Low noise fan replacement for EBM W2S107-AA15-16 115V~ (Datron/Wavetek 4808)?

[carl_lab]

Can anybody recommend a low noise replacement for discontinued Papst EBM W2S107-AA15-16 115V~ (120x120x38mm)?

Unfortunatelly I cannot find a datasheet of the original fan...
So maybe the only way to determine adequate air flow is measuring Datron 4808's internal temperature using the new fan, compared with old fan (it's still working but as loud as a starting jet plane...).

[HighVoltage]

Do you have to have it hooked up to 115V AC?

I had a similar situation in a totally different instrument and installed a German BeQuiet 12 V DC ultra quiet 120x120 mm fan.
But I had 12 V available in my instrument.
 

[carl_lab]

Do you have to have it hooked up to 115V AC?

The calibrator has a transformer with two 115VAC windings.
For 230VAC mains voltage (I'm localized in germany) both windings are in series, for 115VAC use they are in parallel.
I suppose the fan in both cases is in parallel to one or the 115V windings. So that's the typical auto-transformer circuit (="Spartransformator").
Probably I could use a 230VAC fan instead of the 115VAC one if I hook it up to the "serialized" transformer windings.

I'm not sure, if changing to a low voltage DC fan, as you did, will inject additional unwanted electrical noise to the calibrator's guts (current peaks generated by DC motor commutation etc.).

[HighVoltage]

If you have 230V AC available in the instrument, there are plenty of low noise Papst fans available in Germany.
Look for instance at the Papst 4580Z, it has only 30 dB(A) at 1900 RPM and 115 m^3/h in Volume.

[alm]

I could not find a datasheet either. There are instructions how to decode the part number on page 8 of this catalog, but knowing that AA stands for a mechanical design, 15 for an electrical design, and -16 for mechanical variant is not particularly helpful. Unfortunately that catalog does not appear to include any fans with 107mm rotors.

You could measure the current it draws and possibly the rotation speed. That should give you some comparison basis. Without a datasheet it is hard to find a good replacement without measuring the internal temperature at various places with both fans. Even if the air flow is the same in free air, it may very well be different in the presence of back pressure. There is a nice curve in the datasheet, if you can find the datasheet . If you just measure the current at 115V, calculate power and compare to 120mm 115VAC fans from this list you should be able to find a ballpark rotation speed. Compare the shape of the impeller (larger hub and stubbier blades usually means higher static pressure) and you should be able to get a decent estimate of the specifications. Ball bearings are slightly louder but will last longer, especially when blowing up/down or when exposed to high temperatures, than sleeve bearings.

I find it hard to believe that a 120mm fan that 'sounds like a jet engine' runs at only 1900 rpm. In general I would not expect miracles from so-called silent fans without reduction in airflow with some static pressure. It may be that the bearing have just run dry. You could try putting a drop of oil in there to see if that makes a difference.

[carl_lab]

I sent a request to EBM Papst (asking for the original W2S107-AA15-16 115V~ or low noise replacement).
They offered me a 4606Z (180 m³/h free airflow), that is supposed as the successor.
As they couldn't provide me the datasheet of the old fan, I'm not sure if that suggestion is based on comparison of the datasheets of the old and the "new" fan - maybe they just choose the strongest one... 



Unfortunatelly the construction looks nearly identical as the original EBM, no visable signs of aerodynamic improvements (especial wingtips etc.), so I do not really expect lower noise level (45dB).


They also sent a datasheet with airflow diagram, so hopefully I (we) could find an equivalent fan from another maker, that is "low noise"...



Note: Data of the 115V types applies @60Hz, but mains frequency in Germany is only 50Hz.
I'm not quite sure how to recalculate airflow exactly, but I roughly calculate 50/60 of airflow @60Hz, so I'll expect about 150 m³/h free airflow.

Any suggestions regarding manufacturers of low noise AC fans?

[quarks]

I have not yet replaced a fan in my 4808, but I already bought a replacement from Conrad

https://www.conrad.de/de/axialluefter-230-vac-180-mh-l-x-b-x-h-120-x-120-x-38-mm-28fr300-538108.html

and installed some in other gear. So far I like it, because of good build quality and it is quiet compared to the original.

