24/3000 Multiplus Voltage / Frequency Readings Off

Q&A and troubleshooting
, , , ,
1

The Issue I have is:

I have a brand new Multiplus 24v 3kw unit that doesn’t correctly indicate the DC voltage on the battery terminals, and is a bit off on what it shows the input grid frequency is.

My background:

Retired now, but I spent over 34 years in the hydro power industry as an electronics / control system / communications tech, and was a foreman over the meter / protection shop, as well.

I have had training and experience in maintaining and repairing power control systems, and several older style ferro-resonant single phase and 3 phase UPS systems, as well as several PWM style units.

About the only background I have on these smaller units is from helping a friend when he had problems with a 8kw 48v Outback split phase unit in his off grid cabin. He had about 4kw of solar, a small Kubota 7.5 kw gen set, and it had some great big diesel locomotive starting batteries for about 60kWh of energy storage.

System:

My system is composed of a 100/30 MPPT controller, a Victron 500 amp Smart Shunt, and a 24/3000 Multiplus unit.

Energy storage is 400 Amp hours of 24v LiFePo made up of 4 strings of 2 each 100ah 12v batteries. All battery connections are made with #2 awg cable.

They are all fused at the individual battery string positive terminal(s) with marine style 100 amp fuses and each string goes through a 100 amp breaker to be paralleled up on a 400 amp junction block, which is further connected to a 350 amp positive distribution bus via 1/0 awg cable.

Negative #2 awg connections are all paralleled up on a 400 amp block, and the 500 amp smart shunt is connected to that junction block via 1/0 cable and on to the negative distribution block, also with 1/0 cable. The Multiplus is connected to the distribution buses by #1/0 awg cable, with a total run length, (negative + positive) of less than two feet of 1/0 awg.

In terms of performance, it works just as advertised, and quite well at that.

The Problem:

The issue is that the Multiplus is not displaying the DC voltage accurately, and it’s making fine tuning the DC absorption and float voltages a pain.

The Multiplus reads almost exactly 0.4 volts LOW. My Fluke 87 on the DC terminals will read 28.0 volts, the Multiplus reports 27.6 volts.

All of the other Victron Devices match within ± a digit or so of what the Fluke 87 indicates. (As did a 24/1200 Phoenix I had in this system for a short while.)

If the MPPT said the battery voltage was 27.23 volts, the shunt would indicate 27.22, the Phoenix would say 27.24 or in that ball park. The Fluke 87 would indicate 27.23, so basically, everybody agreed with each other.

Swap the Phoenix out for the Multiplus, and it indicates almost exactly 0.4 volts LOW.
Meter, shunt, MPPT all still agree with each other. Multiplus is the odd man out.

Well, no big deal, just apply an offset, you say?

The point is, I have checked THREE other Victron devices and they were ALL DEAD NUTS ON as compared to the Fluke 87. (Which, by the way, was checked several years ago against a NIST traceable calibrator…)

That does NOT sit well with me. Three other Victron devices are right, and this one isn’t??

Ok, fine, lets try the offset…

Ran into the NEXT issue.

When the unit is CHARGING, it OVERSHOOTS the set point that I’ve programmed in. Watching the voltage via a meter directly on the Multiplus terminals shows an excess of well over several tenths of a volt in excess of the setpoint. (The remote voltage sense wires are NOT connected, and it didn’t make any difference when I DID connect them.)

In other words, if I want absorption to sit at 28.1 volts, I’ll set the Multiplus to 0.4 less than that value, (IE: 27.7 volts,) and watch the meter on the DC terminals of the Multiplus shoot up over 28.4, 28.5, and as high as 28.6 volts, and it will HOLD that voltage. It doesn’t just overshoot and drop back, it stays there for a while, which is enough to trip off the ‘charge switch’ in some of the batteries internal BMS. When that happens, the other battery in the string no longer gets to sit and finish the top balance cycle, because there is no longer any current flowing in that string.

So, to keep the overshoot from triggering the BMS in the individual batteries, I have to offset the set point by the measured STATIC offset of 0.4 volts, PLUS the ‘overshoot’ value of almost 0.4 to 0.5 volts. That makes the set point in the Multiplus almost a full volt below where I want it.

One OTHER issue:

Also, the Multiplus indicates 60.3 Hz on the input, and 59.9 Hz on the output, and this is with the pass-through relay CLOSED.

Fluke meter says grid is bouncing around at about 60.02, 60.01 etc. Hundredths of a herz changing values is expected.

BTW, if we had a grid frequency of 60.3 hz, we’d have been scrambling to drop generation…

Anyway, the local Victron dealer wanted me to just swap my unit for one of theirs in stock. I said sure, but FIRST, you drag it out of the box, hook it up and see if it measures ‘off’ like mine does.

