Jump to content

mariushm

Member
  • Content count

    4,091
  • Joined

  • Last visited

Awards


This user doesn't have any awards

About mariushm

  • Title
    Veteran

Profile Information

  • Gender
    Male
  • Location
    Romania
  • Occupation
    Freelancer

System

  • CPU
    AMD FX-8320
  • Motherboard
    Gigabyte GA-990FXA-UD3
  • RAM
    16 GB DDR3 1600 Mhz Low Profile
  • GPU
    Gigabyte Aorus RX 570 4 GB GDDR5
  • Case
    Aerocool XPredator Black Edition (Full Tower)
  • Storage
    128GB Sandisk X400+4TB HGST-NAS+2TB-WD+1TB-WD
  • PSU
    Seasonic X-650 80+ Gold
  • Display(s)
    Samsung T240 (1920x1200 24") + Samsung 2494HM (1080p 24")
  • Cooling
    Zerotherm FZ-120 w/ Nexus RealSilent 120mm fan
  • Keyboard
    Microsft ComfortCurve 3000
  • Mouse
    Logitech MX518
  • Sound
    Logitech X-540 (5.1) + ALC889 onboard
  • Operating System
    Windows 7 Home Premium
  1. What the difference between NVME and m.2

    sata 3 is 6 gbps 6gbps means 6 billions of bits per second going through the cable. There's 8 bits in each byte, so that's 750,000,000 bytes per second. But, for every 8 bits of actual data, there's 2 bits used for error correction, so you actually have a maximum theoretical of 750,000,000 x 8 / 10 = 600,000,000 bytes per second of USABLE DATA. If you use multiples of 1024 as Windows uses when it shows the amount of disk space used, that means a maximum theoretical of 585,937 KiB/s or 572.204 MiB/s But note that just like with USB and other protocols, there's overhead.. there's communication between the sata controller and the drive besides raw data being transferred ( give me data that's in sector 123456 and keep going, ok sent you the data, ok i got the data) and some data "packets" must be a certain fixed size, so there's a few bytes of padding (filler bytes, unused) and so on.... so for very long single file transfers this overhead is very small, for a lot of small files this overhead can be more substantial (but still small, let's say less than a megabyte per second out of <600 megabytes per second) Basically, 6 gbps means you can reach around 550-580 MB/s. USB is even worse, it's practically impossible to actually reach 480mbps (for usb 2.0) or 5 gbps (for usb 3.0) due to the way data is arranged in packets and sent through the cable.. even though physically the wires can transfer 5 gbps worth of bits, by design a device can't really figure out a way to arrange the data it sends to the usb controller in a way that would make it possible to reach that 5 gbps limit. If my memory is correct, I think the maximum possible for usb 3.0 was somewhere around 4.7 gbps ps. The nvme (pci-e) portion of the m.2 connector can use pci-e v2.0 lanes (500 MB/s per lane) or pci-e v3.0 lanes (~970 MB/s per lane max) so a SSD connected to a m.2 connector could have a maximum of 2 GB/s if the lanes are v2.0 , or could have up to nearly 4 GB/s speed. There's also SSDs that can use only 1 pci-e lane out the maximum 4 in the m.2 connector, or on some motherboards you can reduce the number of lanes in the m.2 connector from 4 to a lower number, in order to enable other pci-e slots or give more lanes to other devices on the motherboard. For example if you have a 512 GB TLC drive and you know you won't read and write from it at 2 GB+/s you may want to artificially restrict it to pci-e v3.0 x1 which would still give you up to 1 GB/s in speed.
  2. What the difference between NVME and m.2

    m.2 is the connector, which has the contacts for a pci-e v3.0 x4 "slot" , a SATA connection and other things (usb). A m.2 device can use either sata or pci-e side of the connector for communication. Some budget SSDs may only use the SATA part of the connector for communication, which means max ~ 550-580 MB/s instead of up to ~4 GB/s.
  3. AM2 socket and GTX 550 Ti

    It will run OK. With so old hardware the only issue you may face is having maximum pci-e v2.0 slots, while newer systems have pci-e v3.0 slots. But, the video card is backwards compatible with pci-e v2.0 slots and my guess the performance decrease is less than 1%, much less than pci-e x16 vs pci-e x8 difference. Depending on the video card driver, the processor may limit the performance a bit, recent drivers are more optimized for more cores and higher frequency per core. If the games run slow, i suggest trying out some of the first drivers that support that video card model, often later drivers add more features and patches for various games and slow things down etc.
  4. Calcaulating Wattage

