Not using THAT OPA -- but OPA choice is besides the point IMO.You can stick more caps or ferrite beads to help the ANAL in you
(I posted on ferrites and SMPS elsewhere).
Hi Frog
Very nice project! I can understand from your picture that you have plenty of place for an "external" PSU for the STX and you will not have to place that PSU/transformer directly on the STX board, as I did in my computer. So, will be more easy for you to apply those mods...
Now some more details: First, one will use +/-15v if the final opamp and I/V converter have this working/recommended supply tension. If is about an opamp which will work with +/-12V, then that 15v will be 12... I will advice to build a power supply which can provide a variable/adjustable output, so it can be used for different opamps in final or I/V. It is also very important to have those + and - 15v as equal as possible, or just identical. This +/- PSU have to provide at last 100mA. Better to be dimensioned to 300mA very stable and clean power. Is also important the connection of the GND rail (shorter as possible, and very consistent wire) from your new PSU on the STX board. I think you can have an idea about from my pictures in earlier posts...
Yes, is right as it showed in that picture about how to insert the new +/-PSU
You may note that the only main function of that 12v DC from the computer (Molex) is to provide power to the 5v regulator on the board. As I said before, this 12v is also provide the information to the processor that everything is all right with the power for the entire STX board. You will see that the 12 rail on the board goes to an 7805 regulator (near the clock). Here you can use an 7808 regulator, and then an much better 5v regulator. Shunt or something else with lower noise and so on. This for the clock. From the same 8v you can use another regulator only for the DAC analogue side. Consistent decoupling/filtering capacity right on the pin of the DAC chip or the oscillator is also very recommended.
From the same 8v regulator, you can feed an 3,3 regulator for the processor and the DAC digital side.
If you want, you can just build an high quality 12v, 5v, 3,3v serial PSU and disconnect completely the STX power from the switching PSU for the motherboard you have there. Then you need only to synchronize this external PSU with the computer power up sequence. This can easy be done by an relay controlled by the 12v or 5v computer rails.
To confirm your understanding:
- For seperating the 5V for the DAC and Clock
You cut the output pin of 7805, and use that pad to insert your 5v. You do not have to cut the 5v wire from Molex, but just isolate it that pin (cut the trace) on the board (output of the female Molex).
- For seperating the 3.3V Rail
Yes, remove that ferrite bead. Insert the 3,3v near to the power pin of the DAC chip (but before the ferrite bead placed on that trace). Use here a good/consistent decoupling capacity. I used here ceramic caps 100µ/6v, as on the 5v rails. Is right that picture of my final board is full of texts and not so easy to see the details...
I`m not very sure what you mean about" H6 module"...
You may keep in mind that 3,3v and 5 v have to come from a very good analogue PSU or high quality regulators. Better to have local regulators for every stage/circuit (DAC, clock, and so on). Using an 8v regulator between 12v and the rest of the lower volt rails will improve much the entire quality of the power for the STX board. The +/-15v will supply too the analogue circuits for the SPDIF out/in section through internal connections (layers) inside the board.
I will strongly recommend to proceed step by step with your mods, and observe what is the result of it. So, fix up first the power system. Fix first the 5v stage, then go to the 3,3v. Start the device and see how its working (hear it) after every step. Then proceed to the next mod (I/V converter for example). Then final stage. And so on. Make a plan and follow it. The same "strategy" is about removing and replacing the original components, capacitors on the board, adding of the new better capacitors, and so on. Test the result of that last made mod, and then go to the next step. In this way one can have a very good control, and what can be wrong can be easily fixed up.
Very nice project! I can understand from your picture that you have plenty of place for an "external" PSU for the STX and you will not have to place that PSU/transformer directly on the STX board, as I did in my computer. So, will be more easy for you to apply those mods...
Now some more details: First, one will use +/-15v if the final opamp and I/V converter have this working/recommended supply tension. If is about an opamp which will work with +/-12V, then that 15v will be 12... I will advice to build a power supply which can provide a variable/adjustable output, so it can be used for different opamps in final or I/V. It is also very important to have those + and - 15v as equal as possible, or just identical. This +/- PSU have to provide at last 100mA. Better to be dimensioned to 300mA very stable and clean power. Is also important the connection of the GND rail (shorter as possible, and very consistent wire) from your new PSU on the STX board. I think you can have an idea about from my pictures in earlier posts...
