Office PC Trouble!

[Herman Mcpootis]

43 minutes ago, Lefteh said:

heres a cheaper powersupply 80+ bronze and its evga so good warranty. this saves you 3 bucks

65$ 9 or 10 dollars cheaper than current going by the trend it may dip to 60$ aswell

 

no, stop. the BT isn't a good quality PSU, and he's much better off getting a B450M Pro4-F instead.

 

10 minutes ago, Lefteh said:

also check this out!!!! 20$ ssd https://www.amazon.com/Kingston-120GB-Solid-SA400S37-120G/dp/B01N6JQS8C

 

that's one of the slower Dram-less SSDs.

[brob]

1 hour ago, D3monw3st said:

There won't be any beverages, but dust is definitely a problem. The dust collector should be able to keep it from getting in, no?  

 

Filters are not dust collectors. They will keep out larger particles, but not dust.

 

In my experience it is a mistake to cheap out on office builds. People time is much more expensive than a bit of better hardware.

 

Get a good quality psu like the Corsair in the OP. Low-end Evga products are not that good. 

 

Dual channel does make a difference in office applications. It may only be 5% or 10%, but that extra performance can add up to a fair bit of saved man hours.

[Lefteh]

9 minutes ago, Herman Mcpootis said:

no, stop. the BT isn't a good quality PSU, and he's much better off getting a B450M Pro4-F instead.

 

that's one of the slower Dram-less SSDs.

he asked for 100gb and wants price down dramless is fine it may not have as stellar read and writes or last 20 year but it will easily last 10 years. B450M Pro4-F only has 2 more ram slots and he only needs 2 and same vrm 4 + 2 same features just 5 dollars extra. As for the power supply there perfectly fine ive never had one one them fail on me yet and Ive had 3 for a year now. His system will barrly pull 120 watts.

 

[Herman Mcpootis]

4 minutes ago, Lefteh said:

B450M Pro4-F only has 2 more ram slots and he only needs 2 and same vrm same features.

have you actually taken a look at the board? the HDV R4.0 doesn't even have a VRM heatsink. that's a major difference in temperatures. 

[Lefteh]

does not matter if your not overclocking.

heres a xeon 1620 130w tdp no vrm block and has just a little airflow and it never overheats.

[brob]

5 minutes ago, Lefteh said:

does not matter if your not overclocking.

heres a xeon 1260v 135w tdp no vrm block and has just a little airflow and it never overheats.

 

 

ark.intel.com does not list a xeon 1260v.

[Lefteh]

Just now, brob said:

 

ark.intel.com does not list a xeon 1260v.

i corrected it i had it flipped and its 130w

[Lefteh]

[D3monw3st]

32 minutes ago, Lefteh said:

ALSO heres a cheaper case its 19$ way cheaper than your current case. https://www.amazon.com/ROSEWILL-Computer-computer-transparent-FBM-X1/dp/B072F29JP4/ref=sr_1_1?fst=as%3Aoff&qid=1560577227&refinements=p_36%3A1253503011&rnid=386442011&s=pc&sr=1-1

Oh dang, this case is really cost effective. No PSU shroud, but does that matter? Also, I really like the SSD: O

[Lefteh]

i dont think psu shrouds matter its just aesthetics(and some cable management) heck you could totally have no case but it would not be convenient to move around or start up or external usb.

[D3monw3st]

27 minutes ago, brob said:

 

Filters are not dust collectors. They will keep out larger particles, but not dust.

 

In my experience it is a mistake to cheap out on office builds. People time is much more expensive than a bit of better hardware.

 

Get a good quality psu like the Corsair in the OP. Low-end Evga products are not that good. 

 

Dual channel does make a difference in office applications. It may only be 5% or 10%, but that extra performance can add up to a fair bit of saved man hours.

Ah, I see. I currently use a higher end EVGA PSU and it has given me no troubles which is why I assumed that would be a better switch.

[D3monw3st]

Just now, Lefteh said:

i dont think psu shrouds matter its just aesthetics heck you could totally have no case but it would not be convenient to move around or start up or external usb.

