silence of the fans, making my HP/Tektronix lab gear a lot quieter

[av500]

Hi there,

first post here, stepping into the world of collecting test and measurement gear for no reason

so, I have some gear that I use daily and where the fan noise really got on my nerves, so I decided to do something about it:

HP/Agilent 33120A, option 001: why this thing needs a such a loud fan, I have no idea. added a Noctua NF-A4x20 FLX and used a 47R dropper resistor to make it run even slower



HP/Agilent 6632B: the fan has to move a lot of air when needed, but I replaced the aging Papst fan with an equivalent Sunon MF60251VX-1000U-A99, it's a bit quieter now and there is actual fan speed control so it's bearable. While at it I also added front panel binding posts, including the sense lines and earth GND. had a leftover piece of panel that matched the front panel color almost exactly, milled that to size and aligned the holes with the 0.75" spaced cutouts in the plastic bezel. I opened the the 1k resistors between output and sense on the power board and installed new ones right at the binding posts.






Tektronix TDS3014: how I managed to work next to this loud fan for 15 years is beyond me, replaced the fan with a Noctua NF-A9 PWM (PWM is not needed but was what the shop had) and added a 220R/1W resistor to drop from the 24V fan supply to something below 12V, runs at 65mA now vs the nominal 75mA and is very quiet.




HP/Agilent E3631A: this one comes with a very loud and uncontrolled fan and is very annoying since I need it mostly for dev work. replaced the stock fan with a Noctua NF-A8 PWM and added an intelligent fan controller. I drilled and glued 4 NTC thermistors into the 3 heatsinks, next to the linear pass transistors and the 6V rectifier diode. used a small PCB from an existing project with an STM32F030 to measure the 4 NTCs and PWM control the fan. fan runs between 40% at room temp and goes up to 100% when the heat sinks reach 40 degC, running the PSU at full load for an hour never exceeded the 40 degC. had to be careful when debugging since the fan runs from GND and -16.5V, so used an USB isolator to talk to the CPU while writing tzhe code. now it sits dead silent when powered on and the fan only revs up when really needed...





thanks

[Zucca]

Welcome to TEA!

https://www.eevblog.com/forum/testgear/test-equipment-anonymous-(tea)-group-therapy-thread/

Well done! When I am replacing the fan I always make sure the air flow is at least equal the old one... but I am sure you did the homework before replacing them.
 

[av500]

Well done! When I am replacing the fan I always make sure the air flow is at least equal the old one... but I am sure you did the homework before replacing them.
 

yes, I researched airflow, but in some cases decided to go below the original spec, since my gear sits in a clean lab and not on a noisy factory floor wedged in between tons of other hot stuff and with clogged air vents

[thinkfat]

Welcome to the party, old pal ;-)

[anotherlin]

Are you developing eurorack synthesizer modules?
Bottom left corner of photo looks like euroracks.

[av500]

Are you developing eurorack synthesizer modules?
Bottom left corner of photo looks like euroracks.

I do   see https://vpme.de

[Zucca]

since my gear sits in a clean lab and not on a noisy factory floor wedged in between tons of other hot stuff and with clogged air

"Lab T Ambient Max" is also an importan variable. Unfortunately in summer my lab goes above 30°C so I am scared to cut the air flow.

 

[av500]

and another device, the HP/Agilent 66309D, added a Noctua NF-A6x25 FLX fan and a front breakout panel for all the rear connections

[thinkfat]

I ordered one of those fans as well. I'll put that into my own 66309D to tame its noise. It's one of the things that keep me from using it, the sound it makes is absolutely obnoxious.

[CalMachine]

Very nice!  And a very good idea 

I put 60mm x 25mm Noctua fans in my 2 3458A's and 3245A.  They are damn near silent now.  The internal temperatures of the 3458As dropped a few degrees... so the original fans were pretty shot.   

[Zucca]

https://www.eevblog.com/forum/testgear/keysight-34461a-fan-noise/msg2838026/#msg2838026

[thinkfat]

Just exchanged the fan in my 66309D! What a relief! 

I noticed that the fan in my unit must have been exchanged at a point. There was some noname China fan inside instead of the Papst 612. It was _loud_!

Now, this power supply is alltagstauglich (besides being big, heavy and I have no space on my desk for it).

[av500]

ran it today for an hour with 3A on one channel and 1.5A on the secondary, exhaust vent temp was around 57 degC at 21 degC ambient

[Zucca]

ran it today for an hour with 3A on one channel and 1.5A on the secondary, exhaust vent temp was around 57 degC at 21 degC ambient

+36 Delta T °C is not too much? Looking at the spec the Noctua NF-A6xFLX is 17,19 CFM and the original Papst 612NN is 24,7 CFM.
Just curious, I am not in the position to judge your decision (aka I do not have too much experience with replacing fan...)

[av500]

shorting both rails to ground at full current is the worst case load. and as said before, the fans are way overspecced since these units have to keep working sitting in a test rack running 24/7 and with nobody ever cleaning the dust out of the fans. for the occasional home use I don't see an issue. ymmv

[Zucca]

What is your safe delta T limit?

If you had measured for example 65°C - 70°C instead of 57°C would have been a problem?

Thanks!

