Ideal temperature ? 35C? 40C? 42C?

[Alex Lubensky]

Thank you Patrick, I just made a test strangely, I have to set the fans to 46 so that the sensor is a 42.... And when it arrives at 42 and I put the fans at 42 it goes down to 39..

It depends on the temperature of your location also. In my case (Scarlet MX, Eastern Europe autumn) those figures usually don't match. My sensor usually runs about 38-42 degrees indoors, while my brain set to 55 degrees. Outdoors it may be even lower - about 34 degrees, while the brain still runs 55.

 

[Rich Miro]

So it seems that most people like between 42-44C. How long does it normally take to get your camera warmed up? Do you cover your fan so that you reach target temp faster? any tricks of the trade out there?

 

[Raphaël_FR]

Hi, I think that if you close the fans so that the temperature rises faster you risk creating condensation somewhere. I imagine that RED has created a kind of mild temperature rise to avoid this condensation effect... I would like to know more about the sensor's cooling / heating system, if it works with a Peltier effect module like on the BMC cameras. By pure desire to understand the things that surround me haha

 

[Viktor Sloth]

for me it's near impossible to keep it at 42c...
if it's not 25c plus, it's so difficult.
do you guys use heat pads or something to keep it warm?

 

[Aaron Lochert]

for me it's near impossible to keep it at 42c...
if it's not 25c plus, it's so difficult.
do you guys use heat pads or something to keep it warm?

I live in Arizona.

 

[Jay Silver]

What I've learned over the last few months is that it all depends on where you're going to shoot. The highest setting you can go to on the target temp is 48 degrees. If you set your auto black shade at 44, then you'll have 4 degrees up and 4 degrees down. It will help your camera maintain a higher temp when the ambient temp is so high out there.
Same goes for shooting in the cold. If you're shooting where its' cold, your camera should be black shaded at 4 degrees up from the lowest target temp. having multiple auto black shades will cover you for almost every temp you might encounter.

 

[Antony Newman]

During 'lockdown' I have tasked myself with understanding the compromises that calibrated SENSOR TEMPERATURE has on SENSOR NOISE. (My ultimate aim - to record sound without the 'whine' and 'white noise' of the Monstro's fans).

In my following test

+) I allowed the sensor to stabilise at 43,44,45 & 46c : black shaded at this 'calibration temp' : recorded a clip at the same temperature
+) Then Cooled the sensor temp to 1 degree below the calibration temperature - started recording - manually set fan 20% and recorded a clip with the Temperature rising - noting down the exact points the sensor temp increased.
+) clips then extracted into Resolve and all equally pushed to compare noise.

AJ

 

[Antony Newman]

Thoughts

+) When leaving the camera to STABILISE for 30 mins at a temperature : Calibrate : Record a clip - there was OFTEN MORE NOISE than when recording at a SENSOR temperature 1 degree higher.

+) At 1 deg C higher than the Calibrated temperature - the noise levels for 43,44,45 and 46 were all similarly low.

+) All tests where the SENSOR TEMP went BELOW the Calibrated temp caused an increase in noise (which remained at this same elevated level for a few degrees below the calibrated point)

AJ

 

[Antony Newman]

(Current) ADAPTIVE Fan Control

Watching REDs adaptive FAN algorithm - they appeared to cater for this noise profile.

+) When the SENSOR TEMP falls below the CALIBRATED TEMP - there is a Very rapid drop in 10% increments to try and get the SENSOR TEMP back up.

+) When the SENSOR TEMP is above the CALIBRATED TEMP - there is no 'emergency response' - but a gradual decline - because when the SENSOR is at a higher temperature - noise is not a problem. (It also reminded of a JL comment made about the Monstro liking the heat).

A way to visuaise the SENSOR TEMP under adaptive is that it is 'more' like a the height of bouncing ball == it tends to spend more of its time at a HIGHER temperature than the CALIBRATED (floor) temperature.

