question

Bobby avatar image
Bobby asked

Victron Multiplus II 48/5000 (noisy) Fan Replacement

Hi all, as per the topic, I own Multiplus II 48/5000/70, with MPPT and a 48V battery bank.

Things are working as expected, so no problems with configuration.

The Fan also switches on when having a load of around above 2kw for more then 30-40 seconds (and I also believe that that is in order). However - I am not confirmable with the noise of the actual fan - the thing is loud and I do know that there are many fans on the market that have different bearings in order to lower the noise.

Can anyone advise if the fan in question is 24V DC or 12V DC and confirm the dimensions (if I'm not mistaken, it should be 120x120x38mm)?

Thanks in advance

Bobby

Multiplus-II
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

17 Answers
Bobby avatar image
Bobby answered ·

Ok, so I found a possible replacement(s)...I understand that it is less thick, but that is not that important as diameter and flow rate...
Noise however is significantly less.

Only thing I'm not sure is which exactly fan is inside of the Multi II ?


The replacements (Noctua Industrial 120mm)

Also, the connector PWM has four contacts but it can be modified to two only so that won't be an issue

2 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Hi

You need 110CFM or 186 m3/h flow rate

Model in Multi is a JF1238B2HR-R

24v 110CFM 2800 RPM dBA 43.50

They have s similar sound rating. all depends on the fan blade design if you want quieter fan noise. slower rpm as well with the correct blade design.

There are many other known fan brand names out there.

0 Likes 0 ·

Thats exactly the brand i posted the screenshot for...

Not sure why it didn't show

Based on the db rating, they would generate the same noise level :(

I wonder it the pitch of the noise would be different

0 Likes 0 ·
Pierre-Marie Baty avatar image
Pierre-Marie Baty answered ·

I'm resurrecting this old thread for other readers to share my experience.

I have a Multiplus-II GX and I too find that fan noise absolutely excessive. I have applied stenis' advice and had taped bitumen mat on the inside of the housing cover, to find out that it had only improved the problem marginally -- i.e. when you're in front on the device. The situation is not improved when, as in my case, the enclosure is hung on the wall (a *concrete* wall) that separates the house from the parking garage : it can be heard from everywhere in the house !

I too have searched for fans with similar characteristics and found none that would be allegedly quieter, at least on their datasheet.

The vibrations induced by the fan activity are just absurd ; even its neighbour the Fronius inverter saturated with solar power generates less than half of that noise.

Moreover, the problem has gotten significantly worse since I upgraded my MPII firmware from v470 to v490; with the old firmware, the fan would only kick in when the device was sending to or drawing power from the battery and stay quiet in passthrough mode - now it spins even in AC passthrough mode, when all the power on the output comes from the grid input ! This is ridiculous and infuriating.

In my opinion the last thing that would work, apart from moving the enclosure far away from the house which is unacceptable, would be to create a forced air duct, well fitted to the enclosure's air intakes, long enough to breathe air from the outside, and deport the fan all the way there.

I see no other solution, apart from enabling that "silence fan" configuration assistant which has the big drawback of reducing the inverter power.

This annoyance is, IMO, by far the #1 negative point of these devices, and if I had known they would be so noisy I probably wouldn't have bought one. Buyers beware.

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Rob Duthie avatar image
Rob Duthie answered ·

Hi

Their fans are noisy in those units even from day 1, yes there are 120mmx38 and are 24v.

It's to do with the blade design etc to have a low noise fan, this fan might be middle of the road type? As i side note this is the company that has made capacitors for years, not fans.

They are ball bearing type. Brand Jamicon data sheet attached.JF1238-14.pdf

Regards

Rob D

NZ


jf1238-14.pdf (388.5 KiB)
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Bobby avatar image
Bobby answered ·

I will try and find some replacement with lower noise (db) and attempt to change it...
The noise of the factory fan is just too annoying and loud

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Deon avatar image
Deon answered ·

I have an EasySolar 48V 5000VA and it is extremely noisy

With an uncalibrated phone based noise meter, my Easysolar measurements are:

1 - Silence - 20dB

2 - My workspace - 50dB (EasySolar fans at full speed)
5 meters from
the Easysolar

3 - 1m from blue encl. - 60dB (fans at full speed)
of the EasySolar

4 - Outside the encl. - 80dB (fans at full speed)
at hot air outlet


The current fan JF1238B2HR-R, gives a spec of 43.5dB, which means the position of fans in the blue enclosure is the noise generator. I have not opened the enclosure to understand the noise sources.

I need some advice - do I replace these fans, and will it affect the warranty?

The alternative is to have the Enclosure moved, and the associated costs and additional wiring.