If you want to buy really cheap, have a look here:

https://www.voelkner.de/products/71056/Luefter-Tm-220.html

[HighVoltage]

quarks:
The Conrad suggestion would not work, since he needs the 115V version

carl_lab:
Why not trying the suggested 4606Z and then measure the airflow and noise in comparison to the old one?

I looked in my cabinet of 115V fans but this size is not in there.

[carl_lab]

I think 230V~ fan will work if I change the circuit:

Original: 115V~ fan is connected parallel to one of the 115V~ primary windings of transformer.
New: 230V~ fan is connected parallel to both 115V~ primary transformer windings in serial.

Why not trying the suggested 4606Z and then measure the airflow and noise in comparison to the old one?

I thought about doing so, but just to estimate the airflow it's not exactly what I call low-cost.
It don't need to be low-cost, but I don't expect the new Papst to be significant more silent than the old original.

I could also buy another (low noise) fan and compare airflow or internal temperature.

@quarks:
32dB of the "Conrad" fan sounds promising... 
 

[alm]

Where do you see 32 dB(A)? The only 230V models in the datasheet provided by Conrad with 180 m³/h (105 CFM) air flow are the 4E-230S or 4E-230B. Both are specified as 50 dB(A) at 230 V. Slightly worse than the Papst. Also keep in mind that the 180 m³/h is at 60 Hz. At 50 Hz, the airflow is only ~95 CFM (160 m³/h) according to the curves in the datasheet.

At such a high airflow rate turbulence will likely to be the dominant factor. Bearing noise will be negligible (so you might as well go for ball bearing for reliability). Even if you had a magic fan blade design that would somehow eliminate all turbulence (good luck with that), you would still hear the noise from the turbulence caused by the case and circuit boards.

In my opinion the only way to substantially reduce the noise is to reduce airflow. How much impact this has on reliability, accuracy and drift is something you will have to test. It might be easier to use a DC fan of similar size and power and reduce its voltage until you find something that strikes a reasonable balance between performance and amount of noise you will accept, and then buy an AC fan with a similar design and rotational speed.

[carl_lab]

Where do you see 32 dB(A)?

That's what I found on the Conrad website:



But you're right, that sounds too optimistic.

The only 230V models in the datasheet provided by Conrad with 180 m³/h (105 CFM) air flow are the 4E-230S or 4E-230B. Both are specified as 50 dB(A) at 230 V. Slightly worse than the Papst. Also keep in mind that the 180 m³/h is at 60 Hz. At 50 Hz, the airflow is only ~95 CFM (160 m³/h) according to the curves in the datasheet.

I've had a look at that datasheet, but the fan (blades) and naming was completely different to that what you can see at the Conrad website. So maybe they mixed it up with another fan's datasheet.

I have not yet replaced a fan in my 4808, but I already bought a replacement from Conrad

https://www.conrad.de/de/axialluefter-230-vac-180-mh-l-x-b-x-h-120-x-120-x-38-mm-28fr300-538108.html

and installed some in other gear. So far I like it, because of good build quality and it is quiet compared to the original.

Of course I do not know the original he compared with...

[Armadillo]

It is well known industrial practice to measure the noise by dBA, dbB or dbC, range of frequency filter scales to simulate range of human perception. dBA being the more common.
Hence, [IMO] if a Seller will to even miss out the weighing factor [A,B,C] in their specification sheet, you can safely ignore the Seller.

[alm]

Ah, I missed the technical data on the Conrad site somehow and went straight to the datasheet (which is usually more reliable, since that is information provided directly by the manufacturer). I also noticed the different model numbers. I am not sure which fan you would receive if you ordered it. The certificate of conformity also specifies 4E-230S, by the way. But maybe Conrad have switched supplier without updating the documents?