CALIBRATING using ADAPTIVE temp

If you sent the camera to Adaptive Fan to 45c - and calibrate -
the actually Calibration temp might wandering up to 45.9 degrees C - or even into the 46's

If you then use your camera (as in my testing) the SENSOR TEMP will swing above and BELOW the calibration temp (even on ADAPTIVE) by 0.5 - 1 deg C below the actually calibration temp - which would means increased Sensor noise.

AJ

 

[Antony Newman]

Minimal Sensor Noise using current Firmware and Adaptive Setting

To minimise the ADAPTIVE algorithm dipping the SENSOR TEMPERATURE below the Calibration Temperature (and increase sensor noise) - Manually Set Fan to the HIGHEST fan Speed whilst still keeping the Sensor at the same SENSOR Temp.

Example : Fan % : Temp

43% : 44c
44% : 43c
45% : 43c
46% : 43c
47% : 43c
48% : 42c

If CALIBRATED at 47% fan speed - and Recording with ADAPTIVE in the 43c range and the temp goes down by just a fraction of deg C, the algorithm responds by immediately to raising the sensor temp (and avoiding noise).
This setting would MINIMISE sensor NOISE for a 43c Calibration

If CALIBRATED at 44% fan speed - and Recording with ADAPTIVE in the 43c range - all recording will be at or BELOW the CALIBRATED temperature.
This setting would MAXIMISE Sensor NOISE for a 43c Calibration.

AJ

 

[Antony Newman]

FAN ALGORITHMS - Thoughts

+) The RED SDK : Camera Sensor temperature : Resolution of 1.0 deg C.

If this is stored internally with greater resolution - then the ADAPTIVE 'hunting' (which steers like a canal boat hitting the sides before changing direction!) - would allow the adaptive algorithm to be (say) +-0.1c rather then +-0.5c

If there was a special 'aggressive' ADAPTIVE MODE where the fans go from 0 ...100% in order that a +-0.1c variation could be maintained during the Calibration - this would give approx a 1 deg c headroom Above the calibration temperature compared to the current method.

Suggestion : Change the ADAPTIVE algorithm TARGET TEMP during recording to be 0.5 degree above the Calibration temp (as it would minimise sensor noise)

+) CORE TEMPERATURE

I could not find evidence that the ADAPTIVE algorithm uses the Core Temperature in its steering of the Fan Power to reach the Sensor Temperature. FAN speed ramping only appeared to coincide with Sensor Temperature fluctuations. (use of a 2nd variable would help if a 'critical damping' fn was the aim)

+) QUIET MODE

I am not convinced this more works as intended - on several occasions the fans ramped up high levels even though the Sensor Temp was already more than 1 degree colder than the Calibration temperature.

My guess that people don't use this mode any more? (Or that somehow the internal copy of the target temperature was used in preference to the one exposed in the SDK Target Temp. I note in the manual that this is only supposed to happen when a calibration map is downloaded).

However :As quiet did not exactly meet my needs I did not explore this further or try and track down the root cause.

AJ

 

[Antony Newman]

Suggested new Fan Mode : FAN SPEED LIMIT IN RECORD

The unwanted sounds from the camera are :
+) White noise from top two camera intake fans (approx 3dB louder than air exiting exhaust)
+) High frequency fan bearing sounds (which amps up from about 45%)
+) Very high frequency sound from OLED screen (can be blocked if screen is covered)

The problem with ADAPTIVE is that it MIGHT ramp up to a point where White Noise, or HF Bearing sound is intrusive.
With a chart of absolute sound level vs FAN percentage - What is known is the top speed the Fans can run during recording before the audio is ruined.