2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Nickus de Vos avatar image
Nickus de Vos answered ·

Not specific to Victron, but in general the perceived fan noise of equipment is mostly contributed by vibrations threw the chassis rather than the fan itself. I'll put some good money on those fans being a lot quieter if taken out of the inverters and measured in free space. Often you can get rubber mounting systems for fans which helps a lot, pads, washers, screws etc. A good fan manufacturer like Noctua should sell some of these accessories. Best I think will be to buy a bunch of accessories with a good replacement fan then open up the inverter and see what you can do in mounting the replacement fan with some rubber accessories. Often something simple like rubber mounting screws in place of metal or plastic can make a big difference, or just ensuring that the mounting screw doesn't touch any other part like the chassis or cover.

1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

I tape a bitumen mat on the inside of the housing cover, this reduce the vibration quite a lot - easy and cheap solution :)

0 Likes 0 ·
popnic avatar image
popnic answered ·

@Bobby I had a chance to compare different fans with same voltage, amperage, and size. The Jamicon model - thank you Rob for sharing - that is in MP2 was the one of the most powerful form air flow perspective. From noise perspective you can hear mainly the air flow from the Jamicon fan when it is supplied with DC or high freq PWM from 20 to 24V.

Unfortunately MP2 need a hi flow fan to cool down especially the transformer and also the MOSFET's and choke due to "redesign".

I visit a showroom to discuss the fan noise issue. @Deon similar measurement results. There one unit MP2 was with fan on and it was also loud. I convince a technician to remove the front panel and he did it. We were surprised about the difference. With the front cover off the noise was way less but when he put the cover back, the noise comes back at high level again.

If you change the fan i think that you will loose the warranty but if you still want to do it, in my opinion you have to address to:

- noise source - fan itself - change with a better one with same flow rate;

- fan speed control hardware - low frequency PWM - add a basic RC filter;

- mounting bracket - transmit vibration to the housing - add a damping material under it and between the screws and housing @Nickus de Vos Victron should consider your remarks.

- housing - being half empty it resonate like an acoustic guitar - here you have to be careful because also the housing has an important contribution to heat transfer. @stenis good ideea.

Having some experience in NVH from one of my previous job I can say that this kind of issues are not so easy to solve and has to be considered in the design phase which, unfortunately was not the case.

Good luck!

19 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

I don't understand why Victron don't use a continuously-variable-speed fan (temperature-controlled) like in PC coolers instead of simple on/off -- this wouldn't make any difference at high loads (say, over 60% of maximum power) but could make the units much quieter at medium loads (say, 20%-60%) where they spend most of their life. It's only the last part of the fan speed curve where the noise goes up rapidly, at medium speeds they're *much* quieter.


Any comment from Victron about why this is, and whether it could be added in future via a firmware update?


0 Likes 0 ·

Hi iand

The fan speed is controlled depending on the temperture of the transformer and switching electronics.

One note when at lower speeds they are more noisy due the PWM controller.

Regards

Rob D

NZ


0 Likes 0 ·
Rob, I fitted Noctua PWM fans to cool my batteries. Tested before installing. They're completely inaudible until about 60%. So although older ones may be noisy at low speed, good ones aren't.
0 Likes 0 ·
Rob Duthie avatar image Rob Duthie kevgermany ♦♦ ·

Hi Kev

You say you fitted a Noctua PWM controlled fan in the multiplus and it is a lot quieter the orginal ones? what about airflow rating compared to the original ones. I see that they only do 24v ones in a industrial version. and very simalar air flow rates.

They are half the db noise rating of the Jamicon ones and also have included damping etc.

The fans that are mounted in the Multiplus2 are hard fixed with no damping to the frame to keep vibrations and resonates down. That is what part of the issues appears to be. Just done a quick measure of the PWM when fans is running slow the PWM is 50Hz at 8v DC and has quite a virbration to it, after fitting a crude capacitor filter across the motor connection it does make it less noisy but not by much, it sounds like the fan motor dose'nt like running at the lower speeds not realy designed for PWM applications signals, The fan like the Noctua is better suited to this application of PWM control.

I have one on order here in NZ to replace and report back on my findings

Regards

Rob D

NZ

0 Likes 0 ·

@Rob Duthie I did an experiment on a testbench with Jamicon fan. I played with PWM frequency from 30 to 50Hz as you mentioned and different duty cycle at 24V . I got good results with a 10W 33Ohm series and 1000uF in parallel on the fan side.

I choose 33Ohm from 27, 33 and 47 ohm as a compromise for rpm of the fan and less noise.

27 ohm the PWM is noticeable. With 47 ohm the fan is running smooth but rpm is too low.

I do not want to open my Multiplus II because the option to send it back is still available.

Ron, can you experiment this for me in real conditions?

filter.jpg

Thank you.

0 Likes 0 ·
filter.jpg (15.2 KiB)

Nope, as coolers to shift some air over my batteries. I designed the compartment very badly and there's no convection.

