fenderman
Staff member
- Mar 9, 2014
- 36,892
- 21,407
I have machines that are now about 8 years old.
It depends on your rates. Modern pc's are VERY efficient, drawing about 40-50w under a typical blue iris load.
4-8 cameras is a meaningless number. What matters is the framerate x resolution.
Thanks for the quick reply!
Wow, 8 years! That's great.
My bad. Can you give an estimate for 5MP cameras at 15 fps?
5 MP * 15 FPS * 8 cameras = 600 MP/sec. For this I would recommend an i7-3770 or newer (3rd-gen desktop i7) or i5-8400 (8th-gen 6-core i5) or newer.
This looks like a good one: Dell OptiPlex 9020 | 3.40GHz Core i7 4770 | 8gb DDR3 | 320gb | DVD-RW | WIn 10 | eBay
I can't keep track of which drive bays are available in which dell/hp/etc cases, but for one of that size I'd expect at least two 3.5 inch bays, one already filled for the included 320 GB HDD. The other bay is probably empty with no plastic caddy to mount the drive in.
Also in this specific example it comes with a graphics card which is probably worthless to you; unless you need it for a video output, you could take it out and use just onboard graphics and save a little power.
This looks like a good one: Dell OptiPlex 9020 | 3.40GHz Core i7 4770 | 8gb DDR3 | 320gb | DVD-RW | WIn 10 | eBay
I can't keep track of which drive bays are available in which dell/hp/etc cases, but for one of that size I'd expect at least two 3.5 inch bays, one already filled for the included 320 GB HDD. The other bay is probably empty with no plastic caddy to mount the drive in.
Also in this specific example it comes with a graphics card which is probably worthless to you; unless you need it for a video output, you could take it out and use just onboard graphics and save a little power.
Last edited:
As an eBay Associate IPCamTalk earns from qualifying purchases.
I'm not aware of any 5MP cameras worth buying at this time by the way. You can typically get better price/performance from 2MP or (new) 4MP starlight cams or 8MP cams with a large 1/1.8" or 1/2" sensor.
Thank you for such a detailed response! For this setup, as @fenderman mentioned, I can assume the pc would draw close to 40-50w?
As an eBay Associate IPCamTalk earns from qualifying purchases.
if you can get that setup, but as long as you stay modern (4th, 5th gen or better) and stay away from old business workstations, and avoid running a power hungry graphics card until absolutely necessary.
As an estimate,
- 50w * 24h per day = 1.2kwh per day
- 75w * 24h per day = 1.8kwh per day
This thread just made me want to put my
Kil-a-watt meter on mine. Just the pc and switch are running around 62-65 watts together, and the 21” monitor is only adding 12-13 watts with blue iris on the screen. I would think the number of cameras would also matter since the load of the switch will largely depend on how many cameras it has to power. Each one of my cams add 5-6 watts when plugged in. No significant difference when they go in or mode.
Anyway mine currently only has 3 5231 cams on it and is running on an optiplex with i7 6700 and a nvida quadro graphics card.
Sent from my iPhone using Tapatalk
Kil-a-watt meter on mine. Just the pc and switch are running around 62-65 watts together, and the 21” monitor is only adding 12-13 watts with blue iris on the screen. I would think the number of cameras would also matter since the load of the switch will largely depend on how many cameras it has to power. Each one of my cams add 5-6 watts when plugged in. No significant difference when they go in or mode.
Anyway mine currently only has 3 5231 cams on it and is running on an optiplex with i7 6700 and a nvida quadro graphics card.
Sent from my iPhone using Tapatalk
looney2ns
IPCT Contributor
J Sigmo
Known around here
- Feb 5, 2018
- 996
- 1,336
In my experience, what fails first in a PC is never the CPU itself. In fact, I've never seen a CPU failure.
What inevitably bites the dust are some electrolytic capacitors, used in a switching power conversion circuit on the motherboard or in the power supply.
