prone to catching on things like sleeving is.
2. In-cable caps make their cables more fragile. I can't count the number of times I've popped my PC case's side off to do some work, and some cable or connector is twisted at a weird angle or stuck in a groove in a way I could never achieve even if I tried. The prospect of that happening to cables with integrated caps horrifes me.
3. In-cable caps make their cables more expensive and difficult to replace (no more swapping for cables from a different PSU but same OEM).
4. In-cable caps make cables look fugly if you decide to sleeve them (because you can't sleeve the capacitor bulge).
5. Most PSUs get by just fine without in-cable caps. To my mind, they're a gimmick that manufacturers introduced because their marketing departments wanted slightly higher review scores. Personally I have little faith that they make much difference to the overall performance of the product.

 

[delshay]

Now this thread has got me thinking. Would custom solid core cable make a difference?

 

[dont whant to set it"']

The main difference would be the stifness or lack thereof.

 

[Solid State Soul ( SSS )]

Now this thread has got me thinking. Would custom solid core cable make a difference?

It is always recommended to use the cables that came with your power supplies. I'v read threads of people damaging their power supplies by using third party cables, some PSU's even have a warning on them to only use the cables that came with the unit itself.

 

[eidairaman1]

The main difference would be the stifness or lack thereof.

Thats why its not in automotive or pcs lol

 

[robot zombie]

It is always recommended to use the cables that came with your power supplies. I'v read threads of people damaging their power supplies by using third party cables, some PSU's even have a warning on them to only use the cables that came with the unit itself.

I think this mostly has to do with different pinouts on the PSU side. It's only onthe other side they are standardized. Only the cables that come with the PSU are guaranteed to have the proper pinouts. Of course 3rd party options with the correct pinouts exist, but you have to make sure they're made specifically for the PSU they're being plugged into.

 

[Mr.Scott]

EDIT: I have dual Cornell Dubilier 38000uf capacitor (total 76000uf) with custom built 8 pin that plugs direct into the back of my Seasonic 1200w PSU. It does nothing for the GPU, but CPU gain 1MHz overclock.

Oh Boy!
Totally worth it.

 

[Deleted member 185158]

Oh Boy!
Totally worth it.

lol. (10 char)

 

[Grog6]

You can't really add the 38000uF of capacitance directly to the PS mobo without destabilizing the power supply.

Even the 0.01ohm of resistance of the cables makes a huge difference in stability on a switching power supply, and the cables add that.
Switching power supplies are tuned for a specific maximum capacitance; they won't 'start up' with too much capacitance.
I've designed a bunch of power supplies.
Computer power supplies are really cheaply built, lol. I've worked on every one I've ever bought, before I used it.

Stranded cables are better than solid cables in every case; solid is cheaper, but there are frequency limits based on skin effect.
Solid wire works great for wall power, but would radiate like crazy used in a computer, at several frequencies.

Inside the PS, you'd need to use resistors between the output and the additional caps, and that dumps heat, and lowers voltage/power.

Adding it on the CPU motherboard would be ok, but there's another switcher there to make CPU voltage, and stability there is a concern.
You could design it in, tho.
Also those caps are pretty big, and mobo MFR's don't like huge things sticking up; you either have to load them after SMT soldering, by hand, or use expensive smt ones.

Another soldering step is expensive, maybe a 20% cost adder.

I'd also say the usefulness is varied board to board; cheaper mobos will benefit much more than expensive ones.
And depending on the design of the mobo, it might cause issues; if it makes the input to the CPU power section ring, it may not even start up.

It's a complex thing; this goes over some of the basics:

Foundations (Part 1.A) - Understanding Bode Plots and Stability of Power Suppliestags: bode plot, stability, crossover frequency, phase margin, gain margin

Most experts agree that the best way of designing power supply control loops is through the use of frequency domain design methods. Some people may have studied these techniques at university in...

www.biricha.com

 

[delshay]

You can't really add the 38000uF of capacitance directly to the PS mobo without destabilizing the power supply.

Even the 0.01ohm of resistance of the cables makes a huge difference in stability on a switching power supply, and the cables add that.
Switching power supplies are tuned for a specific maximum capacitance; they won't 'start up' with too much capacitance.
I've designed a bunch of power supplies.
Computer power supplies are really cheaply built, lol. I've worked on every one I've ever bought, before I used it.

Stranded cables are better than solid cables in every case; solid is cheaper, but there are frequency limits based on skin effect.
Solid wire works great for wall power, but would radiate like crazy used in a computer, at several frequencies.

Inside the PS, you'd need to use resistors between the output and the additional caps, and that dumps heat, and lowers voltage/power.

Adding it on the CPU motherboard would be ok, but there's another switcher there to make CPU voltage, and stability there is a concern.
You could design it in, tho.
Also those caps are pretty big, and mobo MFR's don't like huge things sticking up; you either have to load them after SMT soldering, by hand, or use expensive smt ones.

Another soldering step is expensive, maybe a 20% cost adder.

I'd also say the usefulness is varied board to board; cheaper mobos will benefit much more than expensive ones.
And depending on the design of the mobo, it might cause issues; if it makes the input to the CPU power section ring, it may not even start up.

