How to solve airflow issue?

[Ers900]

I have a 230v supply - 100amp panel (UK) - each miner has its own 20amp breaker/fuse with a 16amp commando socket c19 power cable.

Miner intake temps are between 55-58, outlet temps are 76-79.

I have 4x L9 (17gh) I managed to get 3 of them working,

When the fourth one is turned one it’s like a random pick out of the 4 as to which one will overheat and go offline.
Exhaust ducts are pulled outside with an extraction fan.
I believe there is not enough airflow coming in into the miner room in the basement.

Which fan specs would be satisfactory enough to feed 4x L9’s airflow?
For example would a 10” fan (1130m3h) blowing air into the miner room do the job?

Thanks

[FP91G]

I have a 230v supply - 100amp panel (UK) - each miner has its own 20amp breaker/fuse with a 16amp commando socket c19 power cable.

Miner intake temps are between 55-58, outlet temps are 76-79.

I have 4x L9 (17gh) I managed to get 3 of them working,

When the fourth one is turned one it’s like a random pick out of the 4 as to which one will overheat and go offline.
Exhaust ducts are pulled outside with an extraction fan.
I believe there is not enough airflow coming in into the miner room in the basement.

Which fan specs would be satisfactory enough to feed 4x L9’s airflow?
For example would a 10” fan (1130m3h) blowing air into the miner room do the job?

Thanks

Can you post a photo of your mining equipment and ventilation?
Your temperature is excellent, my friend has these ASICs in a noise box and the temperature there can reach 85 degrees, while the ASIC works great.
1 cubic meter = 35, 31 CFM.
Asic has this characteristic of 190 CFM or 333 m3h, but this is practically at maximum speed.
If with the maximum reserve, then I would consider 350x4 or 1400m3h

[Ers900]

I have 2 and 2 connected, the 2 on the right work perfect, as so for the third on the left, but with the fourth one connected, any random 1 miner will go offline. (My guess is lack of air in). The ducts run 5 metres to the exit holes.

I’d want to get some advice and confirmation before I start dishing out money on extra fans etc. hoping that it will be a definite resolution.

https://imgur.com/a/7jFzZmf

https://imgur.com/a/v64d1ai

[FP91G]

I have 2 and 2 connected, the 2 on the right work perfect, as so for the third on the left, but with the fourth one connected, any random 1 miner will go offline. (My guess is lack of air in). The ducts run 5 metres to the exit holes.

I’d want to get some advice and confirmation before I start dishing out money on extra fans etc. hoping that it will be a definite resolution.

I added your pictures. It's a good idea to dry the laundry with warm air, but I would like to see the room for the ASICs and the way to connect the ASICs to the ventilation.
It is very difficult to understand the reason for the ASIC shutdown from this picture. I would start with an open door to the room at least for testing, if possible.
If the ASIC was shutting down due to overheating, you would see this by the increase in temperature or by errors in the logs.
I would also monitor the voltage in the network, it happens that under load it drops and the equipment switches off.

[Ers900]

I have an 100amp fuse box, with 4 miners running it draws at most 70amp I believe.

The miners are here in this (3Metre x 2Metre) room. Pictures shows 3 but I have 4.
https://imgur.com/a/3zHxFB5
Excuse the wooden platforms underneath the miners, I have a proper rack set up now.

I have these air gap vents in the basement entrance door
and same size whole 40x40 in a brick wall next to the door.
https://imgur.com/a/7yhDByL
Distance from the green entrance door to the miner room in the back is about 6-7metres.

All doors are open to allow air to flow into the miner room.
I even tried running 2x 300cfm fans blowing towards the entrance of the miner room. Around 3metres away.
Didn’t have ducks to actually aim it into the room.

[FP91G]

I have an 100amp fuse box, with 4 miners running it draws at most 70amp I believe.

The miners are here in this (3Metre x 2Metre) room. Pictures shows 3 but I have 4.

Excuse the wooden platforms underneath the miners, I have a proper rack set up now.

I have these air gap vents in the basement entrance door
and same size whole 40x40 in a brick wall next to the door.

Distance from the green entrance door to the miner room in the back is about 6-7metres.

All doors are open to allow air to flow into the miner room.
I even tried running 2x 300cfm fans blowing towards the entrance of the miner room. Around 3metres away.
Didn’t have ducks to actually aim it into the room.

It's better now, at least you can think about something.

If you think that you have a problem with the supply ventilation, then with the door open you should not have problems with the air supply. If your door is closed, then ventilation is needed.

This theory needs to be tested with the door open.
I advise you to monitor the voltage, I think that this may be the reason.
Do you have any old refrigeration equipment or industrial or welding equipment nearby?
I also looked at the ASIC logs. I think this will clarify the problem a little.

[Ers900]

Thanks for your reply, would a voltage issue cause the fans speed to increase resulting higher temps then resulting in fault coming up? I believe I’m still under the 80% rule though.

