Antminer S5 vs Spondoolies SP20

[philipma1957]

S-5's are back on sale at a price jump


https://www.bitmaintech.com/


ship on march 30th

431 with ups shipping included to the usa.


or june 9th

401  with ups shipping  included to the usa

[thedreamer]

From what I read, the s5 is having issues and seems quite loud where most people modify or change the fan.
The SP20s can get rather quiet if you get them to sub 700watt and around 1200GH. No mods necessary.

This is an accurate assessment. An SP20 running @ ~1250GHs @ 20% fans is quieter than an S5 with the stock fans.

To your point of "having issues", I'm not sure what specifically you are referring to. The two B1 S5's I purchased have been running very well, performance wise.

I've heard and saw some reports on high error rates at stock speeds, needs to reset multiple times due to dropped connections and the mounting issues due to the top open.

Still thinking of getting one or two to play with though. Or the C1 maybe, but I hate how they made the watercooling an external 'necessity'.

[MegaFall]

Don't mean to revive this thread but have a similar question. Considering the SP20 or the S5. If I wanted to max wattage consumption at 450-550 watts which would get me the most speed at an underclock? I'd like to keep temps under 46c and the noise at about the same of a stock s3+.

Also, can either of these be load balanced to divide up the hashing power and mine on 2-3 pools at the same time?

[Biodom]

Don't mean to revive this thread but have a similar question.
Considering the SP20 or the S5. If I wanted to max wattage consumption at 450-550 watts which would get me the most speed at an underclock?
I don't know if 450W is even possible on SP20. 550W is possible with a good PSU on both Sp20 and S5 with similar speed

I'd like to keep temps under 46c and the noise at about the same of a stock s3+.
It does not work the same way as on S3 as far as temp are concerned. SP20 shows exhaust temp (less than 70-75 is OK) and actual chip temp, for which 85-115C is OK; for S5 the sensor shows some temp, which is roughly at least 20-25C less than actual chip temp-with temp less than 65C being OK, they don't need to be 46C

Also, can either of these be load balanced to divide up the hashing power and mine on 2-3 pools at the same time?
not sure, but here is one link:
https://bitcointalk.org/index.php?topic=966799.0

see ^^^^

[MegaFall]

Don't mean to revive this thread but have a similar question.
Considering the SP20 or the S5. If I wanted to max wattage consumption at 450-550 watts which would get me the most speed at an underclock?
I don't know if 450W is even possible on SP20. 550W is possible with a good PSU on both Sp20 and S5 with similar speed

I'd like to keep temps under 46c and the noise at about the same of a stock s3+.
It does not work the same way as on S3 as far as temp are concerned. SP20 shows exhaust temp (less than 70-75 is OK) and actual chip temp, for which 85-115C is OK; for S5 the sensor shows some temp, which is roughly at least 20-25C less than actual chip temp-with temp less than 65C being OK, they don't need to be 46C

Also, can either of these be load balanced to divide up the hashing power and mine on 2-3 pools at the same time?
not sure, but here is one link:
https://bitcointalk.org/index.php?topic=966799.0

see ^^^^

Well I'm concerned about temp because the 45c that my S3+ runs at heats my bedroom to the point where I need to keep my door opened and close off the heating vent to the room. So if both run significantly hotter then both are a no go.

[Biodom]

Don't mean to revive this thread but have a similar question.
Considering the SP20 or the S5. If I wanted to max wattage consumption at 450-550 watts which would get me the most speed at an underclock?
I don't know if 450W is even possible on SP20. 550W is possible with a good PSU on both Sp20 and S5 with similar speed

I'd like to keep temps under 46c and the noise at about the same of a stock s3+.
It does not work the same way as on S3 as far as temp are concerned. SP20 shows exhaust temp (less than 70-75 is OK) and actual chip temp, for which 85-115C is OK; for S5 the sensor shows some temp, which is roughly at least 20-25C less than actual chip temp-with temp less than 65C being OK, they don't need to be 46C

Also, can either of these be load balanced to divide up the hashing power and mine on 2-3 pools at the same time?
not sure, but here is one link:
https://bitcointalk.org/index.php?topic=966799.0

see ^^^^

Well I'm concerned about temp because the 45c that my S3+ runs at heats my bedroom to the point where I need to keep my door opened and close off the heating vent to the room. So if both run significantly hotter then both are a no go.

you simply can't have either of these machines (unmodified) in the bedroom as they are too loud for this.
You can mod S5 with a couple of low noise fans (there are gazillion of posts about this) PLUS use it at 300 mhz-should be ok for ~1000GH.