Oh, I see. I know that in the current office computers they seem to have some sort of shroud around it, just wanted to ask

[Herman Mcpootis]

34 minutes ago, Lefteh said:

i corrected it i had it flipped and its 130w

TDP is a heat measurement, not wattage consumed. 

[Lefteh]

1 minute ago, Herman Mcpootis said:

TDP is a heat measurement, not wattage consumed. 

As a form of energy, heat has the unit joule (J) in the International System of Units (SI). However, in many applied fields in engineering the British thermal unit (BTU) and the calorie are often used. The standard unit for the rate of heat transferred is the watt (W), defined as one joule per second. This tells you a good estimate of how much power aka watts aka Heat per second you will consume.  TDP is a good estimate of power consumption denoted in watts at least for older xeons as they stick to a very strick power profile as they dont turbo incredibly high or vary alot.

[Lefteh]

now let me show you something very interesting. My xeon will only go to 87 Watts even though its a 130w part. But how could this be? Simple the e5-1660 the 6 core variant is actually 130w and the 4 core variant is a cut-down with 2 cores disabled. (130/6)*4 = 86.666 Watts

proof of TDPs

intels own website

E5-1660

E5-1620

 

so in this case the TDP does lie, but its not for the reason that you might belive as the TDP is very accurate for these types of cpus. There was a time not to far ago that TDP was actually relevant.

 

[Lefteh]

32 minutes ago, Lefteh said:

As a form of energy, heat has the unit joule (J) in the International System of Units (SI). However, in many applied fields in engineering the British thermal unit (BTU) and the calorie are often used. The standard unit for the rate of heat transferred is the watt (W), defined as one joule per second. This tells you a good estimate of how much power aka watts aka Heat per second you will consume.  TDP is a good estimate of power consumption denoted in watts at least for older xeons as they stick to a very strick power profile as they dont turbo incredibly high or vary alot.

and if you didn't get the good estimate this is referring to 100% load as i hope you already understand this.

[Norwegiantweaker]

2 hours ago, valdyrgramr said:

You quoted yourself.  You might wanna mention or quote the person you're talking to rather than make 3 posts in a row with one talking to yourself.  There is an edit button.

i think he mght actually be talking to himself tho

[XR6]

3 hours ago, Lefteh said:

so in this case the TDP does lie

3 hours ago, Lefteh said:

but its not for the reason that you might belive as the TDP is very accurate for these types of cpus

That makes no sense. How can the TDP be 'very accurate' when you admit that it was incorrect (or at least in that case)?

3 hours ago, Lefteh said:

As a form of energy, heat has the unit joule (J) in the International System of Units (SI). However, in many applied fields in engineering the British thermal unit (BTU) and the calorie are often used. The standard unit for the rate of heat transferred is the watt (W), defined as one joule per second. This tells you a good estimate of how much power aka watts aka Heat per second you will consume. 

What are you on about? That's just random crap you copied from Wikipedia, it doesn't back up your claims about TDP.

TDP has nothing to do with power consumption, it's a heat measurement. Some people do use TDP to calculate power draw however it's an inaccurate method of doing so.

[Jurrunio]

@Lefteh just because Intel and Nvidia use "TDP" as a term for power measurement in many press articles doesn't mean it's the right way of using it. TDP is a unit for heat output and is always a bit lower than power draw, period. If TDP also equals to power usage, then CPUs and GPUs are technically calculating (i.e. doing work) without energy consumption (cuz everything out is in heat) which is silly.

 

What's really happening is that they're trying over-rate such CPU/GPU's heat output so you overbuy the cooler. Well, was. Intel's recent move on super high turbo clocks on their desktop CPUs goes the other way.

 

Also package power numbers are wrong. I'm getting 150-200w reading of power draw through the EPS 8pin powering the CPU while package power reading is just 80w in my system. That couldn't be further from the truth since my CPU clearly kicks out more heat than the VRM.

 

6 hours ago, Lefteh said:

eres a xeon 1620 130w tdp no vrm block and has just a little airflow and it never overheats.

How did you know that? Attaching a thermocouple to one of the low side mosfets while running Prime95 on the CPU?