The failure almost always then manifests itself with the system running fine until you power it down and then try to start it up again. At that point, the system fails to restart.
Where I used to work, I had to repair a dozen or so motherboards because we could no longer find the mobos with ISA bus slots that were required to operate data acquisition boards necessary for a number of data logging and control systems.
That was tedious. It honestly took about five hours per mobo to do a complete removal and replacement of every single electrolytic cap on each board, and I was beat from hunching over the old workbench for so many hours when doing them.
But from that, I still have a cache of various values of fancy low ESR, high temp, high ripple-current rated caps in my parts inventory. This has come in handy on a number of occasions to repair all kinds of power supplies and boards where I work now.
Our smallish town lost its only electronic parts store a few years back due to the downturn in coal, oil, and uranium prices. So if I didn't have the parts in my personal inventory, we'd be without various instruments and such for days any time we had a failure because everything has to be ordered in now.
Anyhow. Count on an overworked, over heated, (and under specced) electrolytic capacitor, or a bank of them, being the most likely cause of failure in a modern PC.
You can often spot the bad cap(s) because they frequently bulge or even vent once they completely fail. They may also heat enough to cause their heat-shrinkable plastic wrapper to shrink excessively. Or they can just lose enough capacity that they cannot do their job, but not show any outward signs.
What inevitably bites the dust are some electrolytic capacitors, used in a switching power conversion circuit on the motherboard or in the power supply.
The failure almost always then manifests itself with the system running fine until you power it down and then try to start it up again. At that point, the system fails to restart.
Where I used to work, I had to repair a dozen or so motherboards because we could no longer find the mobos with ISA bus slots that were required to operate data acquisition boards necessary for a number of data logging and control systems.
That was tedious. It honestly took about five hours per mobo to do a complete removal and replacement of every single electrolytic cap on each board, and I was beat from hunching over the old workbench for so many hours when doing them.
But from that, I still have a cache of various values of fancy low ESR, high temp, high ripple-current rated caps in my parts inventory. This has come in handy on a number of occasions to repair all kinds of power supplies and boards where I work now.
Our smallish town lost its only electronic parts store a few years back due to the downturn in coal, oil, and uranium prices. So if I didn't have the parts in my personal inventory, we'd be without various instruments and such for days any time we had a failure because everything has to be ordered in now.
Anyhow. Count on an overworked, over heated, (and under specced) electrolytic capacitor, or a bank of them, being the most likely cause of failure in a modern PC.
You can often spot the bad cap(s) because they frequently bulge or even vent once they completely fail. They may also heat enough to cause their heat-shrinkable plastic wrapper to shrink excessively. Or they can just lose enough capacity that they cannot do their job, but not show any outward signs.
I totally understand what you are saying. It was already in the pc when I got it and it’s not a big power hog. With my systems current draw per day the entire system with cams should cost me $50-$60 per year in electricity. That card probably consumes less then a dollar per month. And my core temps have stayed under 50. I’m sure I’ll take it out but right now I’m just seeing how everything runs as is and getting used to it.
Sent from my iPhone using Tapatalk
I too have rarely seen a CPU fail. Please remember to keep in mind that higher temps have a lot to do with hardware failure.
I have my servers in an AG building with temps that vary anywhere from 36F-110F, therefore I must set up my rigs with special considerations, such as lots of fans and water cooling.***
***Hopefully this satisfies LORD FenderBender's criteria for accurate posts!
I have an Intel i5-3570k with an ASRock Z77 Extreme 4 that has been on nearly 24/7 for over 6 years and it is still running cool as a cucumber. I did use water cooling for the CPU and I think I have over 17 fans on this rig running 24/7 [see pic].
I also am running an i7-4770k water cooled for my BI server. Believe it or not, both the motherboard and the processor died! Intel graciously gave me a brand new CPU [it was still under their warranty].
I think the motherboard was a mini. I switched it out to a AsRock Z97 Killer [found a great deal on a number of these MB's]!