Yes, is right as it showed in that picture about how to insert the new +/-PSU
You may note that the only main function of that 12v DC from the computer (Molex) is to provide power to the 5v regulator on the board. As I said before, this 12v is also provide the information to the processor that everything is all right with the power for the entire STX board. You will see that the 12 rail on the board goes to an 7805 regulator (near the clock). Here you can use an 7808 regulator, and then an much better 5v regulator. Shunt or something else with lower noise and so on. This for the clock. From the same 8v you can use another regulator only for the DAC analogue side. Consistent decoupling/filtering capacity right on the pin of the DAC chip or the oscillator is also very recommended.
From the same 8v regulator, you can feed an 3,3 regulator for the processor and the DAC digital side.
If you want, you can just build an high quality 12v, 5v, 3,3v serial PSU and disconnect completely the STX power from the switching PSU for the motherboard you have there. Then you need only to synchronize this external PSU with the computer power up sequence. This can easy be done by an relay controlled by the 12v or 5v computer rails.
To confirm your understanding:
- For seperating the 5V for the DAC and Clock
You cut the output pin of 7805, and use that pad to insert your 5v. You do not have to cut the 5v wire from Molex, but just isolate it that pin (cut the trace) on the board (output of the female Molex).
- For seperating the 3.3V Rail
Yes, remove that ferrite bead. Insert the 3,3v near to the power pin of the DAC chip (but before the ferrite bead placed on that trace). Use here a good/consistent decoupling capacity. I used here ceramic caps 100µ/6v, as on the 5v rails. Is right that picture of my final board is full of texts and not so easy to see the details...
I`m not very sure what you mean about" H6 module"...
You may keep in mind that 3,3v and 5 v have to come from a very good analogue PSU or high quality regulators. Better to have local regulators for every stage/circuit (DAC, clock, and so on). Using an 8v regulator between 12v and the rest of the lower volt rails will improve much the entire quality of the power for the STX board. The +/-15v will supply too the analogue circuits for the SPDIF out/in section through internal connections (layers) inside the board.
I will strongly recommend to proceed step by step with your mods, and observe what is the result of it. So, fix up first the power system. Fix first the 5v stage, then go to the 3,3v. Start the device and see how its working (hear it) after every step. Then proceed to the next mod (I/V converter for example). Then final stage. And so on. Make a plan and follow it. The same "strategy" is about removing and replacing the original components, capacitors on the board, adding of the new better capacitors, and so on. Test the result of that last made mod, and then go to the next step. In this way one can have a very good control, and what can be wrong can be easily fixed up.
Last edited:
Coris:
With your PSU, how do you...
-- Simultaneously startup PC and sound card?
-- If they have independent pwr switches, are they sufficiently isolated from each other so no current rush can "leak" from one (sound card) to another (e.g., rest of PC)? E.g., stored energy in caps?
Just a note about my preference/support for better (third-party) SMPS for better sound-card performance as an alternative (given cost, hassle of build-time, etc.) ... those better third-party SMPS's (Antec, Seasonic, Nexus, etc. [see my prev. posts for links and details]) also isolate/clean-up the other PC sections (memory chips, processors, HDDs. motherboard components, video card) so that lesser garbage reaches the sound system. And you get the added bennies of better perf./reliability in those non-audio sections.
Bottom line: you can DIY-build a sound-card PC PSU (e.g a linear PSU as in this thread -- quite a time/effort/$$/resource-eating project IMO), and it may indeed lead to better overall sound and a fulfilling, didactic, skill/craft/art-improving experience. But on the grand scale of things, the picture is much more complicated. IMO, it's all about making better use of time/$$/resources/effort while maintaining a DIY "theme".
With your PSU, how do you...
-- Simultaneously startup PC and sound card?
-- If they have independent pwr switches, are they sufficiently isolated from each other so no current rush can "leak" from one (sound card) to another (e.g., rest of PC)? E.g., stored energy in caps?
Just a note about my preference/support for better (third-party) SMPS for better sound-card performance as an alternative (given cost, hassle of build-time, etc.) ... those better third-party SMPS's (Antec, Seasonic, Nexus, etc. [see my prev. posts for links and details]) also isolate/clean-up the other PC sections (memory chips, processors, HDDs. motherboard components, video card) so that lesser garbage reaches the sound system. And you get the added bennies of better perf./reliability in those non-audio sections.
Bottom line: you can DIY-build a sound-card PC PSU (e.g a linear PSU as in this thread -- quite a time/effort/$$/resource-eating project IMO), and it may indeed lead to better overall sound and a fulfilling, didactic, skill/craft/art-improving experience. But on the grand scale of things, the picture is much more complicated. IMO, it's all about making better use of time/$$/resources/effort while maintaining a DIY "theme".