[MrMissile]

Sorry to rehash the thread, but I just bought a EVGA 1200w P2 power supply I found on sale yesterday to couple up on an sp20 or some S5's.  Debating between the sp20 and the S5, it looks as though I might be better running 2 S5s and get better power use per tg/s than combining an S5 and an sp20.  The S5's are easy to find at about $450, but can hardly locate any sp20's for sale and the price is significantly higher than the S5's.  I'd like to stick in the 2 tg/s range since I'm a noob and see how this goes.  Any thoughts about me thinking 2 S5's might be "friendlier" than an sp20 or an sp20 coupled with an s5?

Heat is one thing to worry about since I'm in Texas, that's why I favor the S5 over the sp20 here because lower watts.  

[thedreamer]

Sorry to rehash the thread, but I just bought a EVGA 1200w P2 power supply I found on sale yesterday to couple up on an sp20 or some S5's.  Debating between the sp20 and the S5, it looks as though I might be better running 2 S5s and get better power use per tg/s than combining an S5 and an sp20.  The S5's are easy to find at about $450, but can hardly locate any sp20's for sale and the price is significantly higher than the S5's.  I'd like to stick in the 2 tg/s range since I'm a noob and see how this goes.  Any thoughts about me thinking 2 S5's might be "friendlier" than an sp20 or an sp20 coupled with an s5?

Heat is one thing to worry about since I'm in Texas, that's why I favor the S5 over the sp20 here because lower watts.  

They are about the same at the same speeds. The S5 is 'maybe' 30-40watts more efficient. At that speed you can run 2x of either on that power supply.

But the S5 is WAY louder than the SP20's.

[notlist3d]

Sorry to rehash the thread, but I just bought a EVGA 1200w P2 power supply I found on sale yesterday to couple up on an sp20 or some S5's.  Debating between the sp20 and the S5, it looks as though I might be better running 2 S5s and get better power use per tg/s than combining an S5 and an sp20.  The S5's are easy to find at about $450, but can hardly locate any sp20's for sale and the price is significantly higher than the S5's.  I'd like to stick in the 2 tg/s range since I'm a noob and see how this goes.  Any thoughts about me thinking 2 S5's might be "friendlier" than an sp20 or an sp20 coupled with an s5?

Heat is one thing to worry about since I'm in Texas, that's why I favor the S5 over the sp20 here because lower watts.  

They are about the same at the same speeds. The S5 is 'maybe' 30-40watts more efficient. At that speed you can run 2x of either on that power supply.

But the S5 is WAY louder than the SP20's.

If noise is a big issue you can modify the fan's.   If you use 2 doing push/pull you can use quieter fans as it's splitting the load.

S5 has some advantages at this point.  It can still be bought new, and have a warranty.   

[thedreamer]

Sorry to rehash the thread, but I just bought a EVGA 1200w P2 power supply I found on sale yesterday to couple up on an sp20 or some S5's.  Debating between the sp20 and the S5, it looks as though I might be better running 2 S5s and get better power use per tg/s than combining an S5 and an sp20.  The S5's are easy to find at about $450, but can hardly locate any sp20's for sale and the price is significantly higher than the S5's.  I'd like to stick in the 2 tg/s range since I'm a noob and see how this goes.  Any thoughts about me thinking 2 S5's might be "friendlier" than an sp20 or an sp20 coupled with an s5?

Heat is one thing to worry about since I'm in Texas, that's why I favor the S5 over the sp20 here because lower watts.  

They are about the same at the same speeds. The S5 is 'maybe' 30-40watts more efficient. At that speed you can run 2x of either on that power supply.

But the S5 is WAY louder than the SP20's.

If noise is a big issue you can modify the fan's.   If you use 2 doing push/pull you can use quieter fans as it's splitting the load.

S5 has some advantages at this point.  It can still be bought new, and have a warranty.  

Still loud as HELL.
An SP20 at same speed, cut Fan to 10% and is MUCH quieter than the push pull setup with deltas and other fans on the S5, otherwise they are not sufficient enough to properly cool the S5.

[MrMissile]

Noise will be an issue as well as heat.  For sure I will change the stock fans with deltas and probably only run one S5 during the hottest summer months...but from my understanding of what you're saying the S5's are much louder than the sp20's?  I guess my prorities are pretty much in order of: wattage per tg/s, heat and then noise.  

