Ghetto Mods

[Apocalypsee]

I fixed my Logitech G400 that have intermittent connection, I haven't used it for a year or two. Looks like the problem is simple, its just the wires that comes out of the mouse where it bends that causing the issue. Cut that out and solder the wires directly to the board and voila! It already serves me for 8 years, it should be fine for another 8 more years Solder might look like a patchjob but I checked the connection is all good.

 

[micropage7]

there's nothing wrong if you try to find the cheaper one
but today the quality is piece of junk, oh at least the electronic works well but the mounting, again, it's pretty bad, i need to cut here and there, cut the pcb and adding some padding so the HDD could sit well

 

[freeagent]

So, I didn't have a problem with air flow, or over heating or anything like that.. But I did have a spare fan, a controller, and some free time Space was tight under the gpu, it was almost a perfect fit, I cut a couple of zip ties and put some sticky tack under and pressed down. Its a perfect fit lol. It is a 4700rpm screamer.. I initially tried using a mobo header, but its lowest speed was 2700rpm, it was tolerable, but.. so I have it on a Zalman controller now sitting at 2000 RPM. The glass doesn't get warm at all during extended sessions

 

[droopyRO]

The top-right fan is blowing air in ?

 

[freeagent]

The three fronts are in, top front is in top rear is out. The two rears are pulling on the hsf. It may look a little unorthodox but the system works fantastic. I don’t have to run them at 12v as there is more than enough flow right now. It’s not silent, but it’s quiet.

 

[droopyRO]

Isn't the top-right fan, pulling hot air in, from the top-left fan ?

 

[freeagent]

Nope, there is too much air moving for it to get warm

 

[delshay]

ASROCK 939 Motherboard Capacitors stacking.

Messing around with stacking capacitors three level high. Going for a forth & final level.

1. second layer mirrors the original capacitors (ceramic).
2. Third layer is custom which can be seen from the photo below (ceramic & Tantalum). Some Tantalum can't be seen as their are under those big capacitors.

Seem to gained 2MHz overclock with undervolt (9hr+ with P95) but it's not valid at this time.

NOTE: You can cheat on temperature if you mod the very top left hand component. ie impress someone you got 4GHz on air with just 5c on the board.

Also note all capacitors on bottom row are for DDR Ram including the last four to the right (2.6v+). All other capacitors are CPU core.

 

[Grog6]

Nice work!

I keep a few kinds of solder for doing such things; a few degrees melting point can make a huge difference.

You know you want the smallest values closest to the PCB, with larger values further out; this maximizes the loop inductance vs impedance for a given value.

All ceramic caps are inductive over their resonant frequency, so you want to stack those resonances across a wide range.

This stack would be the trace of the lowermost lines, all the way across; so usable out to ~200MHz, really.

 

[delshay]

Nice work!

I keep a few kinds of solder for doing such things; a few degrees melting point can make a huge difference.

You know you want the smallest values closest to the PCB, with larger values further out; this maximizes the loop inductance vs impedance for a given value.

All ceramic caps are inductive over their resonant frequency, so you want to stack those resonances across a wide range.

This stack would be the trace of the lowermost lines, all the way across; so usable out to ~200MHz, really.

The Tantulum are doing all the damage ie improving performance. I did not think such a small value capacitor will bring that kind of performance.

Their so good I force on the last four I had left. The final layer are polymer capacitors ESR 3 mohms Binned, to be fitted.

This motherboard has upgraded MOSFETS for most of the motherboard with increased power output for the CPU done at the controller.

Most capacitors on this motherboard are much closer 2mohms binned (not the 16v capacitors).

 

[Grog6]

When the board is switching 100A in a few microseconds, a little bit of capacitance can make a big difference.

Inductance is the big problem; a trace 0.1" long is right at 1nH, and can drop quite a bit of voltage on a fast current change.
Luckily, there are a bunch of traces in parallel, one per pin.

A via 0.010" can be 1 milliohm, and 1.5nH inductance, easily.

So the closer the capacitance is to the processor pin, the more effective it is.