The 50Hz explains the noise differences. I didn't try running the Noctua's at low frequency. Just decided to stay with their specs. Apparantly Intel said 30KHz to keep PWM noise inaudible. And that's the PC standard.

What I also noticed was that the smaller fans make fan noise earlier than the bigger ones - if my English makes sense.


0 Likes 0 ·
But I must add that it's important to use PWM at the PC standard of 30KHz.
0 Likes 0 ·
Rob Duthie avatar image Rob Duthie kevgermany ♦♦ ·
Hi Kev

Yes i don't think Victron are running at the higher frequency rate thats why they are noisy at low fans speeds this i have noticed on a few installs i have done with multiplus2 inverters to date. even the older modles the multigrid models as well.

So i have on order a Nocua fan to replace it in my unit to see what the difference it makes. as it has rubber damping mounts compared to the Jamicon brand which is hard mounted to chassis with no damping at all, so it will resonate in the enclosure which interns makes it sound very loud when running at lower speed so i will use my inverter personal as a dummy test unit, as i have a complete system setup with Vicrton gear etc to test these issues in real time applications in my workshop which is connected to the house.

Regards

Rob D

NZ

0 Likes 0 ·

Hi,

Which one did you order?

This one is looking really good:
Noctua nf-f12-industrialppc-24v-3000-q100-ip67-pwm

br,
Andy

0 Likes 0 ·
Hi

Thats the one.

0 Likes 0 ·

Please let us know how you get on.

Some time ago I posed a question to @Johannes Boonstra (Victron Energy Staff) in his YouTube tech talks about fitting a quieter fan, his reply was the fairly predictable yes, but at your own risk and I've a feeling he also said it would invalidate the warranty. Obviously not a problem for you, but if anyone's reading and decides to try, be aware. I haven't taken it further yet. Might do. If the standard fan is low frequency PWM, I guess I could rig up an Arduino to convert that to 30KHz. But mine (compact in an easy solar) is quiet enough unless there's a big system load.

0 Likes 0 ·
Rob Duthie avatar image Rob Duthie kevgermany ♦♦ ·
Hi Kev

The Noctua fan won't realy start until the PWM is at higher output. I used a fan from a Dell server as well which are realy quite fans, the PWM comming out of Victron is crude to say the least, This made the DELL fan noisy as well, by the PWM pulses. On a normal varible regulated PSU super quite even the standard fan was very quite but soon as you add the Victron PWM noisy as.

I have added a 220uf cap across the motor output which has smoothed out the PWM and fan doesnt resonates no where as much. Also if i rubber mount the fan it will be even more quiter.

I do have a PWM generator i can change the Freq and Duty cycle which runs motors on to see were the sweet spot is. What we need is PWM output converter to voltage output filter.

Regards

Rob D

NZ

0 Likes 0 ·
Thanks Rob, good try, explains a lot.
0 Likes 0 ·
Very interesting. I'd be very curious to see what that "crude" PWM signal out of the Multiplus mainboard looks like on an oscilloscope, measured in time and frequency (FFT). Once we know what it really looks like, I bet something can be done to improve the problem !
0 Likes 0 ·
Rob Duthie avatar image Rob Duthie Pierre-Marie Baty ·
Hi

Around about 50Hz when running at slow speed or about 8v.

0 Likes 0 ·
I was more thinking in terms of signal cleanliness and harmonics but thank you nonetheless.
0 Likes 0 ·
Rob Duthie avatar image Rob Duthie Pierre-Marie Baty ·
The PWM is a clean signal, just the motor can't handle the low freq PWM signal, its to low so it makes the motor pulse and be noisy, needs to be set at a higher freq rate to smooth out the pulses. i am not sure if it controller by a micro or not if it is, a software firmware update will fix it?
0 Likes 0 ·

OK so if the signal is clean and the frequency is unsuitable for the motor, and considering the motor is wired with 2 wires (which means it should be able to be driven by a continuous signal), then the lowpass filter (popnic's method) is the best that we can do. That will effectively transform the PWM square signal into a quasi-continuous signal ranging from 0V (PWM = 0%) to Vcc (PWM = 100%)

0 Likes 0 ·
Show more comments
baxter avatar image
baxter answered ·

Hi,


up to know I thougth the fan is just voltage controlled, has it has only to wires (+ and ground).

Are there new multiplus II with a pwm controlled fan?


9 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Hi

The multis have all being PWM controlled 2 wire fans, since day one, As i have owned a the older version and the newest versionof multis and all have the same noise issue at lower speeds etc.

Regards

Rob D

NZ

0 Likes 0 ·

For a PWM controlled fan 4 pins are required, at least for the Noctua Fans (and all fans I use in the PC settings).


So how to connect a 4pin Noctua PWM fan to a 2pin Multi? I guess there will be no PWM signal and the fan is only voltage controlled.