Armed with the existing details parameters that the user already updates in fan modes :

Proposed Parameters in : ‘FAN SPEED LIMIT IN RECORDING’ mode:
+) 'Maximum Record Speed' - is the maximum record speed used when recording (which would PROTECT me from unwanted sound contamination)
+) 'Post Record Delay' - is used as a time after record stops that the FAN speed is allow to ramp up to a faster speed (that will cook the speed of the fan down). The rate of increase to this new speed will be slow initially to rapid (to minimise 'drama' on set / ease up the speed over a few seconds).
+) 'Maximum Preview Speed' - After recording, the sensor might be 49c and I want to get down to 46c : The Fan will increase over 'Post Record Delay' up to this new level (eg 100%).
+) 'Target Temperature' : After Recording - The fans will run at 'Maximum Preview Speed' - UNTIL the 'Target Temperature' is reached. After that point the Camera switches to an Aggressive ADAPTIVE - taking into account the temperature of the Core Temp - and aiming to keep the temperature between 0.0c and 0.5c ABOVE the TARGET TEMPERATURE. In the event that a CALIBRATION is loaded - this Number is updated over the SDK (I don't believe it is currently).

[ +) ALSO : If there is more temperature resolution available - please alter the RCP for temperatures so that external programs can display. This might be useful for those wanting finer control during manual calibration (ie where the adapting Algorithm was not aggressive enough) ]
The benefit of this mode would be for people who want to Record -> cool down -> Record with the minimum time, and Not have to depend on an operator to do it each time.

AJ

 

[Daniel Förderer]

Very interesting information!
Thanks a lot!

 

[Antony Newman]

Thanks Daniel.

Hopefully these results will be enough for Red to verify / and create a tweaked fan mode for those striving for the highest quality audio recorded in Camera.

AJ

 

[Les Hillis]

I had another look into this during the hottest part of Summer a couple of months back.

I also found the 'Adaptive' and 'Adaptive Preview/Quiet Record' don't work properly (or at least as I'd like them to), wherein the fans don't pull the temp far down enough when in 'Preview Mode' and so kick in too soon/too high when in 'Record Mode', so I had to explore the Manual mode.

I'll probably have to look into it again when the ambient temperatures drop to their lowest over Winter, but for the time being I've just made a 42°C & 46°C Calibration to suit the higher ambient temps I'm currently experiencing.

That's partly based on allowing a 4°C range either side of the target sensor temp (ie. keeping the T/E indicator in the green). At 5°C above the target (calibrated) sensor temp the on-screen indicator turns yellow, but within 4°C I haven't seen any degradation in the recorded image so I don't mind if the temp wanders a bit within that range.

Something else I noticed was that the sensor temp was consistently 10-11°C above the ambient temp when the fans were at 100% in Preview mode, as measured with a probe placed at the camera's fan intake. What that meant was, with a 46°C Calibration, I could still record in a 36°C ambient environment.

However, I also found that I'd like to keep the fans running at 40% or less to minimize their noise, and that in turns results in a 17°C difference between ambient and sensor temps when actually recording.

As an actual example -

From -

33.4°C Ambient Temp.
100% Preview Manual Fans.
44°C/62°C Sensor/Camera Temp.
(With 46°C Calibration).

To -

33.8°C Ambient Temp.
40% Record Manual Fans.
50°C/75°C Sensor/Camera Temp.
(With 46°C Calibration).

In - 6:31 mins

So, 6-7 mins record time (or 5-6 mins @ 36°C Ambient Temp.) before hitting 51°C Sensor/Camera Temp and going out of calibration (triggering Yellow indicator).
Then 1-2 mins at 100% fans to get back to 46°C Sensor/Camera Temp.


For me, all the temps we can calibrate at will give a good image within about 4°C of the chosen calibration, so it's all about finding a balance between what ambient temp you're recording in and what fan noise level you find acceptable. For me, choosing a 40% fan noise level, I'd either have to limit the amount of recording time per shot, or control the ambient temp (to somewhere below 33°C with a 46°C Calibration) to give longer record times per shot or eliminate having to cool down between takes.

 

[Andrew Reese]

Would 43-44 degrees celsius also be recommended for Gemini sensor?