Ryujin II Pump Speed Capped at 1770rpm

[s09931]

Hi, my Ryujin II pump speed is capped at 1770rpm. I can lower the speed and it is reported correctly, but if I put the speed at max in armoury crate, the max it will go to is 1770rpm. Isn't the max supposed to be 2800rpm?

PC Specs:
7950x
x670e Hero
32gb DDR5
Strix OC 4090

[duranx3]

I have the same problem, I don't think it's a driver problem since I have two identical AIOs, one with ARGB which is the one that gives me the problem and the other without RGB with noctua fans, the version with ARGB my pump does not exceed 1800 rpm I did the experiment to put the micro fan to 0 and turn off all the fans including the radiator ones so that you can hear if the pump makes more noise from 50% to 100% and there is no difference, I did the same test with the AIO in good condition and if there is an increase in sound according to the RPM, talk to Asus Spain and they told me that my board is compatible now I am in the process of RMA with the page where I purchased my defective unit.

[RogScott]

@duranx3, please update us after your RMA. I still have the same issue -- Armoury Crate reports the speed at about 1700 on any setting 65% to 100%, whereas the stated speed at 100% should be more like 2800.

No cooling problems, so I'm leaning towards this being a problem in Armoury Crate reporting incorrectly (and the pump being just fine).

@Murph_9000 -- Are you still seeing correct speeds in latest AC for your Ryujiin-II?

[Murph_9000]

RogScott wrote:
@duranx3, please update us after your RMA. I still have the same issue -- Armoury Crate reports the speed at about 1700 on any setting 65% to 100%, whereas the stated speed at 100% should be more like 2800.

No cooling problems, so I'm leaning towards this being a problem in Armoury Crate reporting incorrectly (and the pump being just fine).

@Murph_9000 -- Are you still seeing correct speeds in latest AC for your Ryujiin-II?

Yes, I'm still seeing all speeds from 840 up to 2800 rpm correctly reported on the Ryujin's LCD in fixed mode, and reported speeds well above 1700 both in AC and on the LCD in smart mode when running something like CPU Burner or Cinebench (and they are the speeds expected from the curve set in smart mode). I.e. all seems good here on latest AC, and I've still never encountered the 1700 problem.

[duranx3]

RogScott wrote:
@duranx3, please update us after your RMA. I still have the same issue -- Armoury Crate reports the speed at about 1700 on any setting 65% to 100%, whereas the stated speed at 100% should be more like 2800.

No cooling problems, so I'm leaning towards this being a problem in Armoury Crate reporting incorrectly (and the pump being just fine).

@Murph_9000 -- Are you still seeing correct speeds in latest AC for your Ryujiin-II?

As I already mentioned, I have 2 identical AIOs, one is ARGB and the other is not, it does the experiment of putting the pump at 50% and 100% and there is no audible difference in the pump, when in the AIO in good condition (the one that is not argb ) if you notice a noise increase when it goes from 50% to 100%, I'm afraid it's not a software problem and it's from the pump, if it's true that even limited to 1700 rpm it works very well with a 12900k it keeps it cold, but I feel that it does not give its full potential, I have just ordered a third similar AIO and I am going to send the damaged one to RMA.

[duranx3]

I bought a new drive (my third) and it also has the problem, this is worrying enough, I think I'll go with another AIO model, it's a shame.

https://www.youtube.com/watch?v=lSpnhSq_dFY

[RogScott]

@duranx3 --

Yes, very disappointing for such an expensive AIO! My Ryujin is the non-RGB 360 version, by the way.

I watched your video. Everything is the same as what I see. All the Ryujin components shown in AC are the same as mine, including firmware version. The pump speed acts exactly as yours does.

These are Asetek pumps, which are pretty high quality, so it is hard to believe the pump itself could be defective on so many Ryujin units. That still makes me think the problem is with Armoury Crate.

Is it possible that when AC is set to 50%, it is already driving the pump at 100% / 2800rpm but displaying the speed as 1700? Changing the setting above 50% would not cause the sound to change if the pump was already running 2800 when set to 50%.