Last edited:
hollowman
About how simultaneously start up PC and sound card PSU`s is just described in my last post as in may previous ones. Is about to use an relay. They who do not know how to use an relay to do this task, maybe should find another field of interest, go fishing or so...
I have also an suggestion for you: If you can see so well that picture of " the grand scale of things", just do it so further, use better your time and resources and leave alone they who have something else to do or discuss here in this thread.
All your interventions here on this subject shows to me quite clear that you are out of this theme. You hardly understand what all this is talking about here, and your criticism has now substance... You just "talk" in your posts about something else.
Why you do not start your own thread to discuss your problems, about your computer, your founded solutions and so on, or to show to al other that you have a better theme than this one here? You may answer at this question for your self. I personally do not care about such answer.
Sorry, but I think I will use better my time from now on, than answer to your interventions...
About how simultaneously start up PC and sound card PSU`s is just described in my last post as in may previous ones. Is about to use an relay. They who do not know how to use an relay to do this task, maybe should find another field of interest, go fishing or so...
I have also an suggestion for you: If you can see so well that picture of " the grand scale of things", just do it so further, use better your time and resources and leave alone they who have something else to do or discuss here in this thread.
All your interventions here on this subject shows to me quite clear that you are out of this theme. You hardly understand what all this is talking about here, and your criticism has now substance... You just "talk" in your posts about something else.
Why you do not start your own thread to discuss your problems, about your computer, your founded solutions and so on, or to show to al other that you have a better theme than this one here? You may answer at this question for your self. I personally do not care about such answer.
Sorry, but I think I will use better my time from now on, than answer to your interventions...
PC PS philosophy 101
Seriously ... I think you are promoting a misdirection in PC power-supply DIY. One major flaw with it is that it's way too labor intensive.
Another relay? IMO: Rube-Goldberg approach. Not ONLY that but you didn't address the isolation issue AFAIK (see prev. msg).
I've never posted NON-topically WRT your PSU thread. I have criticized, yes, and if your project is robust enough to withstand my criticism ... it may lead to improvements and evolution on all fronts (this includes your skill set as a debater -- don't look a gift-horse in the mouth).
Isn't it better to have the "enemy" on your own home turf? You, then, at least know where "they" stand (if it were my own thread, you may never know about that, given the billions of threads on the Net). This (my posting/criticizing in "your" thread) can help you strategize -- you don't have to necess. change one aspect of your PSU schema or topology. It may be simply down to VERBAL debate. Who can "market" their ideas and opinions more effectively.
Seriously ... I think you are promoting a misdirection in PC power-supply DIY. One major flaw with it is that it's way too labor intensive.
View attachment 283887
Hi Coris
I will start with the +-15V for I/Vs etc. Next step is to seperate the 5V rail for analog 5V from the DAC/Clock.
For the 5V rail i want to use the TL431 Shuntregulator+LM317 but with his own transformator.
See datasheet s30 precision 5V 1,5A regulator.
TL431 Datasheet
The 3.3V seems not so easy. In my case(I have the PCI Version) there is no ferritebead to desoldering(see picture).
Also i have take some measurements at the molex connector. On the 5V line flows 550mA and at the 12V line ca. 135mA.
You say the only think for the 12V line(from molex connector) is to supply the 5V for the DAC. Where becomes the 7812 his power ? When my measurements are right it comes from the 5V from the molex connector but this is a little bit crazy...
Greets,
Frog
I will start with the +-15V for I/Vs etc. Next step is to seperate the 5V rail for analog 5V from the DAC/Clock.
For the 5V rail i want to use the TL431 Shuntregulator+LM317 but with his own transformator.
See datasheet s30 precision 5V 1,5A regulator.
TL431 Datasheet
The 3.3V seems not so easy. In my case(I have the PCI Version) there is no ferritebead to desoldering(see picture).
Also i have take some measurements at the molex connector. On the 5V line flows 550mA and at the 12V line ca. 135mA.
You say the only think for the 12V line(from molex connector) is to supply the 5V for the DAC. Where becomes the 7812 his power ? When my measurements are right it comes from the 5V from the molex connector but this is a little bit crazy...
Greets,
Frog
Last edited:
We MIGHT trust you more if you, e.g., posted some quantitative metrics...e.g., o'scope screen captures.