[av500]

to be honest, I dont know. I dont plan to run this PSU with both channels in max current short circuit anyway

[thinkfat]

to be honest, I dont know. I dont plan to run this PSU with both channels in max current short circuit anyway

Neither will I. Silence is bliss. If my "lab" gets too hot in the summer I'll be in the park under some trees or in a swimming pool, not sweating in front of a power supply.

[Zucca]

According to the 66309D Manual:

Unless otherwise noted, specifications are warranted over the ambient temperature range of 0 to 55 °C.

So if your lab do not go to 55°C it is safe to lower the air flow (and the noise of a fan).

Now I would be interested to understand or calculate of much lower it could be,

Stock Situation: TAmb Max 55°C and 24,7CFM.
User Situation: TAmb Max=T Lab Max=40°

What is the minimum CFM of the new fan in the User Situation?

Interesting....

[thinkfat]

According to the 66309D Manual:

Unless otherwise noted, specifications are warranted over the ambient temperature range of 0 to 55 °C.

So if your lab do not go to 55°C it is safe to lower the air flow (and the noise of a fan).

Now I would be interested to understand or calculate of much lower it could be,

Stock Situation: TAmb Max 55°C and 24,7CFM.
User Situation: TAmb Max=T Lab Max=40°

What is the minimum CFM of the new fan in the User Situation?

Interesting....

Well, you're an EE You'll figure it out. But the relation between cooling effect and airflow is certainly not linear.

I found this: http://www.crydom.com/en/tech/hs_wp_fa.pdf

There's a table with a correction factor for the heatsink rating in K/W against the LFM of the fan. I think as a rough estimate you can convert the CFM of the fans to LFM and the check in the table the difference in correction factor and that is basically your answer.

But by just looking at the table and seeing how little difference an airflow difference between say 400 and 600 LFM makes - I think she'll be alright.

[Zucca]

Well, you're an EE You'll figure it out.

Thanks TF, I will do my homework in the next days too.

[Zucca]

I could not resit...

AF_u = AF_s (Tout_s - Tin_s) / (Tout_u - Tin_u)

where

AF_u = air flow in user situation (new user fan spec)
AF_s = air flow in spec situation (old stock fan spec)
Tout_s =Max air Temp going out the instrument at max power dissipation with stock fan
Tout_u =Max air Temp going out the instrument at max power dissipation with new user fan
Tin_s = Max air Ambient Temp according to spec.
Tin_u = Max air Ambient Temp according to user lab.

To solve the puzzle we need to know either Tout (do a stress test) or the max. Power the fan needs to take away the instrument (you surely know why).

Now back to av500 situation and my question, his max Delta_T measured is 36°C in his max power user situation @ 17,19 CFM (I assume). This means in stock condition (24,7 CFM) at max Temp 55 °C running the same test:

Tout_s= 55 + 17,19 * 36 / 24,7 = 80 °C almost pefectly.

so assuming the HP eng designed for a Tout air max =80 °C, at a  max Tin_u=40°C (most of our hobby Labs) the new CFM must be at least (assuming Tout_s=Tout_u):

24,7* (80-55)/(80-40) = 15,43

or in other words, since he got a new fan of 17,19 CFM, his new Tmax Amb temperature is:

80 - 24,7* (80-55)/17,19 = 44 °C

so I dare to say av500 is safe if his T Lab is below 44°C and he does not dissipate more power than in his stress test.

Sorry for any mistake, I was hungry and now I go the have lunch.

[HighVoltage]

Very good job on your upgrades!
Thanks for sharing the detailed model numbers of the fans you used.

Almost all my Agilent / Keysight gear has been updated to new fans or reduced voltages.
Sometimes a 50 Ohm or 100 Ohm series resistor is all its need to bring the noise level down to acceptable values.

What surprises me mostly is that some Agilent gear does not have fan speed regulation, although
it is so easy to implement in the design.

[nctnico]

I could not resit...

AF_u = AF_s (Tout_s - Tin_s) / (Tout_u - Tin_u)

where

AF_u = air flow in user situation (new user fan spec)
AF_s = air flow in spec situation (old stock fan spec)
Tout_s =Max air Temp going out the instrument at max power dissipation with stock fan
Tout_u =Max air Temp going out the instrument at max power dissipation with new user fan
Tin_s = Max air Ambient Temp according to spec.
Tin_u = Max air Ambient Temp according to user lab.

If you look at graphs which show thermal impedance versus airflow you'll notice there is no linear relation between those. Without checking the actual heatsink specifications it is impossible to determine the new maximum temperature.

However, in my experience the fans in HP/Agilent equipment move way more air than necessary. I have upgraded most of them to temperature controlled fans and the fans only switch on under heavy loads.

[klausES]

In addition to a suitable fan (speed, wing shape, low noise of the commutation, decoupling),
a temperature-controlled speed control with adjustable minimum speed, adjustable slope (start of the slope and hysteresis) is always recommended.
In addition to the design, the intelligent placement of the temperature detection is also important.

As a percentage, in practice you will mostly be in the area of the very quiet minimum speeds, or only slightly above it at a slightly higher speed
and if the requirements increase, power dissipation increases or the ambient temperature rises in midsummer,
(I still test the latter in one small room with fan heater), if the curve is correct, sufficient throughput is guaranteed.