question

In the documentation it says that the units need to be IDENTICAL. I think you knew this information beforehand and you knew this when you placed the order with your supplier.

I assume that you conveyed the importance of the units being identical to your supplier. I also assume that you conveyed to your supplier your circumstances surrounding your access limitations to the site.

So, if your supplier has provided two units which are not IDENTICAL then surely you need to take this matter up with them.

The way that the firmware is implemented should be of little consequences, the fact remains that both versions of firmware need to be IDENTICSL regardless of what features are physically present or what features you have the choice to use or not.

I can understand your frustration especially in your circumstances, but to me, the documentation is clear and correct - the units need to be IDENTICAL, no subtle differences at all. It's out of Victron's control if you supplier does not fulfill you requests.

I don't work for Victron, I'm just expressing my own opinion.

There is no subtle difference barring the serial numbers.

Are you suggesting that I should take it up with my supplier that the units have different serial numbers?

Because that is only knowable difference between these units. They are identically spec'd.

And a repeat order.

Only by performing a firmware upgrade and cross-referencing the firmware version against the firmware list could I establish the phoenix inverters use multiplus chips. There is no documentation that indicates this.

The multiplus chip difference relates to the current capacity of the multiplus transfer switch.

You CAN'T order different Phoenix inverters based on the current capacity of the transfer switch because the Phoenix inverter DOESN'T have a transfer switch.

There is no ordering choice, ergo it is Victron's problem when I order "IDENTICAL" items and get items with subtle, unknowable, undocumented differences.

At the risk of repeating myself there are no transfer switches on this unit, so there is no actual hardware incompatibility.

If you order a multiplus with a 50A transfer switch there no requirement to intrusively open up a static sensitive equipment to confirm the 3rd and 4th digit on a IC on the circuit board.

The expectation is you get what you order, without having to don an anti-static wrist strap.

Now if you order a phoenix it does not have a transfer switch, so there is no ordering option.

Why should it now be an unspoken expectation to open the unit up?

There are 8 phoenix firmware versions, none of which relate to this 48V 5kVa unit.

My contention is if there is only one ordering option, and the incompatibility issues actually relate to an entirely different product range, then it is unreasonable to expect a buyer to be aware of issues at a microprocessor based level by intrusive measures.

The selling point is that it is parallel connectable. Somehow that gets put across very loudly the product support is also put forward.

The argument that somebody should've known is an unreasonable one, when the issue is so subtle and undocumented in this particular product's manual.

After having commissioned literally 1000's of MW of power systems over 30 years that use micro-processor equipment, I have a pretty good idea about what is due diligence from the middleman and what is a reasonable expectation of the manufacturer's responsibility.

But, in any event I'd like to hear from Victron now.

A parallel configuration requires extremely precise balancing.

This can come down to the variation in resistance between two identically branded switches or wire join.

The variations between unit design or part changes can introduce resistance imbalance between the units. And other consequent issues.

The issue here isn't that you are not using the transfer switch functionality of the processor, but that different units were supplied to you when identical units are necessary.

Configuration of parallel systems is for advanced users, and very often with close support by a local distributor or supplier. Then units intended for use in parallel systems are specified when purchasing as identical, and for parallel use.

There is a very clear Warning notice on the front page of the parallel systems manual here:

We do not support untrained and/or inexperienced installers working on these size systems.

...hire an installer that does have experience with these large systems, for both the design and the commissioning.

Regarding feedback from Ben, We are installing in the real world, so it will never be perfect.

The goal is minimise to potential number and consequence of the imbalances and irregularities. A line has to be drawn somewhere, and it was drawn at requiring identical units. I don't think it is unreasonable, though it can of course be frustrating.

You are right that there are other factors outside of the control of Victron, which will impact our equipments performance. There are enough opportunities for unexpected issues through system imbalance as it is.

Nevertheless, this requirement will not be changing. It should be known loud and clear that if you wish to parallel Victron inverters:

• It should be done by an installer with previous experience and support from a trained distributor
• It should be done with parts that have been specified and supplied for a parallel system at time of purchase.

I believe that these systems are sufficiently complicated and expensive that there is some onus on the distributors and installers to do their research on what is required. These requirements were already existing, but in light of this post I am now increasing the prominence at the top of the parallel systems manual under WARNING heading.

Other equipment that is able to be connected in parallel such as MPPT's and DC to DC converters do not have these strict requirements.

Sorry, about dropping the issue, I have been travelling internationally.

I see Ben has picked up where I dropped off.

He has also picked up on the the point that this is an ARTIFICIAL difference.

Victron designer's used two different chips in batches without considering the paralleling incompatibility. The Phoenix inverter manual does draw attention to firmware, but only to the first 2 digits and the last three. Which were identical in this case.

I contend that, yes the unit's are identical and therefore it was a design error to not standardize on the chip.

"and nobody knows what other changes are inside." Don't tout training and make statements like that.

In any event it is twaddle, without further evidence I contend these are two standard multiplus chips, there different applications are clear, when deployed in the Phoenix an unnecessary soft artificial restriction prevents paralleling.

It was an unconsidered result of thinking you could use different chips because the transfer switch functionality is not needed in the Phoenix.

Now, that is QC issue within Victron's design function, it should have been picked up it standard type-testing. Don't pass off an inherent design flaw as the QC responsibility of the supplier or installer.

Unless you can point to a technical reason why those two inverters should have different chips:

"and nobody knows what other changes are inside."

doesn't cut it.

Thank-you M.Vader.