1651069850302.png

https://noctua.at/pub/media/wysiwyg/Noctua_PWM_specifications_white_paper.pdf


0 Likes 0 ·
1651069850302.png (23.1 KiB)

Actually it seems they used an electronic trick, where the Jamicon fan is driven analogically with a PWM square signal that is lowpass-filtered by the fan hardware itself into a quasi-DC signal between 0 (when the PWM is at 0% duty cycle) and 24V (when the PWM is maxed out). The Multiplus fan is not really a PWM-driven fact in this regard, just an *analog* fan that's driven with a PWM circuit, instead of a DAC - probably because it's cheaper.


If you want to plug your Noctua, you'll need to use the blue, yellow and black wires. Tap a 24V DC somewhere from the mainboard to the yellow wire, match the black wire (ground) with the black wire, and send the red wire (PWM) through a dividor bridge that will lower the 24V PWM square signal from the Multiplus into a 5V PWM one compatible with your fan. That's how I'd do it.

*edit* sorry for my bad english ; by "dividor bridge" I meant this : https://en.wikipedia.org/wiki/Voltage_divider

0 Likes 0 ·
baxter avatar image baxter Pierre-Marie Baty ·
Thanks for explaining this.

How weird is that

0 Likes 0 ·

Actually it's a rather common trick.

Choose your resistor values wisely: to bring down 24V PWM to ~5V, you need Z1 to be 4 times the value of Z2 (ex: Z1 = 4k ohm, Z2 = 1k ohm).

0 Likes 0 ·
Rob Duthie avatar image Rob Duthie Pierre-Marie Baty ·

Hi

That will work Ok just trying to avoid any soldering on the PCB and make it a plug in play type fix.

I am trying different T filter to see how to smooth out the pulses to the motor. So far not the best solution, just a 100 to 220uf cap makes the most difference but the fan runs faster of course.

I have used a PWM generator test rig to run a Noctua fan at the lower 1KHz PWM runs very smooth and is better at the 25kHz range better linear control down lower At the lower end of my PWM tester it won't go that low to test at the same as the Multiplus, need to change to re test at the 25Hz range same as the Multiplus.

Regards

Rob D

NZ

0 Likes 0 ·
The fourth wire is PWM in. Three wire fans omit this and simply feed the actual fan RPM back. Speed control is by voltage from the controller, often/usually in the form described by Rob - i.e. a PWM signal 'smoothed' to give a lower voltage overall. Hence they were called PWM, but.... The fan didn't read the PWM signal like the 4 wire fans do.
0 Likes 0 ·
baxter avatar image baxter kevgermany ♦♦ ·

Will the Noctua Fan work with just 2wires like the original one (only +24V and gnd) ? Based on the info I got from Noctua support the fan also supports voltage control without PWM Signal in.

0 Likes 0 ·
Hi Yes it does.
0 Likes 0 ·
iand avatar image
iand answered ·

I still don't understand why Victron do such a lousy job of fan speed control -- every single PC on the planet has proper HF PWM control to give quiet variable-speed fans and the circuits to do this cost next to nothing, so why on earth don't Victron do the same? And mount the fans on little silicone feet like Noctua do, the cost is pennies but it makes a big difference to noise from the case...

2 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

I'm not trying to justify anything but from a different perspective:

Current system works and is reliable and well proven

Many installations are in places where noise isn't a problem

More consistency in parts for maintenance

Other development priorities

Perhaps high frequency PWM wasn't as common when this was designed

Change is high risk-get it wrong and a huge number of units will give failures in the field.

All fans, even the super quiet ones, develop a lot of wind noise at higher speeds. The alternative is to fit much bigger fans at the cost of bigger units.


Having said that, I'd love the unit in my camper to have a much quieter fan. At all speeds.




0 Likes 0 ·
iand avatar image iand kevgermany ♦♦ ·
HF PWM has been ubiquitous in all PCs for many years, the risk of any problems are probably lower than the existing solution, cost adder is close to zero.

The cost adder for bigger fans is negligible in the context of overall product cost -- if you want to be quieter and can fit them in they're quieter, but the product has to be designed for them -- which is too late for MP II and Quattro II.


0 Likes 0 ·
Pierre-Marie Baty avatar image
Pierre-Marie Baty answered ·

Well, it seems clear to me that vibration dampening and keeping sound levels acceptable were
*never* considered at all as possible issues by the designers of these products. It's to a point that many design choices that have been made were in fact the worst ones in this regard.

Now to recover from here... Apart from the forced air duct, or redesigning the enclosure completely and retrofitting the boards into a carefully hand-crafted one, I don't even think there's a proper solution.

I'm still watching this thread in the hope of a miraculous idea, but I'm slowly pushed into thinking it's a waste of time. Currently I live with it.

1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

I contact Victron about fan noise issue. They come back to me thru distributors technical representatives in a day. They were fast and they listen me but we come to the conclusion that the unit is not faulty so the service is not necessary.