    Pretty much that, as a simplification. You'll want to have a power supply that can provide at least that amount of power to the components. A power supply is most efficient when your components load the power supply up to around 60-80% of its maximum power. If your power supply produces less than around 40% of its rating, or it has to produce more than around 80% of its maximum rating, then its efficiency decreases. For example, if your computer components use only around 200 watts, then you could use any power supply, from let's say 400w to 850w ... but a 450w model may be more efficient compared to let's say a 650w model, because its output will be at around the 50-60% load level. The 650w model may be less efficient, but being oversized, it would probably also produce less heat, so the fan it comes with may not even spin, so that 650w model may be more silent. So you see there's tradeoffs Also remember, components in a computer use 3 voltages: 12v , 5v and 3.3v. Most components these days use 12v quite a lot, but very little on 3.3v and 5v. So it's important that the power supply can actually give that amount of power on 12v, not on all three voltages combined. For example, a power supply may output 400w on 12v and 150w on 3.3v and 5v, so it can be labeled as a 550w power supply. BUT, if your components actually need 350-400w from 12v, this power supply would be borderline, for brief moments the components may take more than 400w and the power supply could shut down to protect itself and your components. Also, keep in mind that components don't use the peak power all the time. For example, a video card may use up to 160 watts for example, but in a game like Fortnite which doesn't have super fancy graphics or high fps, the video card may only use 100w. Also, the CPU may use 100w when it's at 100% load for long periods of time, but when you're gaming the CPU won't stay at 100% all the time, the load will fluctuate. So when you're gaming, your power consumption won't be 100w (cpu) + 160w (video card) all the time Also, when you add some components like let's say a DVD writer, you would have to use 10w for example, because that could be the power draw when the unit actually burns discs, spinning them at high speed and using the built in laser to burn the discs, but when the dvd drive idles or just reads discs, the power demand could be less than 1-2 watts.
  5. I want to make a product help me.

    You have to be more accurate with your specifications. What resolution of display? Can you live with 160x120 pixels or 320x240 pixels or do you need 800x600 or something higher? What interface do you want to your display? Do you want colors or can you live with just let's grayscale or black and white display? Can you live with just text and basic graphics or do you want full color, high fps etc etc ? Do you need to add/delete/update tasks through some wireless communication or you want to add through some keyboard or buttons? Do you need a lot of memory for tasks? Basically is 100 KB to 1-2 MB of storage enough for your tasks or do you want something more complex like adding images and music or sounds to each task? If you want something simple like showing a list of tasks and you're fine with less than megabytes of storage, then you could use even something as simple as a 2-3$ microcontroller (arduino or something better like pic16/pic18/pic32 or even some arm cortex chip) and a simple dot matrix lcd display to print lines and basic art work through something simple interface like i2c or spi or parallel i/o. You can do all this on a basic prototyping board, you'll just need the programmer to program the chip. If you need higher resolution or more storage or you want to use a display with hdmi input or something more complex, then it may make more sense to use a raspberry pi and you'll just core the software
  6. The serial or usb output from UPS to PC should be optically (or mechanically, through signal transformers) isolated so it shouldn't matter what's used to power the ups or pc or whatever. Think of it like sending Morse code signals between two hills - the ups can send signals to the PC and the PC can send signals to UPS through the cable, but the two devices are like two separate hills... doesn't matter what powers ups or pc. An UPS will be affected a bit if it's powered from an inverter... well, the charge rate of the internal battery should be affected. Otherwise, it would depend on the UPS technology, the UPS may simply pass through the AC voltage from the inverter to the PC or may tweak it a bit (if it's a good quality line interactive UPS or online UPS) or the UPS may permanently power the PC from internal battery or through an isolating transformer inside the UPS.
  7. How To Relay

    You can buy several kinds of relays. There's a cheaper relay type which is non-latching, which means you have to keep the coil energized to keep relay ON (meaning the secondary contacts connected together and current flowing through them). As soon as the coil is not energized, the relay turns OFF and the secondary contacts no longer let energy flow. There's slightly more expensive relays called latching, which only require you to send a sort of pulse of energy on one of its pins to turn the relay ON (and something mechanically changes inside to have the secondary contacts connected) and then even if you stop that pulse of energy, the relay will be permanently in the ON position. You turn such latching relay OFF by sending another pulse of energy on another pin, for a period of time. So the non-latching relays are let's say 1-2$ and they need maybe 12v at around 30-50mA constantly to keep the coil energized and allow energy to flow to the inverter. You would use a part of your battery power just to keep the relay on. It may make more sense to buy a latching relay which would only need let's say 12v at 50mA for a second to mechanically flip the relay to the ON position, and another such second pulse on another pin to turn it OFF. Here's an example of a cheap non-latching relay that can handle 10A of current on the secondary contacts: digikey link It's $1.3 if you buy one and says it uses 400mW (12v @ ~33 mA) to stay on. Here's an example of a latching relay that can handle 16A of current on the secondary and it's not too expensive at around 4$: digikey link The operation and release time is only 10ms, so you can be fairly sure you turned it on or off if you send power on one of the pins for a bit more than 10 ms, afterwards it stays on or off without having to keep pumping energy in the relay. It needs slightly more energy (around 600mW) and the burst of current would be slightly higher but it's for such a short time that it won't matter. Anyway, what you want to do can be easily done with some general purpose npn transistors and some diodes. It can be done even easier with a microcontroller like an arduino or a PIC. You'll need a transistor with good enough voltage on base pin (at least 30v or so), or you can use 2 resistors (voltage divider) or a zener diode to reduce the voltage on the base pin to some reasonable level. Another option would be to add a diode between the dc-dc converter and the battery and then use a microcontroller to measure the voltage on the left and right side of the diode - on the left you'd have the voltage of the dc-dc converter (so you'd know if there's no voltage, then charger isn't working) and on the right you have the voltage of the battery. You can keep the microcontroller powered 24/7 from the battery, through a small 12v to 5v regulator, since it will basically use mA worth of power while monitoring the two power sources using it's ADC module. This way you can use either a non-latching or a latching relay but a latching one would be great because the micro can send the pulses of energy to turn on or off the relay whenever you think it's needed and not waste energy keeping the coils energized (for example, measure input voltage from dc-dc converter 2-3 times a second and turn off the relay only if the dc-dc converter stops outputting 12v for 30 seconds or so, or turn off the relay only when the battery voltage drops below 10.5v (most relays will work down to around 9v but lead acid batteries can be damaged if voltage goes too low)
  8. Desoldering from a heatsink pcb?