All is well again!
I go out of my way to keep massive air flow and use only the best PSU [Seasonic].
Hope this helps!
Soar
I have my servers in an AG building with temps that vary anywhere from 36F-110F, therefore I must set up my rigs with special considerations, such as lots of fans and water cooling.***
***Hopefully this satisfies LORD FenderBender's criteria for accurate posts!
I have an Intel i5-3570k with an ASRock Z77 Extreme 4 that has been on nearly 24/7 for over 6 years and it is still running cool as a cucumber. I did use water cooling for the CPU and I think I have over 17 fans on this rig running 24/7 [see pic].
I also am running an i7-4770k water cooled for my BI server. Believe it or not, both the motherboard and the processor died! Intel graciously gave me a brand new CPU [it was still under their warranty].
I think the motherboard was a mini. I switched it out to a AsRock Z97 Killer [found a great deal on a number of these MB's]!
All is well again!
I go out of my way to keep massive air flow and use only the best PSU [Seasonic].
Hope this helps!
Soar
Last edited:
fenderman
Staff member
- Mar 9, 2014
- 36,892
- 21,407
There is no need to go overboard with water cooling or extra fans venting. Any decent system will be properly designed by the manufacture. I have over 20 BI systems most are 8-4 years old running 24/7 no issues.
FenderBenderman,
Unfortunately, some of our rigs are inside steel buildings with no A/C. Ag buildings to be specific. So ambient temps today were 104F+. In this scenario, water cooling + lots more fans do indeed help!
J Sigmo
Known around here
- Feb 5, 2018
- 996
- 1,336
When we put PCs in the Soils Prep labs at some of our laboratories, we built enclosures for the PCs that filtered the air through HEPA filters and included small window AC units. The setups provided positive pressure to keep the fine rock dust out, and the AC units kept the ambient that the PCs experienced at normal room temp.
This may or may not be practical in your applications, but it worked great for us. By keeping the entire PC cool, it helped with the capacitor life as well as the actual CPU chips.
Newer PCs use a lot less power than ones from even 5 to 7 years ago. That helps enormously, of course. But manufacturers are now also very concerned with keeping their systems quiet, as well. And sometimes they do sacrifice air flow for noise level. So there's good and bad with the newer PCs.
Between the labs in our company, we had, perhaps, 1000 PCs, and they were all left running 24/7 so the IT folks could do backups and software deployments at night.
Capacitors. Friggin capacitors!
This may or may not be practical in your applications, but it worked great for us. By keeping the entire PC cool, it helped with the capacitor life as well as the actual CPU chips.
Newer PCs use a lot less power than ones from even 5 to 7 years ago. That helps enormously, of course. But manufacturers are now also very concerned with keeping their systems quiet, as well. And sometimes they do sacrifice air flow for noise level. So there's good and bad with the newer PCs.
Between the labs in our company, we had, perhaps, 1000 PCs, and they were all left running 24/7 so the IT folks could do backups and software deployments at night.
Capacitors. Friggin capacitors!
fenderman
Staff member
- Mar 9, 2014
- 36,892
- 21,407
That is an omitted fact. The OP is not placing the pc in such heat. Your post implied that this is something everyone should consider, its not. Its a foolish thing to do.
fenderman
Staff member
- Mar 9, 2014
- 36,892
- 21,407
modern pc's have no such issues in standard office/home space. Aside from the blue iris pc's that run a load 24/7 I have loads of office pc's that run 24/7 for remote access. There has never been an issue with failure. Never had a mobo or cpu failure and one power supply over many years and at least 100 pc. The fans will increase speed when temps rise.When we put PCs in the Soils Prep labs at some of our laboratories, we built enclosures for the PCs that filtered the air through HEPA filters and included small window AC units. The setups provided positive pressure to keep the fine rock dust out, and the AC units kept the ambient that the PCs experienced at normal room temp.