We are only a dozen of "sensitive" people here but remember:

Victron is powered by Multiplus II's and this is why they cannot hear us.

Personally I will not change anything in my Multi. It is a good product but is not for me. I lost the opportunity to send it back so I will keep it while I am looking for another inverter/charger.

Sorry to bothering you guys.

0 Likes 0 ·
Rob Duthie avatar image
Rob Duthie answered ·

Hi Bobby


I have modified a my own Mulitiplus2 unit with a Sunon fan 25mmx120mm model # EEC0252B1-0000-A99 with rubber mounts and with 47uf 63v cap across the fan terminals. And results are, the bad droning noise or PWM noise due the low PWM frequency you normaly heard has gone.

Much quieter unit now, no droning like before. just air flow noise though the unit now which is much more pleasant to listern to.

The air flow is the same as the orignal Jaicon fan, the 25mm thick fan allows for the rubber mounts to used on the original mounting bracket.

If Victron could just change the PWM to a higher frequency rate the problem would go away , and they would have a much quieter inverter on the market.

Regards
Rob D

NZ


fan-mod2.jpg

fan-mod3.jpg

fan-mod1.jpg


fan-mod1.jpg (2.5 MiB)
5 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Excellent ! I'll try this model just like you did as soon as I have time to open that Multiplus-II again. And thank you for the pictures. Out of curiosity, what does the dB-meter say compared to the factory setup ?
0 Likes 0 ·
Rob Duthie avatar image Rob Duthie Pierre-Marie Baty ·

Hi

Didnt not have a dB meter, just it is much quieter than before no droning noise.

0 Likes 0 ·
Thanks Rob for getting to the bottom of the fan noise issue and providing some solutions, I have been wondering why my Multiplus 2 fan noise is like no other piece of equipment I've come across in my 40 years in electronics.

It would be good if Victron could bring out modification kits (hardware and software) to sort this out once and for all.

I'm not sure your noise isolation mounting is as effective as it could possibly be? Is the bolt and nut fully isolated?



0 Likes 0 ·

Yes it is, the rubber damping is the isolation. it way better than the stock fan setup. It boils down to Victron to just change the PWM pulse to much higher frequecy rate and problem goes away, simple as that, i have tested with a higher rate and is much quiter on the standard fan. As there PWM pulse rate is only 25Hz too slow thats is why the fan drones.

Regards

Rob D

NZ


0 Likes 0 ·

I can also confirm that the recommended SUNON EEC0252B1-A99 (EEC0252B1-000U-A99) fan is much more quieter than the original fans of the Multiplus-II.
The fan is available on the SOSElectronic website in EU and is cheap.

I used a 220 uF capacitor because at 47 and 100 uF I could stil hear some "vibrations" from the PWM at low speeds.

0 Likes 0 ·
ecoman avatar image
ecoman answered ·

Hi folks, first time posting here. I also belong to the surprised, noisy Multiplus II owners club, mine is the biggest 10kW version. I almost fell off my chair when I heard the fans starting the first time. That was under a light load of about tad over 2kW which one would think to be easy for a 10kW unit (well, maybe 8kW instead of 10kW name tag) but anyway, only a fraction of max. The unit is brand new, and naturally Victron will not consider faulty, so no warranty claims to be done. As such, I am considering the same modification as you have done - but one thing caught my eye in the SUNON EEC0252B1-A99 (EEC0252B1-000U-A99) fan mentioned in this thread. The fan spec says the following, and I would like to hear how do you read and understand it.

On page 11 of the spec, they tell not to use PWM control with it:


1653322061248.png


Ie, the way I read it is that the fan has only two pins, power and ground, and neither of those should be used with PWM signal. Wonder if it means the fan will die prematurely without warning, or smth else? Should we find another fan which is ok for PWM control?

Thoughts?


1653322061248.png (25.3 KiB)
4 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Yep, that's why the modders use a capacitor+resistor setup to create what's called a "lowpass filter" that effectively transforms the square wave PWM signal into a continuous DC signal that's harmless for that type of fan. If you do exactly what Rob Duthie did, it'll be okay.

I don't think you should take the datasheet advice literally (that the fan speed can't be adjusted). That fan is just a DC brushless motor, and I fail to see how driving it with a variable DC voltage to control its speed would harm it as long as its nominal voltage isn't exceeded. I might be mistaken here though (for example, that fan would be a stepper motor, which would require some complicated electronics driver card, but I'm yet to see a fan like that).

0 Likes 0 ·
Hi

Well the Jamicon fan is a very similar fan design so will have the same issue as it has a bushless design as well, The noise it makes suggest that can't be good for it either. At least the Sunon has better data sheets than the jamicon fan.

Well to date the sunon works a treat and is way quiter.

All Victron has to is change the PWM rate and all will be well.?