    Get a hair dryer or ideally a paint stripping gun and heat the whole board from the bottom to heat the whole aluminum board. Put the pcb between two planks of wood or something so you can have the heat source under the board and have the heat gun (whatever you use) at 10-20cm from the circuit board blowing hot air on it. You want to get the aluminum pcb to maybe 70-90c ,, any higher and you risk the plastic case of the leds melting and the small wires connecting the led diode inside the package to desolder from the actual terminals on the packaging. Then once the whole aluminum sheet is warm, ADD solder to the leads of that regulator and the big tab of the chip and once all three ends flow you can lift the regulator off the board. If you don't need the regulator chip anymore, use some wire cutters to cut the two leads first, for less thermal mass... It's a pain to do anything with cheap soldering irons... a soldering station is not that expensive (less than 100$) and it's an investment that will last you 10+ years if you take care of it... it's a few dollars per year practically.
  9. Create a disk image. That doesn't care about file structure, fragmentation, antivirus programs blocking files from being transferred...
  10. why doesn't Apple use AMD cpu's?

    Apple wants to move to their own processors based on ARM , which would allow running same app on iPhone and their Macs and make it possible to "move" running apps between devices etc.. it takes maybe 1y+ to engineer a product line and transition between processors and all the stocks related (cpus sockets boards parts stock for repairs) i suppose it was much easier to keep going with intel and convenient integrated graphics until they get their own cpus working
  11. I need help finding older drivers

    SM Bus controller is related to chipset, it should be part of the chipset drivers. According to manual the motherboard uses nForce2 SPP (Ultra 400) northbridge and nForce 2 MCP-T (Deluxe Model)/MCP southbridge, so you should find the chipset drivers on nVIdia's site (legacy, chipsets etc) or any other motherboard with the same chipsets. Try Drivers , Legacy , nForce 1/2 , select operating system and so on on nvidia.com website. The network controllers ... see manual page 9, there's realtek 8201bl and broadcom ac101L ... the broadcom ones are n the website, if you can't find them listed for your particular OS, try downloading the ones for other OSes, the drivers may be in packages that are not just for the OS they're listed for. The video card... that's nvidia.com ... go and download the one for that series... Audio ... if it's not on Asus website, try nvidia.com , legacy , nforce audio or if that doesn't work, should be in nforce 1/2 chipset drivers ... And again, if I may suggest... you could also simply find a motherboard that uses the same chipset combo but made by other manufacturer, and those drivers should work
  12. 3,5mm audio female pinout

    There aren't 7 pins , there's 3 or 4 pins, and the rest are grounded (connected to the ground pin) and there just for mechanical support, so that jack won't break or wiggle when you insert and pull out the audio cable. The audio jack is normally ===[ GROUND ] | [ LEFT ] | [ RIGHT ] ) so one pin is left channel, one is right channel and at least one is ground (common for both audio channels) If you really want to be exact, you can get a male stereo jack, plug it into the female connector and then use a digital multimeter in continuity mode to figure out which pin is what.
  13. It's an IT store like Newegg but in Romania: https://www.pcgarage.ro/ You can use Google Translate if needed: example link
  14. comparison from a shop i access often differences are highlighted
  15. Forum Hosting Advice?

    Just look at the top 10-50 forums out there and at the bottom of the pages you should see what forum software is used for them. You can use anything, from free stuff like Simple Machines Forum (SMF) to more commercial stuff. It depends also on who's going to use your forum, what's the target.. a more technical audience like let's say circuit designers engineers etc may not care about the forum software because they care about content more (see for example EEVBlog forum which uses free SMF forum software) but for other purposes you may prefer more complex forum scripts for features like signatures, avatars, bling etc ... for example if you want to attract a younger crowd which cares about that, like what's here on this forum. You can start a forum even on a 1-2$ per month shared hosting account, and you can move it to something more professional once you get a lot of members and page views. Forums do a significant amount of database queries, so the higher the member and viewer count, the more the database server will be loaded, and in a shared hosting environment you will see you forum start to become slower once the number of members goes up (and the number of posts).
×