EPEVER Tracer 3215BN

[Weasel]

I got my hands on an EPEVER tracer and i am a bit astonished a little dissapointed and also slightly relieved. I say relieved because it doesn't beat a bluesolar

In short for those who want to know quickly, It is a very high quality, like in reeeaaly very good MPPT. as far as i could tell It has good conversion efficiency and is excellent bang for buck.But It has tracking issues that can be mitigated as long as you configure your panels the way it likes, basically PWM rules apply.

The one i got here is the 3215BN and its a 30A MPPT with a 20A load output accepting 150V input and outputs to 12/24V. i was obligated to take it apart before i tested it  and i was pleasantly surprised. it is really well built well designed with very high quality components, dare i say surpassing the likes of victron, perhaps just on components but still.

The beefy inductors are potted nicely in the heatsink casing, several components are heat transferred to the main heatsink with thermal pads, not talking about the mosfets which of course would be, things like control chips, even a cap, its a nice touch. Speaking of caps, they are none other than Nippon Chemicon KZ#.  Comparing them to the same series ones i have here they don't appear to be fake either. Its got nice big terminals, just use ferrules with them because the are screw type. All in all i am thoroughly impressed with its quality and at the price you cant go wrong.

But, and its a big but, its tracking is rather poor. It adjusts quickly in response to amps availability changes but tracks very slowly with respect to the voltage, only after an hour did it reach the voltage "goal" i set for it, and the higher that voltage the longer its going to take. This becomes a real problem when you take into account that it starts up and locks in the voltage as soon as the PV voltage passes the battery voltage by a few volts. if this doesn't sound like a problem to you let me give you an example. 

if you use 3 60 cell panels at ~32V that does 7A each you could be tempted to string them up in series for 96V to overcome cable losses, after all the unit can handle 150V. That should be a 672W string (96 x 7). now if i'm charging a 24V battery bank i should be getting 28A into the bank when 7A is available. But what happens is as the sun rises the voltage on the panels rises slowly, as soon as it goes just above the battery voltage by about 4V it starts drawing current, it does a sweep trying to find the best operating voltage once, and after that never lets off and keeps the voltage low.  In this example case lets call it 30V; meaning its going to be stuck under 250W for a while, very slowly ramping up until it eventually reaches 96V operating where it can do +600w. Now, i said it changes voltage slowly, from what i saw it was about 3 minutes for a 1V change, so three hours before it would reach the optimal voltage? i didn't test this 3 hour ramp up exactly, i tested for a little over an hour, but the problem is there. I should also say if you where to isolate the panels and connect again at the full 96V it would start tracking quickly and at the correct voltage no problem, its that slow voltage ramp up that is the problem. 

I saw i video of someone complaining the tracer didn't track nicely in the morning and well, this explains it. As i mentioned at the beginning the solution would be to stick to PWM rules and keep the panel voltage above the battery voltage but not far above it. a 60 cell panel is perfect for a 24V bank, its why we have 60 cell panels actually, and the three as per our example would be in parallel. Any light would get them up to 37V ocv very quickly and the epever would work beautifully.

Finally, my ultimate opinion is this MPPT is a firmware rewrite away from being quite brilliant.

https://drive.google.com/open?id=0Bz4hQaGz4RJpNndkNExMdEZKZEk

https://drive.google.com/open?id=0Bz4hQaGz4RJpVFdDZk9CRDJmQXM

 

[___]

So it sounds like it has good hardware and needs better firmware? :-)

[Hunter85]

good morning, I have epever 3210 An on my camper and it works perfectly with a 350 watt panel, I live in the camper for 6 months.

I think epever makes good products at a fair price. Victronenergy has the same specifications and the price is inflated due to the European brand often.

I think it's perfectly normal that the solar controller takes an hour to reach the maximum constant voltage if the battery is not 100%. in fact, at the beginning the battery absorbs the maximum current, but the voltage is not maximum. the voltage increases as the absorbed current decreases. in this phase the controller works in MPPT mode. I think that all the algorithms of the mppt controllers and the latest generation of chargers work like this, this is also written in the manuals. it's not a defect of that controller

[___]

The issue seems to be that it tracks really slowly to a new maximum power point once it has found one on the initial scan. The easy solution to that is simply to do a periodic rescan from zero... that is what lots of controllers do. But finding a new MPP, tracking multiple points... etc... that's what all the R & D is about at the moment.

[Hunter85]

I did not understand very well what you're talking about because my 3210AN seems very quick in the exchange.

the company epever writes that their controllers work like this, as in the picture I posted.

it's correct??

[Gabriel_2018]

I'm sure Tracer is quick. 

May be @Weasel is considering that Vmp of a string of three 60 cells is 96 V and that's only happens when temperature cell is 25°C.

The slow increase of Voltage is typical in absortion stage. The MPPT increase PV voltage in order to reduce the charge of the battery.

If ambient temperature is 20 °C, the Vmp of that string Will be round 75-80 V, far from 96 V.

Edited November 6, 2018 by Javi Martínez

[Hunter85]

44 minutes ago, plonkster said:

The issue seems to be that it tracks really slowly to a new maximum power point once it has found one on the initial scan. The easy solution to that is simply to do a periodic rescan from zero... that is what lots of controllers do. But finding a new MPP, tracking multiple points... etc... that's what all the R & D is about at the moment.

I did not understand very well what you're talking about because my 3210AN seems very quick in the exchange.

the company epever writes that their controllers work like this, as in the picture I posted.

it's correct??