[brob]

https://www.anandtech.com/show/13544/why-intel-processors-draw-more-power-than-expected-tdp-turbo

[Lefteh]

 

@brob in your link it states that turbo boost technology 2.0 only has 2 states pl1 and pl2, but in the article you quoted " Intel has had this in place for a number of generations of processors, and most of it didn’t actually matter, as the power draw for the full chip was often well below the PL1 value even at full load. "The E5-1620 has base of 3.6 GHz with turbo 1/1/2/2 basically in non stressful aplications eg only 2 cores loaded the xeon is always at 3.8 GHz when the xeon is loaded to 3 or 4 cores it only has a 100MHz boost so its 3.7GHz. Now lets take for instance the i7-2600k as 1620 is really only a slighty higher cache (2MB) version with same max turbo. using this article we can see that the load on cinebench R11.5 measured through the ATX12V connector is 86W . Now lets look at how the 1620 performs with that same benchmark.

HMMMMMMM oddly accurate only few watts less power than whenaida64 is stressing which is mostly due to not being as fully stressful or maybe the cache is being stressed harder or perhaps its not reading the wattage past a certain point, who know i need to buy a current clamp to test my self. @Jurrunio can also test this if he remove all overclocking and sees what he should get as he has a 2600k. Its clear that overclocking messes with the power draw reading for CPUID as its not using normal states anymore.

8 hours ago, Jurrunio said:

How did you know that? Attaching a thermocouple to one of the low side mosfets while running Prime95 on the CPU?

Have you ever held your finger directly to a VRM? You can tell the difference between a decent number of temperatures its a built in functions humans have if you didn't know. If There was a VRM temperature measuring device embedded on the board most likely if it was over heating it would throttle the clocks, if there isn't then it could have the potential to get ridiculously hot and thus detectable by my finger. So either way its going to be detected by one or the other. Also if you want me too i will buy a thermal probe give me some good recommendations so i can test it more scientifically.

 

And if your wondering why i posted 3 posts in a row its the simple fact that when i edit a post you people don't even seem to read the edit like how @brob  responded because when i edit a minute after and then 2 minutes after they respond having the old edit not seeing the correction. @valdyrgramr  and @Norwegiantweaker i hope this clear this up for you.

 

The reason for all this is the fact that an extra heat sink vrm is not worth the money for a 2200g that is only doing office work. The reason that the B450 is being used is to squeeze little bit more performance out of the RAM as A320 is locked to 2666. Here a article about the power usage of the 2200g the write states that this is the power usage of the entire system.  Here is Toms hardware that has power for as well but it doesn't include the entire system  around 54W avg to 65.5W max. Seeing as the the original poster stated that the work that is going to be done is CPU heavy, no IGPU tasks. The ryzen 2200g voltage at load is 1.36v which would mean 48.16 Amps draw through the Vcore VRM. 4 phases 12.04‬ Amps per phase. Looking at a B450 3+3 phase from buildzoid we can see that a beefy 3 phase vrm at 75A outputs 10.3W. taking this over estimate (since the vrm become more inefficient the more current you push through it) of the power used at 75 and scaling it down to 48 Amps we get 6.6 Watts total heat on vrm. Now spread over 4 phases its 1.65W per phase which in honesty doesn't require a heatsink with some small amount of air flow(coming from cpu cooler).

 

[brob]

5 minutes ago, Lefteh said:

Have you ever held your finger directly to a VRM? You can tell the difference between a decent number of temperatures its a built in functions humans have if you didn't know. If There was a VRM temperature measuring device embedded on the board most likely if it was over heating it would throttle the clocks, if there isn't then it could have the potential to get ridiculously hot and thus detectable by my finger.

 

I would strongly advise against holding a finger on a VRM of a motherboard hosting a six+ core cpu. Keep in mind that 60C is going to burn you but not necessarily cause thermal throttling.

[Lefteh]

3 minutes ago, brob said:

 

I would strongly advise against holding a finger on a VRM of a motherboard hosting a six+ core cpu. Keep in mind that 60C is going to burn you but not necessarily cause thermal throttling.

its a 4 core if your wondering and from what i can tell i haven't burned my finger, it does feel hot but its not boiling water hot.