These are really nice for high frequency:

JMK107B7225MAHTR Taiyo Yuden | Capacitors | DigiKey

Order today, ships today. JMK107B7225MAHTR – 2.2 µF ±20% 6.3V Ceramic Capacitor X7R 0603 (1608 Metric) from Taiyo Yuden. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

These are good, but they don't last as long, 1000hour rated.

F380J476MMAAXEH3 KYOCERA AVX | Capacitors | DigiKey

Order today, ships today. F380J476MMAAXEH3 – 47 µF Molded Tantalum Polymer Capacitor 6.3 V 0603 (1608 Metric) 300mOhm @ 100kHz from KYOCERA AVX. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

 

[delshay]

When the board is switching 100A in a few microseconds, a little bit of capacitance can make a big difference.

Inductance is the big problem; a trace 0.1" long is right at 1nH, and can drop quite a bit of voltage on a fast current change.
Luckily, there are a bunch of traces in parallel, one per pin.

A via 0.010" can be 1 milliohm, and 1.5nH inductance, easily.

So the closer the capacitance is to the processor pin, the more effective it is.

These are really nice for high frequency:

JMK107B7225MAHTR Taiyo Yuden | Capacitors | DigiKey

Order today, ships today. JMK107B7225MAHTR – 2.2 µF ±20% 6.3V Ceramic Capacitor X7R 0603 (1608 Metric) from Taiyo Yuden. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

These are good, but they don't last as long, 1000hour rated.

F380J476MMAAXEH3 KYOCERA AVX | Capacitors | DigiKey

Order today, ships today. F380J476MMAAXEH3 – 47 µF Molded Tantalum Polymer Capacitor 6.3 V 0603 (1608 Metric) 300mOhm @ 100kHz from KYOCERA AVX. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

Thanks, but it's funny you should point those out as I do have capacitor from both those company but not those series.

This motherboard has capacitors that run much closer to the device voltage. The goal here was to get capacitors near as possible to the voltage output.
I do believe this gives better filtering/performance.

The capacitors on the back of the socket are 47uf 4v Tantalum, so you are spot-on with the value here. The big capacitors on the back of the CPU are from TAIYO YUDEN (also 4v), but not from the series you pointed out. I'm using 4v capacitors as it makes it also compatible for the DDR Memory (CPU side).

 

[Grog6]

As long as the voltage is above the applied voltage, it's no problem.
Electrolytic caps want to be within 60% of the maximum, to maintain the "Electrolytic barrier", too low of a voltage will cause excess leakage current.

I found recently that a 16V capacitor used at 1.8V lost it's barrier after about 4 years, and needed to be reformed; I replaced it with a 3V polymer cap.

Avoid Z5U caps; X7R dielectric is much better for Frequency response, and fast currents.

Those are examples of what I use in those kinds of places.

 

[delshay]

As long as the voltage is above the applied voltage, it's no problem.
Electrolytic caps want to be within 60% of the maximum, to maintain the "Electrolytic barrier", too low of a voltage will cause excess leakage current.

I found recently that a 16V capacitor used at 1.8V lost it's barrier after about 4 years, and needed to be reformed; I replaced it with a 3V polymer cap.

Avoid Z5U caps; X7R dielectric is much better for Frequency response, and fast currents.

Those are examples of what I use in those kinds of places.

I will be using 2v polymer as the last layer (CPU CORE ONLY). I can also solder this on the side so it's sits lower. Performance is higher than what I said in this thread, but something is overheating. It's not the VRM or CPU, it's pointing towards the controller or a small chip near the PATA socket.

I currently have it running again with P95, so I can grab a screenshot.

I also found you can also cheat on the voltage applied to the CPU tricking CPUZ & other programs.

 

[Grog6]

What's the controller chip?

If you have access to a thermal camera, it's a wonderful tool to troubleshoot things like these.

SOME digital cameras have IR filters that can be removed to make them somewhat of an IR camera; anything that will see heat is good.

Amazingly enough, our cheeks and palms are IR sensitive, if there aren't any other sources in the room.

Be careful; I had "LM7805" branded backwards on a finger for a year, once.
I smelled something hot; I found it!.
Lol.

