I guess if there are no input capacitors in the design then "they" will not wear out due to excessive temperature or ripple current. From an engineering standpoint not the way I would handle it. Does Victron engineering ever post on this forum?
They use a hybrid h bridge design. That thread also gets into calculating DC ripple. Wiring unlimited gets into theory of it and making sure you have sufficient battery and DC wiring. In this particular failure the battery bms failed, wasn't the inverter at fault.I guess if there are no input capacitors in the design then "they" will not wear out due to excessive temperature or ripple current. From an engineering standpoint not the way I would handle it. Does Victron engineering ever post on this forum?
I like that simplicity for ease of repair but you bring up a good point . I wonder if the ripple is worse than other inverters that have more capacitorsYes, there is a very large empty volume in the bottom of the 5kVA 48V Multiplus II UL unit. I would add as many large Aluminum electrolytic capacitors in there across the 48V input bus that would fit without severely blocking airflow. On one of the reviews (Current Connected) commented on not even being able to see any input capacitors under the heatsink. This will create very large low and high frequency ripple currents in the DC input cables all the way back to the battery. Also, causing stray "H" field radiated emissions and possibly interfere with other parts of the system or radios. Additionally, the poor caps (must be small if there are any) that are in the inverter input are very likely getting beat up with 120Hz ripple current far exceeding their rating. This will become a maintenance item for frequent re-capping the inverter. Other markets for military or aerospace applications have stringent requirements for input and output conducted emissions. I would be curious what Victron's expected life would be for these capacitors?
I am familiar with the "H bridge" topology. It is the ideal circuit to synthesize a sinewave in the positive and negative polarity. The FETs are controlled from a high frequency PWM that is fed from a sine reference signal. As the output voltage rises up the sinewave, the on duration of the FETS conduction time builds up to the maximum at the peak of the output waveform, then back down the other side. This creation of the sinewave draws maximum input current at the sine peak and minimum at zero crossing. (consider no output at zero crossing requires no input current, while output at the sine peak requires full input current) This process creates a large 120Hz (2nd harmonic of the line frequency) ripple half sine component in the input current. This half sine of current is drawn through the FET H bridge input stage from the input capacitors. If the capacitors are not present or undersized this current must come from the battery creating an antenna producing radiated emissions from the DC cables. Additionally, the PWM frequency (estimated around 25KHz) ripple current will be present at a lower amplitude. The inverter wants to see the DC input as a low impedance voltage source and containment of this ripple current at the point where it is created is the proper electrical design method. Victron engineering has to be aware of this, but the large capacitors required are not cheap and if not selected correctly their life is limited.They use a hybrid h bridge design. That thread also gets into calculating DC ripple. Wiring unlimited gets into theory of it and making sure you have sufficient battery and DC wiring. In this particular failure the battery bms failed, wasn't the inverter at fault.
![]()
Victron MPII Inverter Overload
There are 4 rows, of 6 each (depending on inverter size). A H-bridge looks like this: Just imagine that for each of the 4 FETs in the picture, you have up to 6 in parallel. In the older allumium-case models you could even see the heatsink/powerpack arrangement looked a bit like an X, with...energytalk.co.za
Given some of these units or their previous models have been working for decades in the field I'm guessing they've done their due diligence.I am familiar with the "H bridge" topology. It is the ideal circuit to synthesize a sinewave in the positive and negative polarity. The FETs are controlled from a high frequency PWM that is fed from a sine reference signal. As the output voltage rises up the sinewave, the on duration of the FETS conduction time builds up to the maximum at the peak of the output waveform, then back down the other side. This creation of the sinewave draws maximum input current at the sine peak and minimum at zero crossing. (consider no output at zero crossing requires no input current, while output at the sine peak requires full input current) This process creates a large 120Hz (2nd harmonic of the line frequency) ripple half sine component in the input current. This half sine of current is drawn through the FET H bridge input stage from the input capacitors. If the capacitors are not present or undersized this current must come from the battery creating an antenna producing radiated emissions from the DC cables. Additionally, the PWM frequency (estimated around 25KHz) ripple current will be present at a lower amplitude. The inverter wants to see the DC input as a low impedance voltage source and containment of this ripple current at the point where it is created is the proper electrical design method. Victron engineering has to be aware of this, but the large capacitors required are not cheap and if not selected correctly their life is limited.
Given the situation, it is best to keep the battery cables as short as possible, positive and negative conductors adjacent and touching - twisted is best. This will help to reduce radiation and minimize inductance. Be aware this ripple current will be imposed on the battery as a ripple voltage with an amplitude a function of its impedance and the peak-to peak ripple current.Given some of these units or their previous models have been working for decades in the field I'm guessing they've done their due diligence.
You can monitor DC ripple via a USB to ve bus mk3 cable adaptor connected to a Victron multiplus or quattro inverter.Given the situation, it is best to keep the battery cables as short as possible, positive and negative conductors adjacent and touching - twisted is best. This will help to reduce radiation and minimize inductance. Be aware this ripple current will be imposed on the battery as a ripple voltage with an amplitude a function of its impedance and the peak-to peak ripple current.