Yeppers. Right out of the box, it shows that with a voltage of say, 27.0 volts DC on the input, the Multiplus reports 26.6v.

Same thing with the input frequency. Shows 60.3 Hz. Output shows 59.9 Hz.

I HAVE updated the firmware in my Multiplus to the latest firmware, (according to Victron Connect,) so if it’s a problem with older firmware, well…

So, Victron, do you have an issue with calibration of this line of product, or?

Steve

2

Rma it

3

The local dealer is more than willing to replace it, except I made them test theirs first before I rip mine off of the wall…

And it has the same problem.

4

Have you tried wiring the remote voltage sensing? - not that I’m suggesting that you have volt drop external to the inverter.
BTW, you could check to see if there is any voltage difference by measuring the voltage at the remote sensing terminals before wiring. There is a 120 ohm resistor from B+ to remote sense + and same in the negative. IF for some reason the remote sensing circuit is taking some current this could cause the small drop.

1 Like
5

Yes, I hooked it up. Made no difference. Removed it.

BTW, the MPPT device works exactly as programmed. Voltage on the battery goes to EXACTLY what I setup in the unit. No overshoot. Drops back into float at the correct point.

The only ‘gotcha’ with the MPPT is that you can’t use the ‘tail’ current to trigger the switch from absorption to float if you have a load on the system. It has no way of differentiating current going to the battery and current going to power whatever the load is.

The solution is to add the Smart Shunt. I have the MPPT and the shunt setup in the Bluetooth Network mode so the MPPT can ‘see’ what the actual battery current is. Works like a champ.

Now, having mentioned that, I could try adding the VE Bus ‘Smart Dongle’ to the Multiplus, add it to the local Bluetooth network and see if the unit uses the voltage values from the shunt to ‘override’ its’ internal sensing.

However, I should NOT have to spend another $100 to attempt to FIX what appears to be a design / calibration flaw.

I say ‘attempt’ because I don’t know for certain if the Multiplus will use the shunt values and over ride its’ internal sensors. I have been thinking of getting one so just so I don’t have to go out and plug into it to see what it’s doing, and I MIGHT be willing to part with the $$ to do so if it becomes obvious that Victron isn’t going to cough up an answer / fix.

Another BTW…

Given that the VE Bus appears to be RS-485 based, (and a relatively ‘slow’ 2400 baud,) I was just going to run in a length of Cat5e to the computer in the house instead of getting the Smart Dongle, but if it will ‘cure’ the voltage issue…

6

I Checked my MP2 24/3000’s this morning. DC voltage is 0.23V low. (measured against MPPT,s smart shunt and BMS).
The solution seems to be to use the smart shunt in conjunction with a GX device with DVCC enabled. This is a necessity for Lithium batteries anyway, as the charge current needs to be attenuated as the cells approach full charge.
With the MPPT’s it is possible for a processor to ‘write’ the correct voltage to the MPPT unit, overwriting the internal measurement. This overcomes the volt drop in the mppt leads at full current (in my case about 100mV). I don’t think that this is possible for the MP2 inverters, though they do follow the DVCC (Dynamic Voltage and Current Control) from the GX device, so that would suggest it is possible. I’ve not looked into the VE-Bus programming for this yet, as I don’t often use the Multiplus for battery charging above 50%; my situation is 100% off grid, and I run about 99.95% solar, with about 50 hours gen time a year.

1 Like
7

I think the solution is for Victron to get off of their collective rears and FIX the problem, not ‘mask it’ by installing external gear.

I have NO intention of buying a Cerbo or any more gear from a company that doesn’t seem to want to respond to an obvious calibration or engineering screw up that appears to cross several device lines.

The sales people must like this though. Buy the gear and then find out to make it work the way it should, you’ll need this ‘add-on,’ oh and this add-on for this function…

I’m NOT talking about buying a Cerbo or whatever to get the functions that they ONLY can support, like internet access, advanced programming, display functions, etc.

Or the need to add on something to support something like the tail off function of the MPPT, where in order to use that function, you have to have the shunt to measure actual battery current.

Adding a device to ‘mask’ the problem is NOT a solution, in my opinion.

Most of the power supplies I have run across have a tendency to ‘back off’ of the current flow the closer they get to their set points, unless you have something like a precision four lead device with ‘remote’ sensing built into the supply, and a very tight ‘slope’ in the regulation circuitry.

It should be pretty easy to build in a programmable set point trigger for a ‘slope’ factor, (DVCC as Victron terms it,) in something like this unit, since it’s all microprocessor controlled. Set it to trigger at say, 0.5v below the supply set point, and then back off the current by some programmable factor as the voltage gets closer to the set point. Maybe something along the lines of what a PID controller does.