 

[delshay]

What's the controller chip?

If you have access to a thermal camera, it's a wonderful tool to troubleshoot things like these.

SOME digital cameras have IR filters that can be removed to make them somewhat of an IR camera; anything that will see heat is good.

Amazingly enough, our cheeks and palms are IR sensitive, if there aren't any other sources in the room.

Be careful; I had "LM7805" branded backwards on a finger for a year, once.
I smelled something hot; I found it!.
Lol.

Controller chip L6714

 

[Grog6]

That's a pretty good controller, FWIW.

4 phases, nice features.

https://www.mouser.com/datasheet/2/389/stmicroelectronics_cd00097798-1205905.pdf

I'd put a heatsink on it, jut because; it will drive the mosfet gates 2A, and it runs 150khz nominal.

I'm having a heck of a time replacing electrolytic caps on a early core2 mobo; the solder won't come out of the holes.

I dug out the big weller iron, so it will unsolder now, lol.

My standard Iron won't even warm up the ground plane, lol.

 

[dorsetknob]

I'm having a heck of a time replacing electrolytic caps on a early core2 mobo; the solder won't come out of the holes.

My standard Iron won't even warm up the ground plane, lol.

Probably so called lead free Solder ,>> melting point is higher than lead solder.>>> most Irons wont get hot enough to melt LF Solder

 

[Grog6]

This one has Asus's "Thermal Armor", which is a thick ground plane on the bottom, with no "wagon wheels" around the holes, so hard to heat.

I got the big Iron out:

This will Work!

Normally, I use a Weller temperature controlled workstation, but the ground sides won't clear so I can get the new caps in, so this.


The Iron I have is actually older than me, from 1955.

This one IS lead free, so that's no fun either; I use "Chipquick" solder to remove stuff, but this one's being difficult.

I use Kester "44" 63/37 to go back with.

 

[delshay]

That's a pretty good controller, FWIW.

4 phases, nice features.

https://www.mouser.com/datasheet/2/389/stmicroelectronics_cd00097798-1205905.pdf

I'd put a heatsink on it, jut because; it will drive the mosfet gates 2A, and it runs 150khz nominal.

I'm having a heck of a time replacing electrolytic caps on a early core2 mobo; the solder won't come out of the holes.

I dug out the big weller iron, so it will unsolder now, lol.

My standard Iron won't even warm up the ground plane, lol.

Thanks, but I already have the PDF on the controller, this is why I was able to up the power & disable vdroop. I think I have found a fault but it's not heat related, investigating.

Getting solder out of the hole with no damage is a work of art. I use a pointed soldering tip & set it to max power 480w. The tip goes partly though the hole & come out the other side. It's best to pre-heat the motherboard then you can use the soldering iron at lower temperature. To remove electrolytic caps I just twist them from the top of the motherboard, then desolder the legs. This method is for electrolytic caps only, do not use this method on any other type of capacitors.

Low powered soldering iron will do damage to the motherboard, you need a fairly small tip but high power 450w+. How much power you need depends on tip size. Large tips can also damage the motherboard. If you have a large tip, then the easiest way to get the solder out of the hole is to add more solder onto the hole. Don't worry about the extra solder as it will suck out the solder out of the hole in combination with 450w+ soldering iron & solder wick. You may have to repeat this a few times, but it does work also with no damage. Try to get the solder out from the top side of the motherboard first, but keep switching sides.

With a big soldering tip you should be getting the solder out of the hole on the first or second attempt. If not, then something is wrong, maybe wrong type of tip.


EDIT: SOLDERING IRON TIP TEMPERATURE (without preheater)

Here if I use a large tip I need 420-440c tip temperature
If I use the tiny pointed tip it's 480c

 

[Grog6]

OK; I ended up using the big iron, a heat gun as a preheater, and my biggest soldapult solder sucker, and I cleared about half the holes.

"Thermal Armor" indeed, lol.

I've used the method you mention, but with multilayer boards you run the risk of pulling the barrels out of the via holes.

If the connection to the interstitial layers goes bad, the board is dead.