Regards

Rob D

NZ

0 Likes 0 ·
Thanks for quick responses folks, I share your views on filtered DC being certainly being more gentle on the fan than Victron design bare PWM square wave at 25Hz making the fan cry out loud. Adding a capacitor will result in higher average voltage and thus higher fan speed (just more cooling than designed :o) ). Reading your text (Rob), you mention only the capacitor across terminals, no resistor. Is that how you ended up setting it, or did you actually add the resistor also (not shown in pics). A resistor could be used to keep the fan speed the same as without RC circuit, yet top revs would be lower compared to original setup when PWM at 100% - so I guess your setup is just fine. I'm just curious on whether you used a resistor or not, and how do you see pros and cons of it?


Thanks.

0 Likes 0 ·
Hi

Not used a resistor as it did not make much of a difference, but only used a 47uF value cap only a made a slight increase in speed, but very much quiter overall. compared to before. The rubber mounting improved it much better again.

Regards

Rob D

NZ

0 Likes 0 ·
sharpener avatar image
sharpener answered ·

Perhaps it would be good to add a series resistor anyway as @popnic suggests, so that the source of the PWM waveform (whatever it is) doesn't drive directly into a capacitive load. Depending on the source impedance, that could result in very high peak currents which would not be good for either the driving circuitry or indeed the capacitor. Like @RayMiller I too was surprised by the very "agricultural" fan noise from my new MP II, for its size it is certainly the noisiest fan I have encountered in a long career in electronics.

2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

ecoman avatar image
ecoman answered ·

Hi Rob, all, and thanks for sharing your implementation, I'll be doing the same mods once back at my remote facility. I will also take my oscilloscope with me to see the signal waveforms and mean voltages before and after changes and can post my findings here.

BTW, does anyone know if the noise originates from just one fan or both in the Multiplus II 10kW model? (would expect Victron to run both fans with the same low freq pwm, though...). Do the smaller models have two fans like the 10kW one, or just one?

Looking at the fan specs, it shows 24V, 207mA, which works out to be the 5W. Thinking of the serial resistor and the impact of it on the power the fan receives, I did a rough math below. This is lacking the effect of the capacitor boosting the speed, and the current value is an estimation, so not exactly accurate but should provide some clues on how the serial resistor size impacts on fan output.

Voltage resistor resistor pwr max fan voltage fan current fan power % of nominal
24,0 V 0 ohm 0,0 W 24,0 V 0,207 A 5,0 W 100 %
24,0 V 5 ohm 0,2 W 23,0 V 0,198 A 4,6 W 92 %
24,0 V 10 ohm 0,4 W 22,1 V 0,191 A 4,2 W 85 %
24,0 V 15 ohm 0,5 W 21,3 V 0,183 A 3,9 W 78 %
24,0 V 20 ohm 0,6 W 20,5 V 0,177 A 3,6 W 73 %
24,0 V 25 ohm 0,7 W 19,7 V 0,170 A 3,4 W 68 %
24,0 V 30 ohm 0,8 W 19,1 V 0,164 A 3,1 W 63 %
24,0 V 35 ohm 0,9 W 18,4 V 0,159 A 2,9 W 59 %

Ie, it would seem that already a 33 ohm resistor would cause the fan to receive only ~18V instead of 24V, and reduced current, resulting in about 60% of the fan power compared to what the factory setup with 100% duty cycle 24V feed would provide. While the fan spec doesn't tell how much less airflow that would mean, I'd guess it is going to be significantly less, probably in direct correlation with fan power and that would discourage using resistor values that high.

As such, it would seem that the implementation by Rob with no resistor is the safe one from maintaining the same (in fact, better) airflow, although a small resistor (say up to 10 ohms) should not cause too dramatic power drop, especially considering the boost from the capacitor which might compensate enough to keep the fan power and thus airflow close to factory spec.

Also, if using only 10 ohms resistor, a 1W model would be fine, as the power dissipation on the resistor would never exceed 0.4W. The 1W rating would actually be ok all the way to 30+ ohms. Then, is there a value of using such a small resistor even to protect the driver circuitry from the change in load impedance...?

Thoughts?

1 comment
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Hi

The resistor will get hot and over time this will fail as i have seen in the past with other fan circuits. The capacitor by itself is simple and effective and no heat is generated.

I will be fitting another fan to a clients Multiplus2 unit which i am installing in case they complain about the noise, i will check the impeadance of the output fan ciruit to see what i can measure etc.

Regards

Rob D

NZ

0 Likes 0 ·
ecoman avatar image
ecoman answered ·

The new SUNON EEC0252B1-000U-A99 fans arrived, but no rubber mounts included, so need to find or create ones. Tested at home without PWM, just straight 24V corresponding to 100% pwm signal, with and without 10 ohm/3W serial resistor.