It would be an interesting exercise to compare the MK3 reading with calibrated lab test equipment. Of course, the inverter does not have direct access to the voltage ripple present on the battery terminals.You can monitor DC ripple via a USB to ve bus mk3 cable adaptor connected to a Victron multiplus or quattro inverter.
Kudos to you ricardocello for being aware of this issue and doing something about it!! You understand the design deficiency and are taking action!! The capacitor ripple current ratings effect on life are specified at the maximum rated ambient temperature. If your ambient temperature maximum is substantially less, you could encroach somewhat on the ripple current rating (it becomes a device life trade off). Or better yet add another one or two caps in parallel.My obligatory ripple threads:
Thread 'Guess what I’m going to do with these capacitors?'
https://diysolarforum.com/threads/guess-what-i’m-going-to-do-with-these-capacitors.95916/
Thread 'Victron Ripple Improvement Testing'
https://diysolarforum.com/threads/victron-ripple-improvement-testing.88093/
The risk is excessive DC ripple at the inverter inputs, causing over heating of the input capacitors.It would be an interesting exercise to compare the MK3 reading with calibrated lab test equipment. Of course, the inverter does not have direct access to the voltage ripple present on the battery terminals.
Thanks.Kudos to you ricardocello for being aware of this issue and doing something about it!! You understand the design deficiency and are taking action!! The capacitor ripple current ratings effect on life are specified at the maximum rated ambient temperature. If your ambient temperature maximum is substantially less, you could encroach somewhat on the ripple current rating (it becomes a device life trade off). Or better yet add another one or two caps in parallel.
Be aware that adding capacitors can actually make dc ripple worse. Don't try and patch it, figure out the root cause.Kudos to you ricardocello for being aware of this issue and doing something about it!! You understand the design deficiency and are taking action!! The capacitor ripple current ratings effect on life are specified at the maximum rated ambient temperature. If your ambient temperature maximum is substantially less, you could encroach somewhat on the ripple current rating (it becomes a device life trade off). Or better yet add another one or two caps in parallel.
I’ve measured with an o-scope in multiple locations, including at the inverter DC lugs.Be aware that adding capacitors can actually make dc ripple worse. Don't try and patch it, figure out the root cause.
Slap some ecoworthy 5kWh rack mounts in there for $800 or so delivered and should be golden,I’ve measured with an o-scope in multiple locations, including at the inverter DC lugs.
It’s way better. But your point is valid if a resonance with the inductance of the cables occurs.
The root cause is that there is too much total resistance in the path to the batteries,
Or I just need more batteries in parallel. The second thing is way easier to fix, and I will eventually.
Believe me, I’ve been tempted by all of these cheap battery review threads and Will’s review videos.Slap some ecoworthy 5kWh rack mounts in there for $800 or so delivered and should be golden,
No shame in waiting for block building to be built first, some of these fires are catastrophic.Believe me, I’ve been tempted by all of these cheap battery review threads and Will’s review videos.
I really want to build a battery shed outside the house first, concrete block and small hvac.
These batteries make me worry.
The root cause is that an AC voltage provider demands current from its source that is directly related to the instantaneous power delivered throughout the entire sine wave cycle. This is an inherent characteristic and the only way to deal with it is filtering. One system design option is to power as many high power loads from DC. Example: water heating or direct EV charging.Be aware that adding capacitors can actually make dc ripple worse. Don't try and patch it, figure out the root cause.
Victron’s design is a hybrid in the sense that they use high frequency AND a toroidal transformer.This conversation is way above my head but this was just posted in another thread today by @RCinFLA and it brings up the question- does the fact that the Victrons have a large transformer help absorb the ripple?
@Solarpete
Post in thread 'Low or high frequency inverter' https://diysolarforum.com/threads/low-or-high-frequency-inverter.99853/post-1355634
Sorry, is the Victron design different from say xw pro or sunny island? Or do they all work the same way?Victron’s design is a hybrid in the sense that they use high frequency AND a toroidal transformer.
Old thread:
Post in thread 'Inverters High or Low Frequency ?'
https://diysolarforum.com/threads/inverters-high-or-low-frequency.6300/post-103169
RCinFLA is by far the expert in this area.Sorry, is the Victron design different from say xw pro or sunny island? Or do they all work the same way?
Let's conjure him! @RCinFLARCinFLA is by far the expert in this area.
I have finally come to the realization the MP II 5kVA 48V inverter is really a 4kW inverter before we derate it's output power due to ambient temperature and power factor. Consider 40C ambient operation with a PF of .60, this will give you a 2220W inverter. A PF of .6 is not unreasonable for motor loads like a heat pump water heater or ground source heat pump. This now makes sense the UL cert. document specifies 33A.120×33=3960 or very close to 4000W?
View attachment 278046