But I find 180 m³/h at 1800 rpm, while both the Papst and the "Innovative Industrial Co., Ltd." datasheet that Conrad provide specify about 3000 rpm for the same air flow, very optimistic. My guess is that they copied the data from a lower speed fan, similar to the 4E-230S-17 from the datasheet they provide (1700 rpm, 55 CFM, 35 dB(A)).

[texaspyro]

At such a high airflow rate turbulence will likely to be the dominant factor. Bearing noise will be negligible (so you might as well go for ball bearing for reliability). Even if you had a magic fan blade design that would somehow eliminate all turbulence (good luck with that), you would still hear the noise from the turbulence caused by the case and circuit boards.

The noise generated by a turbulent air flow goes up at the EIGHTH POWER of the velocity.   It's one of the highest exponents you will find in nature.   

[Armadillo]

At such a high airflow rate turbulence will likely to be the dominant factor. Bearing noise will be negligible (so you might as well go for ball bearing for reliability). Even if you had a magic fan blade design that would somehow eliminate all turbulence (good luck with that), you would still hear the noise from the turbulence caused by the case and circuit boards.

The noise generated by a turbulent air flow goes up at the EIGHTH POWER of the velocity.   It's one of the highest exponents you will find in nature.

HVAC Fans air flow are turbulence in general but are way below the speed of sound.... We are not talking about Fighter Jet Engine here.
Though I agree that the acoustical enclosure plays a important part in the overall noise generation, but I cannot agree with the usage of the power of 8 here. I think you are talking about lighthill simplified approach but I would appreciate the reference materials you based and referred to.  Thanks in advance.

[TimInCanada]

The exact airflow of the fan may not be too critical.  The operating temperature specification for this unit is 0 - 50C.  The current fan is sized for operation at an ambient temperature of at least 50C, and probably more since the unit is designed for rack mounting. 

Also, check the airflow path inside the cabinet.  From this photo it looks like the fan may only circulate air in the rear half of the cabinet where the power electronics are.  It looks like there is a thick divider going across the cabinet:



If that's true, the fan airflow would have little or no effect on the precision components.

It would be an inexpensive test to put in a quieter fan and see how that works.  A thermocouple could be placed, for instance, near the voltage reference parts to measure what effect a quieter fan would have on local temperature. 

[Samogon]

Put obstacle in front of quietest fan and all extra buck go to waste.
In Datrons 4708 4808 fan blows into heatsinks assemblies. And i would like to have formula to estmate sound level, but afraid this can be done only by math sw.
But i am in to test new fan, really tired to hear that small jet in my lab.

[cdev]

I have an 115 v. 80 mm Mechatronics fan (UF80A12) that I use to suck away the fumes when I am soldering. it pushes a very large amount of air for its size and despite that is unusually quiet.

It has a sleeve bearing.

Ive used Mechatronics fans before and they are good quality, quiet and push a lot of air.

[carl_lab]

Ive used Mechatronics fans before and they are good quality, quiet and push a lot of air.

Their 120x120mm AC fans with airflow 180m³/h are about 46-47dB(A).
Papst is 45 dB(A).

http://www.mechatronics.com/products/search.php?type=ac_axial&v=115&cfm=100%3A200&mm=120

Has anybody seen an AC fan with visible aerodynamic improvements (wingtips, airflow break-away edges etc.?).
I've got a notion that all this is available to DC fans only (in most instances fans for PC cases)...

[quarks]

the fans I have bought, came with data information 180m³/h and 28-32dB(A)

[alm]

Are they also 1800 rpm like the Conrad spec? Do you have a model number or a link to a datasheet?

[quarks]

here is what is written on the box

[alm]

Does the label on the fan match the picture on the Conrad site (Artiko RAH1238S1)? Because according to the datasheet (top row is RAH1238S1) it is 2700-3200 rpm and 46-50 dBA (depending on supply voltage).

[quarks]

The label is as in the Conrad picture.

[alm]

Then I think we can assume the Conrad specs are wrong (I trust a manufacturer's datasheet over what Conrad claims). If you want, you could measure the rotation speed (having the fan tick against something light and looking at the audio spectrum might be the easiest). If the datasheet is correct, it should be much higher than the 1800 rpm claimed by Conrad.