Without serial resistor and free airflow I get 24V and 192mA = 4.61W power consumption (100%). With airflow obstructed (fan against the table), current increases to 215mA and power to 5.2W

1654199027525.png

With serial resistor current drops to 170mA, power draw to 4.08W (89%); voltage over the fan to 22.1V, thus 3.76W for fan (82%), and 0.32W (7%) for resistor, which warms a bit but hardly noticeable with finger. With multimeter temperature probe I get a reading of 40C, correlating with finger reading. Based on above numbers, the actual resistor value is 11.2 ohms.

With obstructed airflow current increases to 190mA, power draw to 4.56W (99%) of which fan gets 4.16W (90%) and resistor 0.40W (9%). Resistor temp now 43C. The resistor was in standstill air, not in fan airflow. In MP2 the resistor would inevitably get some airflow, yet it is a warmer environment than +23C now. Tested also with the second fan, got the same readings - so these fans seem to be identical twins. (I ordered two fans as MP2 has two and I don't know if it is one or the other or both making noise - assuming worst case).

Based on quick testing I figure I could use a small serial resistor, lifetime should not be an issue due to minimal temperature rise. Will see if the resistor makes any difference in results, will decide once I get to try the alternatives. Surely without a serial resistor the circuit is most reliable as only one component to fail and no chance of cutting the circuit like failed resistor can cause; yet both have the same chance for shorted capacitor which can also happen.


1654199027525.png (27.2 KiB)
2 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

I used these rubber mounts from RS Comp # 126-3939 they come in a pack of 8

and put a pice of foam on the bracket as per my thread photos.

Regards

Rob D

NZ

0 Likes 0 ·

Thanks, the same seem to be available here also, will get those rather than re-inventing the wheel.

0 Likes 0 ·
ecoman avatar image
ecoman answered ·

An update, things seem different with the 10kW model of MP2 - and so far I have not succeeded in making the device quieter. First, the fans are different, and there are two of them. Second, the capacitor makes no difference on the noise the MP2 generates, I tried 47, 100, 220 and 330uF capacitors parallel to both fans with only marginal impact on noise.

Ths lower power fan is JF1238B2HR-R, 24V 0.32A = 7.7W fan, and the higher powered one is JF1238B2TR-R, 24V 0.60A = 14.4W fan, both JAMICON as in other models. However, the 5W fans I bought based on the information from smaller systems would not most likely not be adequate replacements, and would not be able to generate the same airflow and cooling.

On the signal waveforms, I didn't capture them at all power levels, but here some samples when the inverter was under a modest, bit over 2kW load. The first one is without any capacitors connected, standard Victron factory setup and showing just one of the fans voltage. To be noted that both fans were driven with exactly same kind of voltage and frequency PWM signal

1655978988324.png

The second oscilloscope screenshot shows the voltage of both, each with 47uF capacitor connected...

1655979428983.png

...and the last one with 330uF capacitor on fan power feed:

1655979562184.png

As you would expect with the largest capacitor, the peaks smoothen and the lowest voltage is no longer zero but a slope towards zero volts, until the next peak comes in. BTW, the scales in the oscilloscope were 5V vertical (in single graph displays, 10V in two channel displays) and 20ms horizontal division. As you can see, we get about 5 full phases in 8 intervals on the screen - ie. 8x20ms = 160ms for 5 pulses, which makes about 32 ms pulse interval, or about 31Hz frequency. I don't know yet if the PWM frequency changes along with the load, but at least in this case it seems to be slightly different than the 25Hz mentioned earlier, but maybe the same control circuitry with "approximately" 25Hz...

Just for the sake of it, here few more samples of the same signal with slightly different load - as you can see, there are varying number of peaks on top of the pulse, not clear if intended or just a by-product of the fan control circuit. Whatever the case, those seem to be 4 peaks in approx 13..14ms, or ~3.5ms per phase, or about 300Hz frequency. That's another nasty frequency to listen.

1655980262161.png

1655980212229.png

The next thing I'll try is to disconnect the Victron fan control completely, and drive the fans with plain DC power from a lab power supply to see if a plain DC voltage on the fan (mounted in the frame as is) would generate noise, and if so, which level. That should establish the reference which can be achieved with Victron standard fans. If still noisy, should see if a quieter fans could be found. If clearly less noise, then I'm thinking that it should be possible to convert this PWM to plain DC (0-100%) analog voltage. There are converters which change PWM to 0-100% voltage level, but the ones I've seen so far are all either 0-5V or 0-10V output, have not found any with 0-24V output. Yet it should be possible to convert to that range with a simple circuit.

Thoughts from your experiments and findings, and what do you think about changing PWM to just DC level 0-100%?