It's a complex thing; this goes over some of the basics:

Foundations (Part 1.A) - Understanding Bode Plots and Stability of Power Suppliestags: bode plot, stability, crossover frequency, phase margin, gain margin

Most experts agree that the best way of designing power supply control loops is through the use of frequency domain design methods. Some people may have studied these techniques at university in...

www.biricha.com

You are correct. If I add two many of those capacitors the PSU refused to start, but you have to take also into consideration the load as well. The Antec truecontrol 550 (very old PSU) will accept 5 (190000uf) & light load. Adding those capacitors was an experiment, but i have left two of them in on my main computer which works well.

EDIT: When you say "destabilizing the power supply" in what way. I'm seeing nothing here even on the scopemeter.

 

[Grog6]

The problem is the charging of the capacitors.

Adding an extension to the 24 pin cable may help you use more.

If you had the datasheet for the chip that drives the switcher in the PS, you can put in a slightly larger "Soft-Start" capacitor, to extend the charging period.

The down side is that there's a not insignificant chance of a loud bang, and some bad hardware.
If the power supply loses it's stuff badly, there's no telling how much voltage it will put out.
Putting out 12V on the 3.3V or 5V lines will be really bad for the motherboard. :O

I use programmable loads until I get stuff stable, then try real hardware.

You can buy cheap stuff like this to play with, and modify:

DC Constant Current Electronic Load Battery Tester XY-FZ35

Snap-in panel meter/Load for testing and conditioning batteries. Input Voltage: 5.0-30.0VDC (Polarity protected) Load Voltage: 1.5-25.0VDC (Polarity protected) Load Current: 0.00-5.00A 0.01A Resolution Load Power: 35W Max Current Regulation: ±(1%+3digits) Voltage Regulation: ±(0.5%+1digit) Over...

www.mpja.com

A switcher should start correctly with a 5A load, but you'd need multiple ones for various PS outputs.

Stay away from the AC inputs, and you'll be fine. Like I said, any "oopsies!" and the mobo may be trash.

 

[silentbogo]

For me it's just a gimmick which can introduce yet another point of failure.
Most PSUs that have spare budget to add caps inside cables are usually the ones with already excellent regulation and ripple suppression, and add zero benefits comparing to "non-capacitated" equivalent.
Also, most manufacturers that do add caps usually do so only on 12V rail(s), which makes even less sense due to receiving devices(e.g. MoBo/GPU) not using 12V directly and have switching regulators of their own with a bunch of input and output capacitors to do the same thing.

Why not put the caps inside the psu right before entering the cable?

All PSUs have caps right next to cables, and most modular PSUs also have caps on the PCB with modular connectors.

You are correct. If I add two many of those capacitors the PSU refused to start, but you have to take also into consideration the load as well. The Antec truecontrol 550 (very old PSU) will accept 5 (190000uf) & light load. Adding those capacitors was an experiment, but i have left two of them in on my main computer which works well.

Your old antec truecontrol has no PFC, so it does not care about what's going on with the input. For APFC-enabled PSUs this will do more harm than good.

 

[delshay]

@Grog6

What type of capacitor are manufactures fitting inside the cable?

Oh & Seasonic 850w titanium also accepts max two of those 38000uf capacitors.

 

[Bill_Bright]

For me it's just a gimmick which can introduce yet another point of failure.

I disagree - at least about it being a gimmick. I don't believe highly reputable brands like EVGA and Seasonic are into "gimmicks", nor do they need gimmicks to entice users to buy their products. Companies that make inferior products need gimmicks. The reviews AND in-the-field experiences show EVGA, Seasonic and a few select others make products that truly are top quality.

One concern of mine with these in-cap cables is it means the cables are proprietary. I don't like anything that is proprietary as it just means it will cost me more money and will limit my options. While the component-end connectors of these cables must comply with the ATX Form Factor standards, the cables and PSU-end connectors are proprietary. This means you cannot swap these modular cables with any other PSU - including PSUs from the same maker!

That means we must label and store unused cables for each modular PSU we own or are responsible for separately - forever! If we fail to do that, we may end up using the wrong cable on a PSU. Additionally, there is no guarantee in 4 or 5 years a replacement cable will be available, should the original cables get lost.

I have 5 PCs here with modular PSUs - all different models. I am responsible for dozens more. I does not take long to become a logistical nightmare keeping track of that inventory - especially for small businesses that have no real IT person on staff.

IMO, we need the ATX Form Factor standard to dictate "end to end" standards for modular cables too.

 

[Grog6]

I've never seen that, sorry.

Seems clunky to me; I'd design them in, if it were me.

I just know power supplies.

 

[silentbogo]

I disagree - at least about it being a gimmick. I don't believe highly reputable brands like EVGA and Seasonic are into "gimmicks", nor do they need gimmicks to entice users to buy their products. Companies that make inferior products need gimmicks. The reviews AND in-the-field experiences show EVGA, Seasonic and a few select others make products that truly are top quality.

Bill, as I mentioned before, these capped cables only appear exclusively in high-end PSUs, and effectively do nothing to improve an already perfect ripple suppression. Hard data from reviews only supports it. It's similar to RAM waterblocks for DDR4. Yes, in theory they do work and may improve things, but in practice bring very little to justify the hassle.
For example, Seasonic Prime w/ fancy cables has pretty much identical ripple and transient response to its non-capped Seasonic Prime or Focus Plus counterparts, but has a slightly higher price tag. In case of Prime Ultra, Seasonic decided not to use cable caps at all, but the PSU itself has over 2x less ripple on all rails. So, even they understand that there is no point in upselling capped cables when your PSU design already has a sub-20mV ripple on 12V rail.