8

I’m not trying to excuse Victron, or promote the use of the Cerbo - unless you need the remote monitoring - and you can do this on either an R-pi or Beaglebone. Probably the Multi’s measurement circuits were designed when measuring the voltages and currents were not so critical - and they haven’t been improved to match the more stringent requirements of today…
With lead acid batteries, +/- 1/2V in 24V was no big deal, with lithium it’s critical. However, with lithium batteries you tend to have a very accurate BMS system providing the data - so again, the measurement errors in the multi are less critical.
With the MPPT’s these can be set to transition to float after a time period spent at the Absorption voltage, rather than transitioning on the tail current.
With the BMS (I designed my own), it is possible to set the slope of the ‘tail off’ of the current as you approach full charge, I’ve seen the results when a certain large company forgot to do this, and didn’t throttle back until 100%…the results were predictable…
Happy New Year…

1 Like
9

Yep. I understand the voltage tolerances. Been working with lead acid for years. Everything from standard flooded, VRLAs, to gels, etc.

The batteries I use, (Wattcycle 100ah minis,) obviously have their own BMS in each, and I paid an additional $20 per battery to get them with Bluetooth so I can see what they’re doing. However, there is no way, (currently, pun intended,)

to communicate those values to anything other than an app on a phone.

I settled on those because of several tear down and testing vids on You Tube. They are quite well put together internally, and when they were tested, performed slightly better than advertised, including cold charge and high current cutoff.

So far, they’ve performed very well, and I have discharged / charged them several times, (20% remaining,) and measured all the battery voltages during the cycle. They have all tracked within several tenths of a volt of each other when under load in both ‘directions.’ (Actually, most of the time, they are spot on compared to each other.)

The internal BMSs have shown about 5% maximum capacity differences between batteries during parts of the cycle, but it usually balances out at both the discharge end and the charge end.

I had one battery in a string showing slightly less voltage than its mate, and decided to see just how much it took to get them ‘equal.’

Hooked up a 0 - 55 volt 5 amp power supply set to 14.3 volts, and clipped it on. Pulled 5 amps for about 3 minutes and the current rapidly dropped off as it got closer to that voltage set point, Within about 5 minutes, it was down to 200 ma, and the individual cells were just creeping up over 3.55 volts. The BMS will disconnect charging if it gets much over 14.45, or if one of the cells is over approximately 3.62v.

Pretty good. Not enough to mess with, really, but I’ll keep an eye on them.

Picked up a Smart VE Bus dongle today, and it DOES overwrite the ‘local’ voltage value in the Multiplus. It’s a bit fussy in terms of giving it time to communicate with the Multi, but it does work.

It also ‘agreed’ with all the other Victron gear and the Fluke 87, within ± a digit or two.

I set the Multi absorb to 27.25, float to 27.0 1hr, and did a quick short discharge / recharge to see what difference it made. Connected and communicating, the voltage came up to about 27.35 or so. Tenth of a volt overshoot, and tolerable. Unplugged the RJ45 connection to the Multi, and after a bit of time, it starts charging again, and the voltage climbs up to 27.87. Wait 5 minutes or so, and plug the RJ45 back in. It took a minute or so to get the communication going again, and the charger promptly cuts off and the voltage starts slowly creeping back down. So, the beast IS listening to the Smart Dongle, but if the communication gets hosed, then it will go right back to its old tricks.

I also noticed that I had trouble when the MK3 is hooked up along with the Smart Dongle. RTFM, and sure enough, there is a note about the Multi being able to only communicate to ONE device at a time…

Nuts. I was going to run a chunk of Cat5e into the house and monitor things from the PC using the MK3. Oh well…

10

Update… I joined the Smart Dongle to the local Bluetooth Network just to see what it would do.

I also let the Multi go into float, (set for 27.00) and with the Smart Dongle connected, it went to 27.05 to about 27.1 volts, verified by the shunt and MPPT voltages. Unplugged the Smart Dongle, and waited. And waited. Finally, after a few minutes, the charger started pushing current, and the voltage went up to about 27.55 volts at which point the current dropped to less than an amp, went to zero current, and then slid down slowly to a discharge of about -0.25 amps. I’m just going to leave it in this state for an hour or so, and see where it eventually settles down to.

1 Like
11

Ok, without the smart dongle, voltage settles down to 27.4, so the charger is obviously off by 0.4 volts, as expected.

Plugged in the RJ45 connection, and this morning, 29.98v so using the smart dongle is a ‘fix’ for now.

Hey, Victron, ya wanna send me a hundred bucks to compensate for having to buy another device to ‘fix’ the issue?

12

Dang, Mike, I just re-read your message about checking the voltage AT the remote voltage terminal points. It didn’t click that you wanted me to measure whatever I find there without the wires hooked up…

I’ll do that!
Thanks,
Steve

13

Checked, and the voltage present on the remote terminals is within ± a digit of what is measured directly on the DC terminals…
Nuts.