But some units apparently show correct speeds, like the one @Murph_9000 has. Perhaps some units have a different Asetek pump internally, and AC does not correctly detect that, and then doesn't control or read speed correctly?

Above are just speculation because I have no way to check the actual pump speed -- I can only see what AC tells me. But again -- it is hard for me to accept that so many Asetek pumps could be defective, which leads me to suspect Armoury Crate. I suppose it could be a firmware issue, but I'm guessing your other pump (the one that correctly shows 2800) is using the same firmware version.

What are the hardware IDs for your Ryujiin? Are they the same for the one that does 2800 and the one that limits at 1700? My hardware IDs are:
USB\VID_0B05&PID_1988&REV_0200&MI_00
USB\VID_0B05&PID_1988&MI_00

[Murph_9000]

Yeah, I get the same hardware IDs in Device Manager:

USB\VID_0B05&PID_1988&REV_0200&MI_00
USB\VID_0B05&PID_1988&MI_00

AC tells me my Ryujin's firmware is 0104.

With multiple people reporting units with a max reported speed of 1700, it's got to be either a bad batch of them, or a design variation. Something like a different motor, e.g. with a different number of poles, could do it; or a different motor controller/drive. It could also be a difference on the measurement side; I don't see anything to really point at it being the motor running slower vs the speed being reported slower.

[RogScott]

Murph_9000 wrote:
With multiple people reporting units with a max reported speed of 1700, it's got to be either a bad batch of them, or a design variation.

I have done some more research since my last post, and am at the same conclusion. I think it is "specific Ryujin units". I no longer think it is an Armoury Crate issue.

• I can see from one of the log files that AC is "commanding" the Ryujin to set the correct duty cycle for the pump. And my Ryujin follows that -- with expected pump speeds -- from the minimum 840rpm to about 55%, or 1710rpm. It just doesn't go any faster when AC commands a duty cycle above 55%.
  
• It seems unlikely to be a "bad batch of pumps". If the motor responds correctly to a pwm duty cycle up to 55%, there is just no logical reason why it wouldn't respond correctly when the duty cycle increased above 55%.
  
• I have seen a lot of other cases with other hardware where the rpm signal was incorrectly interpreted, often due to a pulse-per-rotation problem. However, that usually results in a reading that is half or double the actual rpm across the entire range, and that is not what we're seeing here.
  
• I considered that perhaps -- due to a shortage or mistake -- Asus substituted some max-1710rpm pumps in certain units. But that doesn't make sense, since the speeds are correct up to 55%. If they substituted a pump with a max speed of 1710, then I'd expect 55% duty cycle to be closer to 940, not 1710.
  
• The issue has been reported across multiple AC versions, different motherboards, different processors.
  
• The issue does not exist for some units, with exactly the same AC versions, Firmware, Motherboards, Processors, etc.

The explanation that fits the best would be some kind of problem with the onboard PWM controller circuitry on some Ryujin units, such that when commanded to generate a PWM wave over 55%, they just keeps generating 55%. Something like an incorrect value component on a run of the pcb. Just another guess, but it fits observed data the best so far.

In any case, it does look "unit specific" to me, so an RMA is likely the only solution. However, an RMA could also end with a replacement unit that has the exact same behavior (which is what seems to have just happened to @duranx3).

I am currently satisfied with the cooling performance (on a 12900k), so running 1/2 rated speed is just an "annoyance" to me and I cannot justify the hassle of an RMA. Someone with a hotter chip would be impacted more. For the price paid, the quality control should definitely be better -- and the RMA process should include cross-shipping (i.e., advance RMA) at the very least.

[Murph_9000]

I take it the graph and table of speeds generated by the auto-tuning function are also capped at the 1700-ish speed? Here's what mine looks like for the pump:

[RogScott]

Murph_9000 wrote:
I take it the graph and table of speeds generated by the auto-tuning function are also capped at the 1700-ish speed? Here's what mine looks like for the pump:

The speed table for the pump looks as it should, going up to 2800, but I do not think that table is dynamically generated like it is for the fans. Because it is shown and is identical after a new install, before ever running the auto-tuning, and it is always 100% identical (every entry) for everyone.

I think the table for the pump is hard-coded, and not generated or updated by auto-tune.