1655978988324.png (48.0 KiB)
1655979428983.png (61.8 KiB)
1655979562184.png (68.2 KiB)
1655980212229.png (48.1 KiB)
1655980262161.png (48.5 KiB)
5 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

By just putting in the rubber mounts and changing out to the fan to a Sunon fan the reasonants goes away much quiter inverter allround i have being running this for about 4 months now with no issues and piece of mind. And just a 47uF 63v cap across the motor wires. just try this nothing to loose, dont get to hung up on the PWM issues this won't go away until Victron acknowledge the have a poor PWM control for there fan cooling very simple.

Sunon EEC0252B1-0000-A99

I used these rubber mounts from RS Comp # 126-3939 they come in a pack of 8

and put a piece of foam on the bracket as per my thread photos.

0 Likes 0 ·
Yes, I've got the same rubber mounts in my hands right now, the problem is that for the MP2 10kW model they cannot be changed without unmounting the whole unit and undoing the screws in the backside of the unit. Given the weight of 50kg and tight space, it is no easy feat to get it off the wall, but doable if must. The other problem is that the new 5W fans I have are only 1/3rd of the power of the original and cannot push as much air against the counter pressure, so even if changing the fans I need to find ones which can keep up the same airflow in the same chassis. I now have a DC power supply on site, so I can check if plain DC feed will run the fans quietly.
0 Likes 0 ·
Thanks for all your work Rob, I was slow to move, looks like the fans are in hot demand and not in stock in most places I have access to in Australia, RS supply is August and Mouser 2023.


I also note the Sunon fan bearing is magnetic offering lower vibration transmission, ideal for Victron Inverters.

All we need now is to be able to change the fan by just removing the top cover, as opposed to disconnecting the inverter and lowing 10's kg all the wall.


0 Likes 0 ·
Hi Ray

The fans are avaible In NZ from my supplier they are in stock. HiQ Electronics.

I have done a couple now for my clients to keep them happy.

Yes i prepare them before installing them, yes you have to remove the mount to mount the new fans on to,only done the 5Kw models to date.

You could do from the top just need a right angle screw driver to remove the old fan.

Does the 10 kW model use the same Jaimcon fan?

Regards

Rob D

NZ

0 Likes 0 ·
ecoman avatar image ecoman Rob Duthie ·

While the fans in the 10kW MP2 are also Jamicon, unfortunately they are not the same 5W fans as in smaller MP2 models. The fans in the 10kW model are much more power hungry, 7.2W and 14.4W according to the specs posted earlier (although the math gives 24x0.32A=7.7W instead of spec'd 7.2W). Below the model numbers

JF1238B2HR-R, 24V 0.32A = 7.7W fan

JF1238B2TR-R, 24V 0.60A = 14.4W fan

0 Likes 0 ·
ecoman avatar image
ecoman answered ·

Googling for the specs of those fans, this is what I found:

1655983906984.png

1655983729318.png

Which says the bigger fan alone produces almost 53dB at full speed, while the other is about 10dB quieter.


1655983729318.png (115.9 KiB)
1655983906984.png (137.5 KiB)
3 comments
2 |3000

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Wow, a 4000rpm fan is going to make a *lot* of noise in the MP II 48/10000 -- and presumably the Quattro II 48/10000 when it emerges. Which is a pity, because that's what I was intending to use... :-(

Does anyone know what the fans in the original Quattro 48/10000 are -- are they quieter?

Ecoman, how did you find out which fans the MP II 10kVa uses?


0 Likes 0 ·

Simply by opening the box and looking :o). Ie, I bought the MP2 10kW model without any clue about the loud noise it makes, only to discover it once installed. My previous inverter (Studer HPC 4.4kW) was silent, and it never crossed my mind an inverter with 2x the output would make such loud noise.

Here the pictures from inside the MP2 10kW model, you can see the lower rated fan on the bottom left side, and the more powerful one further up on the right side of the unit. On the top there is a massive toroid transformer which makes up most of the unit weight (49kg).

img-20220611-185118.jpg


Photos of the fans below. Those cannot be replaced without taking the unit off the wall, as it seems the mounting brackets are screwed through the back panel. Will update if I find another way to do it or find a way to otherwise dampen the sound a bit. For now, I have only the 5W fans I thought could be used, but apparently not.

Top fan:img-20220611-165211.jpg


Bottom fan:

img-20220611-165221.jpg
0 Likes 0 ·
iand avatar image iand ecoman ·

Looks like the high-speed (4000rpm 0.6A) fan is cooling the heatsink (by pulling air through it) for the power MOSFETs which is behind the PCB in the bottom right corner, which is where most of the heat dissipation will be. What's annoying is that Victron could easily have used a slower-speed 140mm x 25mm fan (like the Noctua industrial one) which would have been a lot quieter and given more airflow... :-(

The low-speed (2800rpm 0.3A) fan is just generally blowing cool air up through the case and across the transformer.

I don't suppose anyone with an original Multiplus/Quattro 48/10000 has done the same thing and looked inside? ;-)

0 Likes 0 ·