[Samogon]

Lowest i can find on Digikey with relatively close CFM is 46dBA.
There is direct dependency of noise and CFM you cant trick nature in the same package size. I allow 10% max for blades shape, but not in magnitudes.

[quarks]

Just meassured 2600 RPM at 230V, so the data is unfortunately wrong and that is unaceptable.
Now I question also the 32dB(A), but still I remember the fans I replaced were significantly louder.

[carl_lab]

Does anybody know how to measure noise level (distance, angle, environment etc.)?
Maybe I can borrow a noise level meter.

[HighVoltage]

I use an iPhone app these days. Probably not so accurate but also not so bad for just comparison purpose.

[Samogon]

Just meassured 2600 RPM at 230V, so the data is unfortunately wrong and that is unaceptable.
Now I question also the 32dB(A), but still I remember the fans I replaced were significantly louder.

Ok so i can order 46 dBA

[alm]

Does anybody know how to measure noise level (distance, angle, environment etc.)?
Maybe I can borrow a noise level meter.

For an accurate measurement you should be measuring at a decent distance (1m is the standard I think). That makes it tricky to measure low speed (noise) fans without special equipment. Ideal environment is an anechoic chamber. Look at what these guys have done.

I do not think that angle is particularly critical at that distance. Ideal is probably something like 30-45° off axis. This document contains some relevant information.

[Armadillo]

Wonder why you guys are even bothered with unqualified noise figure of unqualified flow at free air delivery especially when the installation is not going to be free air delivery?
I can brochure a page saying 1 dbA at 1800 CHM but if upon questioning I tell you that I measured it 100m away, OK.   

Even if the figures of noise at FAD are sworn to be in compliance to certain international standards of compliance in testing, I don't find it of any use here unless you going to free blow the small little fan.

Cheers!

[alm]

Because the manufacturer is not likely to have measured the noise of their fan in a Datron 4808, so free air is the only figure you have in the datasheet? How else would you compare fans without buying multiple and testing yourself?

[Armadillo]

Because the manufacturer is not likely to have measured the noise of their fan in a Datron 4808, so free air is the only figure you have in the datasheet? How else would you compare fans without buying multiple and testing yourself?

Well, Datron did and also considered the pressure losses and induced noises amongst other engineering aspect in their selection.
But admittedly, new discovery is also a good contribution to the community and thanks for the efforts to be contemporary over the instrument manufacturer.
Do share your discoveries.     

[carl_lab]

Then I think we can assume the Conrad specs are wrong (I trust a manufacturer's datasheet over what Conrad claims).

Finally I got the requested datasheet from Conrad and Voelkner:

[Samogon]

That is more realistic parameters
Last night i have installed
www.digikey.com/product-detail/en/UF12A12-BTLR/1570-1015-ND
It is bit quieter will measure temperature inside unit

Well it is much quieter:
With original fan i had 90 dB(A)
With new 62 dB(A)
Definitely airflow twice lower, will monitor temperature if it is sufficint and DATRON engineers were really cautious

[Samogon]

After 30 minutes run with 100VDC output enabled i have 35C in the heatsinks
With ambient temperature 23.5C
Intake 24C
Exhaust 32C

[carl_lab]

Can you post some photo where your temp sensor was located?

[Samogon]

Plunged right into fins of heatsink.

[carl_lab]

Plunged right into fins of heatsink.

I wasn't fearless enough to do so ("DANGER High Voltage" sticker on the beige plastic shield), so I placed it between heatsink of HV block and connectors of the first daughter board:


Picture shows only location (montage of your pic).

Using original fan I measured 33°C after a few hours at ambient temperature of 23°C.

I'll order the fans from Conrad/Voelkner next week.

BTW: what's the 50V/400V switch for?

[Samogon]

Great news! So we are only 2-3 degree apart. I think we should be good.
50/400 swith enables high voltage option, so user cant go higher than 100Volts if remember correctly.

[carl_lab]

I was just wondering about the 400V notation because calibrator spits out 1000VDC.