[Jurrunio]

6 hours ago, Lefteh said:

@Jurrunio can also test this if he remove all overclocking and sees what he should get as he has a 2600k. Its clear that overclocking messes with the power draw reading for CPUID as its not using normal states anymore.

I'm reading from intel XTU, if that's wrong, no software will be accurate. Btw it's 80w package power during gaming when multimeters read 150w, 135w package power in Prime95 smallFFT when multimeters read 200w. At stock clocks I'm getting 50w in gaming (100w multimeters), 80w in Prime95 (150w multimeters).

 

6 hours ago, Lefteh said:

Have you ever held your finger directly to a VRM?

my board has VRM heatsinks so I can't do that

 

6 hours ago, Lefteh said:

You can tell the difference between a decent number of temperatures its a built in functions humans have if you didn't know. If There was a VRM temperature measuring device embedded on the board most likely if it was over heating it would throttle the clocks, if there isn't then it could have the potential to get ridiculously hot and thus detectable by my finger. So either way its going to be detected by one or the other. Also if you want me too i will buy a thermal probe give me some good recommendations so i can test it more scientifically.

I do have a K-type and on my board's core VRM (put right beside a low side mosfet, squished down by the thermal pad under the VRM heatsink), it would go past 80C within 10 minutes of Prime95 and still climbing at that point. As far as I can tell there's no mosfet temperature read out so don't buy my board if you care about software mos temperature readings.

 

6 hours ago, Lefteh said:

when i edit a post you people don't even seem to read the edit

When you edit there's box you can tick to tell sb you edited your post. It's also your responsibility to show what you've edited

 

6 hours ago, Lefteh said:

The reason for all this is the fact that an extra heat sink vrm is not worth the money for a 2200g that is only doing office work. The reason that the B450 is being used is to squeeze little bit more performance out of the RAM as A320 is locked to 2666. Here a article about the power usage of the 2200g the write states that this is the power usage of the entire system.  Here is Toms hardware that has power for as well but it doesn't include the entire system  around 54W avg to 65.5W max. Seeing as the the original poster stated that the work that is going to be done is CPU heavy, no IGPU tasks. The ryzen 2200g voltage at load is 1.36v which would mean 48.16 Amps draw through the Vcore VRM. 4 phases 12.04‬ Amps per phase. Looking at a B450 3+3 phase from buildzoid we can see that a beefy 3 phase vrm at 75A outputs 10.3W. taking this over estimate (since the vrm become more inefficient the more current you push through it) of the power used at 75 and scaling it down to 48 Amps we get 6.6 Watts total heat on vrm. Now spread over 4 phases its 1.65W per phase which in honesty doesn't require a heatsink with some small amount of air flow(coming from cpu cooler).

HDV or Pro4 is an argument between you an @Herman Mcpootis, not me, I'm just here to clarify your flase information in power draw.

[Lefteh]

11 minutes ago, Jurrunio said:

I'm reading from intel XTU, if that's wrong, no software will be accurate. Btw it's 80w package power during gaming when multimeters read 150w, 135w package power in Prime95 smallFFT when multimeters read 200w. At stock clocks I'm getting 50w in gaming (100w multimeters), 80w in Prime95 (150w multimeters).

 

my board has VRM heatsinks so I can't do that

 

I do have a K-type and on my board's core VRM (put right beside a low side mosfet, squished down by the thermal pad under the VRM heatsink), it would go past 80C within 10 minutes of Prime95 and still climbing at that point. As far as I can tell there's no mosfet temperature read out so don't buy my board if you care about software mos temperature readings.

 

When you edit there's box you can tick to tell sb you edited your post. It's also your responsibility to show what you've edited

 

HDV or Pro4 is an argument between you an @Herman Mcpootis, not me, I'm just here to clarify your flase information in power draw.

Okay thank for the info i think im going to go with a MAX6675 arduino thermocouple to SPI and then just use serial to PC and log the data.

I did not know there was a box to tell people that the post was edited thanks for telling me.

 

Also any recommendations for current clamps? Ive always wanted one but i didn't want to spend $100 to $200 on one. you can just PM i don't think we need to send messages anymore on this post and me even making this post on this page is contradictory to what i just said.