At least these don't use the "Shadow Process" to make vias, those would be gone already.

 

[infrared]

I have a great soldering iron for times like this, it was my grandads, and then my dads, looks like like a little fire poker but the tip is a big chunk of copper (~250g) that you heat with a blow torch, haven't found anything that puts up any resistance against that! Good luck with it Grog

 

[delshay]

Forth & final level capacitors in.

Custom soldering iron tip required to get them in. The idea was to add three, but I forced in five. Two are turned on their side.
It's not possible to add another stack as it will hit the back of the PC case, which is a pity as I have found 10uf 3v capacitors 0603.

 

[delshay]

When the board is switching 100A in a few microseconds, a little bit of capacitance can make a big difference.

Inductance is the big problem; a trace 0.1" long is right at 1nH, and can drop quite a bit of voltage on a fast current change.
Luckily, there are a bunch of traces in parallel, one per pin.

A via 0.010" can be 1 milliohm, and 1.5nH inductance, easily.

So the closer the capacitance is to the processor pin, the more effective it is.

These are really nice for high frequency:

JMK107B7225MAHTR Taiyo Yuden | Capacitors | DigiKey

Order today, ships today. JMK107B7225MAHTR – 2.2 µF ±20% 6.3V Ceramic Capacitor X7R 0603 (1608 Metric) from Taiyo Yuden. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

These are good, but they don't last as long, 1000hour rated.

F380J476MMAAXEH3 KYOCERA AVX | Capacitors | DigiKey

Order today, ships today. F380J476MMAAXEH3 – 47 µF Molded Tantalum Polymer Capacitor 6.3 V 0603 (1608 Metric) 300mOhm @ 100kHz from KYOCERA AVX. Pricing and Availability on millions of electronic components from Digi-Key Electronics.

www.digikey.com

UPDATE:

First I was wrong about MOSFETS not getting hot, their very hot, but one phase is running cooler than the others, so i'm not sure if it's even working.

Most important is the PSU cable. This is getting warm when running P95. Did not notice it until I accidently touch it with the back of my hand. I'm not going to cut the power, in-fact i'm already working on a fix so I can increase the power. PSU fan is also trying to spin-up faster.

MOSFETS are 70A, but PDF docs states it can do 100A. Not sure what the power output is at this time, but I just want more power.
FX-60 is already doing 3.1GHz with p95 & it is clear it can go faster with p95.

The MOSFETS I have fitted to this motherboard have rapid rise & fall time (5ns) with very quick turn on/off time. Upgraded MOSFETS are fitted to CPU/Northbridge/Southbridge/Hypertransport/Memory & some other minor rail.

I have already started to added those 10uf 3v Tantulum 0603 to northbridge/southbridge & Hypertransport.

PSU: Antec TrueControl 550W

 

[Grog6]

Have you got an oscilloscope?

I'd look at the inductor waveforms, and then the gates on all the mosfets.

See if there's a gate resistor that's cracked, or open.

The 4 phases on that chip should be 90 degrees from each other, and it uses 2 mosfets per phase, but the 4 mosfets on the lower side won't run as hot, IMHO.

See how the mosfets are connected to the chip on P23 of the PDF file above.

Half the mosfets being cool may not be an issue; on the P5wd2 premium I'm working on, the lower mosfets aren't heatsinked, but I'm adding small ones.

If you hit the current limit for the switching, it will turn off the phase early; you can go too much capacitance.

Take a look at Table 10 in the pdf; see what kind of current limit it's running.


If the cables get hot from the PSU, you can buy pins and wire and make your own cables; they have pins that handle 16AWG wire, but each pin is only rated 9A.
That's why there are multiple wires per voltage. Most PSU's use 18awg cable.
Solder another connector on the bottom of the board, on a short pigtail, if you need more current than the pins can handle.

Here's a link to all the different power supply cables, pinouts, ans Pin numbers:

All about the various PC power supply cables and connectors

All about the various PC power supply cables and connectors

www.playtool.com

These pins are for 16awg wire:
Molex 39-00-0168 Fem WM3115-ND at digikey