Hi
First I have to correct my quite confusing phrase in my previous post about the role of 12v rail in powering the board. My thinking were to say that after one isolate the final and I/V stages from the board power line, then the only load for the 12v rail is only to create 5v for the involved circuits. This is for STX, but for ST may be a little bit different. You may find out self about this. Else, of course in normal way this 12v rail is used to feed the 7812, and to feed 7912 for the minus side needed of the opamps.
How is quite easy to be understand from my posts here, I`m referring in this thread mainly to the Esence STX, which I had, I have and I modified at least. I did not mod an ST board yet, so I can not say that I`m very known with details about ST board, but the DAC stage, analogue output stages, clock are the same as in STX. It is possible that 3,3v is made from 5v rail on an ST board, and that because the current on that 5v rail is so high. You may do a research to find out about more precise details on this field.
Something else I wish to precise is that this mod project involving my STX board is closed down for to day. I`ve done it, it works just fine, very satisfied with the results, and ended. I found out afterwords that PCM1792 have some limitations, but it have a very big potential if the rest of the system is how it have to be around this chip. Now my interest are focused in a little bit another direction, but in the same field of high fidelity, DACs and so on.
This STX mod project become quite old for me now, and that because is possible that I will not be as accurate as in that time I was involved and I was working actually on this mod. I will recommend that you read carefully my previous posts. It can be there informations, which I can hardly remember right now, today.
As my time is also limited, I can not go back to may old posts in this thread to verify if an answer to a question is very accurate or in conformity with what were wrote before. I just do not have that time to do whole this work, and as I said, this project is closed for me now.
So, you may please read carefully my previous posts, specially about the 5v rail. I can remember that this rail is also used to a small circuit, which send too informations about the power on the card, to the main processor. I`m a little bit unsure now about if is very well to just cut the the 5v wire from Molex connector or isolate the input pin on this rail. But is not any problem if one just lift up the output pin of the 7805 regulator, to insert his own 5v rail. That little circuit which is possible to use the 5v rail from computer and send information to the processor, is for sure before that 7805 regulator. So, do a research on your ST board to find out more precise about this.
Referring to the 3,3v rail, you can very well cut this line which goes to the DAC (digital side), and feed the DAC only with an 3,3v regulated from your own 5v rail with a shunt or another high precision/quality regulator. The current need it only to feed the DAC chip will so be quite low and you will not need for a heat sinks or big components for this task.
As a principle about modified the power supply system for this board circuits, you can let as it is the original power system, and cut/remove the power only for the targeted chip (DAC, clock, final & I/V opamp), and insert there your quality regulator. This will be less risky, and can bring the best results. To be everything much easier, I will recommend you a very good regulated and high parameters 8 - 10 v PSU beside the board in your enclosure. From this PSU you can well feed a dedicated 5v regulators for the DAC and another one for the clock. From one of those 5v regulators, you can make 3,3v for the DAC digital side. Those dedicated regulators will be quite small and can be placed as close as possible to the targeted chip. So you can have at last a PSU for +/-15v and another one for an mainly 8 - 10v rail. Use an relay to synchronize those two PSUs with the computer start up sequence, and that`s all!
Last edited:
As a note to my previous post, I wish to precise the follow for those who come in in this thread:
This thread is meant to provide ideas and suggestions as some results and impressions about modifications of this sound card. The rest of the work belong to they who will find those ideas here worth to proceed to it. I do not need and even care about one or another will believe me or not. Read what is here, think self, do it self, and assume your own risks about this work. Find your own way in this work. Do one come here with own informations, and results, this is only positive and they are welcome. In this way the thread will develop it self and will can bring more informations to more interested people.
I have no intention to prove something to one or another one here, and I will not use my time to present here all kind of detailed informations. Find your self what you need!
I think I`ve did enough publishing much enough informations, detailed pictures and I answered to they who asked about.
This thread is meant to provide ideas and suggestions as some results and impressions about modifications of this sound card. The rest of the work belong to they who will find those ideas here worth to proceed to it. I do not need and even care about one or another will believe me or not. Read what is here, think self, do it self, and assume your own risks about this work. Find your own way in this work. Do one come here with own informations, and results, this is only positive and they are welcome. In this way the thread will develop it self and will can bring more informations to more interested people.
I have no intention to prove something to one or another one here, and I will not use my time to present here all kind of detailed informations. Find your self what you need!