[FP91G]

Thanks for your reply, would a voltage issue cause the fans speed to increase resulting higher temps then resulting in fault coming up? I believe I’m still under the 80% rule though.

The speed of the fans is regulated by the controller based on the temperature of the chips. But you have a great temperature. What is written in the logs? Why don't you check them?
When connected in parallel, the voltage at all points is the same, you need the simplest voltmeter to check this.

[Ers900]

33 chain=0 msg="is_temp_return @ design 110 valid return nun.+ time="2025-04-29
12:28:28" level=error pid=24033 chain=0 error="asic num error" msg="detected asic num less than design(0<110)"
time="2025-04-29 12:28:28"
level=notice pid=24033 chain=0 msg="detect chip num 0 design num 110"
time="2025-04-29 12:28:29"
level=notice pid=24033 chain=0 msg="is_temp_return @ design 110 valid return_num:*
time="2825-84-29 12:28:32"
level=error pid=24033 chain=0 error="asic num error" msg="detected asic num less than design(0<110)*
time="2025-04-29 12:28:32"
level=notice pid=24033 chain=0 msg="detect chip num ® design num 110"
time="2025-04-2912:28:33"
level=notice pid=24033 chain=0 msg="is_temp_return 0 design
time="2025-84-29 12:28:36"
level=error pid=24033 chain=0 error="asic num error" msg="detected asic num less
time="2025-84-29 12:28:36"
level=notice pid=24033 chain=0 msg="detect chip num 0 design num 110"
time="2025-84-29
12:28:36 level=error pid=24033 chain=0 msg="failed getting chip temp"
time="2025-04-29
12:28:36" level=fatal pid=24033 chain=0 msg="read no sensor chip temp
time="2025-04-29 12:28:36"
level=fatal pid=24033 chain=® msg="uneffective temp exceed limit (3), uneffective_value_count = 4!"
time="2025-04-29 12:28:36" level=notice pid=24033 chain=0 msg="chain 0, chip temperature array:
-64
-64 -64 -64 -64 -64 -64 -64 -64 -64 -64
NE


This log was from back when I started the 4th one in the miner room, I feel like it threw off the airflow balance.
Only 1 random one out of the 4 would turn off and 3 would remain running at a stable level.
(So for example could be miner number 2 or miner number 3, the problem isn’t specifically with 1 of the miners.)

[FP91G]

ChatGPT:
"The logs from your Bitmain Antminer L9 indicate several issues related to the ASIC chips and temperature sensors. Here’s a breakdown of the key messages and their implications:

ASIC Number Error:
The repeated error message error="asic num error" msg="detected asic num less than design(0<110)" indicates that the miner is detecting 0 ASIC chips when it is designed to have 110. This suggests a failure in the ASIC chip detection process, which could be due to hardware issues, such as:
A malfunctioning or disconnected ASIC chip.
A problem with the control board that communicates with the ASIC chips.
Power supply issues that prevent the chips from functioning properly.
Chip Temperature Detection Failure:
The log entry level=error pid=24033 chain=0 msg="failed getting chip temp" indicates that the miner is unable to read the temperature from the ASIC chips. This is further supported by the message level=fatal pid=24033 chain=0 msg="read no sensor chip temp", which suggests that the temperature sensors are not functioning or are not connected properly.
The temperature readings of -64 for all chips are indicative of a sensor failure, as this value is typically used to represent an invalid or uninitialized temperature reading.
Ineffective Temperature Exceed Limit:
The message level=fatal pid=24033 chain=0 msg="uneffective temp exceed limit (3), uneffective_value_count = 4!" indicates that the miner has encountered too many invalid temperature readings (in this case, 4 readings of -64), which exceeds the acceptable limit. This could lead to the miner shutting down to prevent damage.
Possible Causes and Solutions:
Hardware Issues: Check the physical connections of the ASIC chips and the temperature sensors. Ensure that all components are securely connected and that there are no visible signs of damage.
Power Supply: Verify that the power supply is functioning correctly and providing adequate power to the miner. Insufficient power can lead to components not operating as expected.
Firmware Issues: Ensure that the miner is running the latest firmware. Sometimes, firmware updates can resolve detection and communication issues.
Cooling System: If the miner is overheating, ensure that the cooling system (fans, heatsinks) is functioning properly to prevent thermal shutdowns.
Replacement of Faulty Components: If specific chips or sensors are determined to be faulty, they may need to be replaced.
Next Steps:
Inspect the hardware connections and power supply.
Update the firmware if not already done.
Monitor the miner for any changes after making adjustments.
If issues persist, consider reaching out to Bitmain support or a professional technician for further diagnosis and repair."

This is either a power supply problem, when there is not enough power and the controller starts working with an error, or a cooling problem. The second is checked with an open door, and the first with a multimeter. Sometimes in such cases Vnish firmware is recommended, but you need to be sure that there are problems with power supply.