From what I've researched, the sp20 pulls about 1.3-1.4 tg/s at 1100w, and an S5 pulls about 1 tg/s at 600w.  Is that an accurate comparison?

Thanks for all your help, info and posts.  I'm new to all this stuff but eargerly gathering my sources to make this a little hobby for me.  I'm new to this social media stuff too, I like to fly under the radar without having facebook etc so if I make a "social mistake" please excuse me...

[Finksy]

I would be more concerned with the impact of heat with the SP's than the S5's.  I found the SP20's with auto settings (or even at 10%) and underclocked, the chips would get much hotter than I was comfortable with (>100*C).  If you lose 1 chip on the SP20 due to heat stress, you're down 25% of your hashrate.  If you lose 1 string of chips on an S5, you'll lose considerably less.  I've never heard S5's with aftermarket fans, but the SP20 even at lower fan settings was still very annoying sounding.

Efficiency between the two should be comparable for your intended use & hashrate, so go with whichever one you can get for cheaper (likely S5), and put the money into a push-pull setup to quiet it down, there are many people who have done this on the forum.

I don't have efficiency numbers first-hand with the S5. But the SP20 can get ~1450 GH/s on 1100 W, and efficiency goes up quite linearly as hashrate goes down.

[Biodom]

Noise will be an issue as well as heat.  For sure I will change the stock fans with deltas and probably only run one S5 during the hottest summer months...but from my understanding of what you're saying the S5's are much louder than the sp20's?  I guess my prorities are pretty much in order of: wattage per tg/s, heat and then noise.  

From what I've researched, the sp20 pulls about 1.3-1.4 tg/s at 1100w, and an S5 pulls about 1 tg/s at 600w.  Is that an accurate comparison?

My SP20E hashes at 1230-1235 GH at 680-683w at the wall
S5 (sold all after hashing for a while with them) were ~1150 at 620-630w at the wall or ~1180-1200 overclocked at ~650w at the wall
However, SP20E produces 1230-1235 GH at fan setting 10, which is quiet, while S5 default is SUPER loud and modded with two quieter fans (either deltas or Noctua/Scythe at the cost of ~$40) is still a little louder than SP20 at fan 10.

Basically a draw with slight edge to SP20 due to interesting interface that is fun to play with.
S5 is slightly (~5%) more energy efficient at low speeds for SP20, but Sp20 is slightly more productive.

Re chip temp-we have no idea what chip temp S5 has since they don't measure temp at the chip, but rather at the board.
SP20 chips are OK at up to 115C.

[MrMissile]

I would be more concerned with the impact of heat with the SP's than the S5's.  I found the SP20's with auto settings (or even at 10%) and underclocked, the chips would get much hotter than I was comfortable with (>100*C).  If you lose 1 chip on the SP20 due to heat stress, you're down 25% of your hashrate.  If you lose 1 string of chips on an S5, you'll lose considerably less.  I've never heard S5's with aftermarket fans, but the SP20 even at lower fan settings was still very annoying sounding.

Efficiency between the two should be comparable for your intended use & hashrate, so go with whichever one you can get for cheaper (likely S5), and put the money into a push-pull setup to quiet it down, there are many people who have done this on the forum.

I don't have efficiency numbers first-hand with the S5. But the SP20 can get ~1450 GH/s on 1100 W, and efficiency goes up quite linearly as hashrate goes down.

That's the setup I was hoping for, Finksy.  Sounds like a good small  efficient setup to start out.  Thanks for all your help!

[philipma1957]

if you underclock sp20s to 900gh they are more efficient then s-5's at 1150gh  about .47 watts vs .51 or .52

you can run 2 easy at under 900 watts.  and 8% fan speed.  sound is meh, but not good.

s-5's can be fan modded to be okay, but never good.

pretty much a push if you can get quality working sp20's

[thedreamer]

Noise will be an issue as well as heat.  For sure I will change the stock fans with deltas and probably only run one S5 during the hottest summer months...but from my understanding of what you're saying the S5's are much louder than the sp20's?  I guess my prorities are pretty much in order of: wattage per tg/s, heat and then noise.  

From what I've researched, the sp20 pulls about 1.3-1.4 tg/s at 1100w, and an S5 pulls about 1 tg/s at 600w.  Is that an accurate comparison?