I think I`ve did enough publishing much enough informations, detailed pictures and I answered to they who asked about.
frog993, let me explain how power supply is arranged. Strart with molex +12V and + 5V
1.) +12V thru three 3.3 Ohm resistors feeds the 7805, that in his turn supplies analog part of DAC and auxilary consumers (ADC and relays)
2) +5V is converted by the HF invertor into +15V, that feeds the 7812 and into -12V (in fact, it would be better to have -15V and 7912 then, while there is place for 7909 on the PCB). Thus, we have splitted power supply (+/-12V) for I/U and output stage as well as for the headamp.
+3.3V comes directly from the PCI and feeds digital part of PCM as well as DSP.
The picture belows shows the power paths.
1.) +12V thru three 3.3 Ohm resistors feeds the 7805, that in his turn supplies analog part of DAC and auxilary consumers (ADC and relays)
2) +5V is converted by the HF invertor into +15V, that feeds the 7812 and into -12V (in fact, it would be better to have -15V and 7912 then, while there is place for 7909 on the PCB). Thus, we have splitted power supply (+/-12V) for I/U and output stage as well as for the headamp.
+3.3V comes directly from the PCI and feeds digital part of PCM as well as DSP.
The picture belows shows the power paths.
are you more Wittgenstein?
Have trouble trudging thru your 2nd-lang. English. But your English is better than my Norwegian.
The problem with your "picture" approach is that it cannot be Search Engined (or searched via Find util.). You may need to re-read Orwell's 1984 and Bradbury's Fahrenheit 451 ... the picture/comic approach has it's limits. Or are you more Wittgenstein?
Letter grade for above: C+
Have trouble trudging thru your 2nd-lang. English. But your English is better than my Norwegian.
The problem with your "picture" approach is that it cannot be Search Engined (or searched via Find util.). You may need to re-read Orwell's 1984 and Bradbury's Fahrenheit 451 ... the picture/comic approach has it's limits. Or are you more Wittgenstein?
Hello andreudze
Thanks for your picture and explaining the powerpath. This explains why on the 5V line flows so much current.
When cutting the 3.3V powerpath from the PCI bus(see picture) the DAC is the only consumer on the 3.3V line there ?
Have you build seperate supplys for your xonar soundcard ?
Thanks for your picture and explaining the powerpath. This explains why on the 5V line flows so much current.
When cutting the 3.3V powerpath from the PCI bus(see picture) the DAC is the only consumer on the 3.3V line there ?
Have you build seperate supplys for your xonar soundcard ?
Last edited:
Frog:
About your attached image ... is this a SCANNER photo? (i.e. you put the Xonar on the flatbed of a scanner)
Nice work on labeling and tracing.
This is definitively the most important in this case: how the picture were made...
Thanks for your valuable contribution to this thread!
Scanner photography phil. 101
Actually, yes, but not in the way you may be thinking (= not outside the box).
Many will reach for their DSLR or iPhone or other digital camera to snap a shot when -- if they have a tradit. flatbed scanner on their desk -- right there in front is the easiest, highest-quality solution. And it can produce VERY high-quality results.
scanography- Scanner Art
etc.
Similarly, some will take a linear-topology, or wheel-reinventing approach to a PC sound-card PS -- which involves tons of labor, $$, man-hours, soldering fumes, employing risky (floating ground) topology, etc. -- when an excellent-quality (SQ-wise), much more reliable, comes-with-warranty, little-fuss option is avail. The better-rated (and tested, using objectives science and methodology, publicly disclosed) SMPS I noted earlier.
The PS in your design involves risky design choices, and risky trace-cutting of the orig. SC. Also, you provide little or no objective measurements of your results.
In the end none of my criticisms matter if some of the primary goals of DIY are met: e.g., Pedagogy and Didacticism -- actually learning/acquiring/retaining NEW knowledge skills and science. I don't find much of that that in your project.
I do find some of that at better DIY PC/hardware sites:
Hardware Secrets Power Supply Test Methodology | Hardware Secrets
Seasonic X-Series 650 W Power Supply Review | Hardware Secrets
etc.
This is definitively the most important in this case: how the picture were made...
Thanks for your valuable contribution to this thread!
Actually, yes, but not in the way you may be thinking (= not outside the box).
Many will reach for their DSLR or iPhone or other digital camera to snap a shot when -- if they have a tradit. flatbed scanner on their desk -- right there in front is the easiest, highest-quality solution. And it can produce VERY high-quality results.
scanography- Scanner Art
etc.
Similarly, some will take a linear-topology, or wheel-reinventing approach to a PC sound-card PS -- which involves tons of labor, $$, man-hours, soldering fumes, employing risky (floating ground) topology, etc. -- when an excellent-quality (SQ-wise), much more reliable, comes-with-warranty, little-fuss option is avail. The better-rated (and tested, using objectives science and methodology, publicly disclosed) SMPS I noted earlier.
