Z9 performance and general discussion


#1

Hello all.

Just wanted to report my initial findings with the Z9 I received today.
The default hash board clock is 500MHz. Every 10% increase in clock resulted in a 10% increase in average hashrate and a 10% increase in DC power consumption on the hashboards. Each one is around 6A at 500MHz and 9A at 750MHz. The controller runs around 0.5A but varies a bit.

I noticed other Z9s on luckpool. Some are hashing a fair bit faster than mine. I wonder if they have exceeded the clocks in the drop down box to achieve it.
This one gets over 17k consistently
https://luckpool.net/zen/miner.html?zniW5xFyjvaToSh3RasBmUsf2KyqoMVcALm


#2

Then again, it could just be a difference in the pool reporting and mine hasn’t been online long enough.


#3

There are 2 options in my opinion:

1.) It’s just a better unit some get. Mine is right now at 743 Mh/z and mines at 16.45 kh/s so at 750 Mh/z it would get as well close to 17 kh/s…

2.) The hashrate of that miner isn’t displayed correctly on the pool and isn’t uncommon at all.


#4

Adding some additional info as i test several pools on first day:

Temperature at overclock to 743 Mh/z and 16.45 kh/s on my end: 48C at night, 55C at day with fan on 100%

Test mining pools so far:
Suprnova: ZEC, BTCP, ZEN all work perfect, just some payout display problems as it seems
Miningpoolhub: ZEC, seem to work perfectly


#5

Luckpool on zen as well. And flypool for Zec


#6

Vosk had his up to 16k at 750M so some may be that high. My 1st is just over 11k stock. I imagine we should OC these to the limit and mine as much as we can before the difficulty spikes?


#7

silicon lottery

750MHz on one gave me 17 and on another it only gave me 15.5

Just remember to monitor those temps in the day as summer approaches.


#8

What is the safe upper zone for temps? Is it 70C like GPU’s?


#9

I would assume so. They are a similar process


#10

Pretty sure even 80C wouldn’t matter much. My S9’s run without problems at 90C for months now …
But i doubt the Z9 mini will get into these ranges higher than 60 - 65C.


#11

I have no idea how different these are to an s9. Those cutoff at like 110°. I think I remember people running s9’s in mid 80°s and being just fine. I might gamble when its winter, right now I’ll stick to 70’s


#12

I’am not a fan of big pools so i will pass on flypool and antpool. I suggest all others to do the same so the mining pool distribution is more fair without having 1 or 2 big pools. Just a thought when choosing a miningpool to mine on…


#13

Mine is only hitting 70C on 1 Chip and highest PCB is 55C at 750M. It is doing about 16.5k, so not too bad. Fan steady at 4000.

Which stat is hardware errors? I’m getting discarded shares, but even get those at default. It must be HW?


#14

poolside usually with something this new. try on suprnova. not getting rejected shares there.


#15

About the hardware errors:

=> Go to the Miner Status Tab
=> go down from Summary, Pools to Antminer
=> here you see some different stats, chain, temp, etc.
=> about in the middle/center you see a column HW, these are the Hardware errors

only some hardware erros don’t matter, often it’s just something with the fans. Just in case some day, when they get too much, restart your Asic and eventually lower the frequency 1 step, but it seems this will not be of any need with the Z9 mini as it looks more than perfectly stable on every frequency.

And no, discarded shares are not HW’s (hardware failures). They are just discarded, test with some other pools, i suggest as well suprnova at this point. But eventually it’s as well to where you are located as a european pool for an european might work better for them as a chinese pool.


#16

Just noticed that when i try to configure the api access to the Z9 the Miner Status Tab is gone completly. The miner still works perfectly but nothing more on the Miner Status menu is visible. But on the api application all stats are visible perfectly.

Just in case someone else is running them with an .api application.


#17

Are you resolve your problem ? Changes since Overwinter with Z9 ?


#18

i didn’t solve the .api problem yet, it’s an internal one not related to mining at all, just an internal display problem.

No problem with Z9 and Zcash mining. Mining it on miningpoolhub without any problems the last 12 hours.


#19

Ok thanks Alex, i ask again after 12 hours :wink:


#20

FYI:

The main core voltage converter is:


with integrated switching fets and drivers:

The switching clock is internally generated and configured via pmbus. From what I see this is actually a pretty well done VRM layout. Obviously can’t determine much about ripple or gain and phase margin but hey… it looks nice.

I haven’t pulled my hash board out yet to take my own photos, but from some posted on bitcointalk it appears that these may have a single reference clock from a crystal oscillator. It might be possible to change the crystal or loading capacitance to push the frequency up a bit and get more overclocking range.