My SP20E hashes at 1230-1235 GH at 680-683w at the wall
S5 (sold all after hashing for a while with them) were ~1150 at 620-630w at the wall or ~1180-1200 overclocked at ~650w at the wall
However, SP20E produces 1230-1235 GH at fan setting 10, which is very quiet, while S5 default is SUPER loud and modded with two quieter fans (either deltas or Noctua/Scythe at the cost of ~$40) is still a little louder than SP20 at fan 10.
basically a draw with slight edge to SP20 to interesting interface that is fun to play with.
S5 is slightly (~5%) more energy efficient at low speeds for SP20, but Sp20 is slightly more productive.

Exactly what I get. I have (had) 20x SP20's and 2 S5's (for testing).

The S5's on my killawatt meter use 630Watts at stock speeds and temps in the 60C range, so fan at FULL. LOuD AS HELL and annoying, even in my garage. AMbient temps are 70F and below.
IMO the open design of the S5 is just BAD unless you're in a temp controlled data center.

The SP20's I run at 490-520Watts and get between 980-1060GH/s. With fan at 5% in my office, I used to run two of them at 70F Ambient. Temps on the miners never got above 70C, and the DIES never above the 85C minimum shown.

When I turned them up to 1150GH/s @ 600watts with the fan at 10% the temps still did not go past 77C and although I had to turn the TV volume higher in my office, I could still work.
Now I run them at 550/1.1K in the basement at 30% fan speed and no worries.  

[MrMissile]

So maybe I ought to really look into buying the sp20's instead of the s5's?  Sounds like they are quieter w/o need for any modifications.  I'm sure the initial cost is a little higher, but I thought an sp20 needed 1150 watts to run per the manufactures' website?  Now I'm confused...is there a low speed on the sp20 or do you guys have them tweeked to run at lower watts?

[thedreamer]

So maybe I ought to really look into buying the sp20's instead of the s5's?  Sounds like they are quieter w/o need for any modifications.  I'm sure the initial cost is a little higher, but I thought an sp20 needed 1150 watts to run per the manufactures' website?  Now I'm confused...is there a low speed on the sp20 or do you guys have them tweeked to run at lower watts?

I have to say, the interface and design IMO on the SP20's is the nicest I've seen on ANY miner. PLug and Play, under and overclock settings easily changed, fan everything.
If you want to adjust the Fan to below 10%, there are 2 ways to do it, either vie the webgui or SSH into the unit. I can even help with that if you want later on.

[Biodom]

So maybe I ought to really look into buying the sp20's instead of the s5's?  Sounds like they are quieter w/o need for any modifications.  I'm sure the initial cost is a little higher, but I thought an sp20 needed 1150 watts to run per the manufactures' website?  Now I'm confused...is there a low speed on the sp20 or do you guys have them tweeked to run at lower watts?

you tweak them through interface by adjusting several parameters, such as power, starting voltage and maximum voltage for each loop (they have four) independently.
Check phil's thread (unofficial SP20) for guidance. there are hundreds of setups documented there. each machine is slightly different, so optimal values might differ in your case vs what is posted. the consensus is pretty much that one can use either fan 10 or auto fan to achieve very efficient mining with proper voltage settings.

I say that for a large scale operation S5 might be better, but for a small home based, i personally would go with sp20 (at a reasonable price) just due to so many ways you could change things and see what happens. S5 is also good, apart from that notorious bug (only in some machines) when they overheat when disconnected from the internet. Loud noise can be dealt with by fan modding, but default S5 in the same room (office or bedroom) is basically intolerable. In all fairness, at default settings (1150w/ 1650GH), Sp20 is loud as well.

[alh]

One other thing to remember: Watts = Heat

Assuming you put two 500W miners in the same room, it will heat up the room the same as a single 1000W miner. They all have their wrinkles in terms of chip temperatures, fans, air flow and such. However, the total heat production is based on total watts consumed.

Just because I am a worry wart, I would think two S5's on a single 1200W power supply is "on the edge" in terms of power draw. Each S5 is "about" 600W at stock speeds, so two is the limit of your power supply. While I expect that the EVGA P2 is a very well built PSU, You should probably consider down-clocking the tso S5's if you put them on the same PSU.

It's a real shame that there are only used SP20's. They are great box, and very flexible in terms what you can choose (i.e. Hot/Fast/Loud or Cool/Slow/Quieter).