This may or may not be practical in your applications, but it worked great for us. By keeping the entire PC cool, it helped with the capacitor life as well as the actual CPU chips.
Newer PCs use a lot less power than ones from even 5 to 7 years ago. That helps enormously, of course. But manufacturers are now also very concerned with keeping their systems quiet, as well. And sometimes they do sacrifice air flow for noise level. So there's good and bad with the newer PCs.
Between the labs in our company, we had, perhaps, 1000 PCs, and they were all left running 24/7 so the IT folks could do backups and software deployments at night.
Capacitors. Friggin capacitors!
Just plug the pc in and forget about it.
For what it's worth, I believe you will find water cooling to be less reliable, due to the number of failure opportunities.
For an "old fashion" heatsink cooler with Fan:
All that being said I have both a 10+ year old heatsink cooler (Thermaltake Big Typhoon that is still running with the original fan) that I just blow crap out the fan and heatsink every year. It is working fine.
I removed a 6+ year old AIO water cooling loop from my 2600k and it seemed to still be working fine, but it definitely sounds like its half empty so idk why it wasn't causing problems beforehand.
If you stick with an old fashioned cooler, and you clean it semi-regularly, if it ever fails most modern motherboards will shutdown before serious CPU damage occurs, buy a new fan to replace the bad and you are back in business. I think you will find "old fashioned" to be more reliable AND cheaper. I always worried about the size and weight (nearly two pounds) of the high performance heatsink coolers, but I've had that one hanging off an i7-920 (since it is mounted in a tower gravity is constantly applying pressure to one side of the chip) since we bought it and it hasn't broken the chip yet.
For an "old fashion" heatsink cooler with Fan:
- CPU cooler (fan plus heatsink) - metal heatsinks don't fail they are just a hunk of metal, but they can become clogged with dust and fuzz from environment and become less and less effective.
- CPU paste can deteriorate over time
- the fan that provides cooling will eventually fail, faster in a high dust environment, but bearings wear out ( a quality ball bearing fan will fail slower than a sleeve bearing, but faster than a fluid bearing). Fans also accumulate crap on their leading edges reducing effectiveness.
- Cold plate mounts to CPU with CPU paste (same failure chance as above)
- Cold plate most likely has the pump integrated, but even a standalone pump adds an additional failure point more than above because the little motor that drives the impeller that pushes the fluid isn't required above.
- the radiator itself is just a hunk of metal, probably fine, but has same problem as heatsink in accumulation of dust and crap reducing effectiveness
- the fan that pushes air through the radiator will have the same risk of failure as a traditional fan, same types of problems affect all fans, quality fans last longer, dust free environments help.
- the tubing and liquid that transfers the heat from the block to the radiator will leech out of even the best tubing, eventually your water cooling loop will not have enough fluid to work effectively. In a custom water loop, people have to add liquid over time to address this loss.
- the liquid itself can become less effective over time, due to biological or debris accumulation over long periods of time.
All that being said I have both a 10+ year old heatsink cooler (Thermaltake Big Typhoon that is still running with the original fan) that I just blow crap out the fan and heatsink every year. It is working fine.
I removed a 6+ year old AIO water cooling loop from my 2600k and it seemed to still be working fine, but it definitely sounds like its half empty so idk why it wasn't causing problems beforehand.
If you stick with an old fashioned cooler, and you clean it semi-regularly, if it ever fails most modern motherboards will shutdown before serious CPU damage occurs, buy a new fan to replace the bad and you are back in business. I think you will find "old fashioned" to be more reliable AND cheaper. I always worried about the size and weight (nearly two pounds) of the high performance heatsink coolers, but I've had that one hanging off an i7-920 (since it is mounted in a tower gravity is constantly applying pressure to one side of the chip) since we bought it and it hasn't broken the chip yet.
Last edited:
Yeah I'm all for air cooling. Liquid cooling adds cost, complexity, more points of failure, and isn't really more effective in any significant way.