The PS in your design involves risky design choices, and risky trace-cutting of the orig. SC. Also, you provide little or no objective measurements of your results.
In the end none of my criticisms matter if some of the primary goals of DIY are met: e.g., Pedagogy and Didacticism -- actually learning/acquiring/retaining NEW knowledge skills and science. I don't find much of that that in your project.
I do find some of that at better DIY PC/hardware sites:
Hardware Secrets Power Supply Test Methodology | Hardware Secrets
Seasonic X-Series 650 W Power Supply Review | Hardware Secrets
etc.
Last edited:
For one or another reason you have a high DC level on the output of that channel (where you register that spike). If you still have the isolation caps on the output (just before RCA connectors), then it happened something wrong when you changed out that I/V caps. Here we talking about decoupling caps of the involved opamp. Review once again if everything is all right (traces, no soldering drops between components), and if the issue persist, then just resolder the original caps.
One may be very carefully with SMD components. It get destroyed quick if the soldering time is too long (not more than 2 s), or the temperature of the soldering iron is not adequate. Specially, the isolation material used in SMD caps are very sensitive to temperature excess.
You may also measure the power tensions on the respective caps to confirm if OK. Measuring on this board while it is placed in to the motherboard is very difficult. I will recommend to find an (no matter quality, but DC filtered) 12vDC power supply and an 5v one, use an Molex connector to supply the power (simultaneous12v and 5v) to the board outside of the computer (on the bench). Then measure the power rails, the tension on the decoupling caps, and (important) the output levels on RCA connectors. On RCA should not be any DC level (0 or 0,something). If one removed already the isolation caps, then this DC levels on output RCA (offset) are very important to be very near 0 and equal as possible between channels.
If you may take and publish a picture of your board after cap changing, it will be better to understand what about...
Last edited:
A worthy project?
How much risk/effort makes a PC-based project WORTHY?
Constantly (de)-coupling (de/re-energizing) of the PSU and/or sound boards -- or any other similar components for that matter -- from the chassis or motherboard is not without important risk or effort.* PC DIY -- unless the setup is very strategically located on a workbench -- is too manual and mechanical.
Maybe you have deep pockets, or unlimited time, or REALLY enjoy "DIY" ... so these risks are unimportant.
* Personally ... I don't even like to turn PCs/laptops off as I feel cycling power stresses out PSUs and other PC components; and the often-used daily/nightly shut-downs and re-power-ups are further sources of electrical, mechanical and thermodynamic stress. [Simply go to any PC repair shop and poll the most-repeated time of failure (or research this issue/topic online). Answer: power-up. Indeed my most reliable PC or laptops are "always on", and only cycled if/when absolutely necessary or beyond my control ("Act of God", Rolling blackouts, etc.)
As an alternative to the PS project, why not invest the $$/effort into better (DIY if you like) power/line-conditioning projects?
Coris noted ..."If you may take and publish a picture of your board after cap changing, it will be better to understand what about..."
How much risk/effort makes a PC-based project WORTHY?
Constantly (de)-coupling (de/re-energizing) of the PSU and/or sound boards -- or any other similar components for that matter -- from the chassis or motherboard is not without important risk or effort.* PC DIY -- unless the setup is very strategically located on a workbench -- is too manual and mechanical.
Maybe you have deep pockets, or unlimited time, or REALLY enjoy "DIY" ... so these risks are unimportant.
* Personally ... I don't even like to turn PCs/laptops off as I feel cycling power stresses out PSUs and other PC components; and the often-used daily/nightly shut-downs and re-power-ups are further sources of electrical, mechanical and thermodynamic stress. [Simply go to any PC repair shop and poll the most-repeated time of failure (or research this issue/topic online). Answer: power-up. Indeed my most reliable PC or laptops are "always on", and only cycled if/when absolutely necessary or beyond my control ("Act of God", Rolling blackouts, etc.)
As an alternative to the PS project, why not invest the $$/effort into better (DIY if you like) power/line-conditioning projects?
Last edited:
Yes, something is wrong with those 3,2v... It is possible that one of the caps you changed is conducting, and so is unbalanced the +/- power rails. I do think that is not a catastrophic fault... Just verify again the "operation area", and take a look at the +/- rails on I/V stage. I hope you will find the fault.
Is better to not reuse the SMD components after desoldering...