High voltage (600V) regulator/limiter for solar panels

[zac]

I have a solar array with an open circuit voltage of 640V and a maximum power point voltage of 510V and a charge controller that can tolerate up to 550V input.  So, I'm trying to figure out the best way to limit the voltage to 550V.  In practice, the array would only exceed 550V during unusual conditions (such as very cold and high intensity sun such as immediately after a storm). 

The array has a maximum output current of 10 A (about 5 kw).  The simplest solution seems to be a shunt regulator.  The simplest form may be a string of series connected high power (50 watt) zener diodes with the combined breakdown voltage a bit under 550V.  The shunt regulator would never need to dissipate the full output power of the array since even a modest load (10-20%?) should drop the voltage below the 550V threshold.  I have minimal experience with high voltage electronics.  Anyone have any ideas or experience with this?   

[kripton2035]

Why not split the solar array in half ? take care of the max power input in your inverter but it would have less losses than shunts.

[zac]

Why not split the solar array in half ? take care of the max power input in your inverter but it would have less losses than shunts.

Splitting the array is not an option because of how it is wired.  The other limitation is the charger/inverter only has 2 mppt tracking inputs and both are in use.  

The power loss in a shunt that clamps voltage to 550V in this array would be minimal since the voltage could only exceed this in very vary circumstances.  The main likely scenario is during a sudden sunny episode after a storm in cold weather.  In that situation, the mppt tracker would likely have reduced the load and allowed the voltage to rise to a relatively high level with low solar generation.  A sudden large increase in sun could cause the array voltage to exceed 550V and approaching the 640V open circuit voltage before the mppt tracker increases loading.  This is about the only circumstance in which the shunt would dissipate energy.  I'm uncertain exactly how much energy the shunt would have to dissipate in that situation to clamp the voltage to under 550V, but I'm guessing it should be well under 20% of the array wattage rating so 1 kw at worst.  

So, I'm considering a couple strings of 50 watt zeners in parallel.  I would use at least 2 strings for redundancy.  Another possibility to reduce dissipation in the zeners is to include a light bulb/heating element in the string.  One concern is voltage accuracy of the zeners since typical high power zeners only have 5% tolerance.  

Posted on: May 14, 2025, 03:04:49 15:04 - Automerged
It is curious that some particular zener models of similar wattage are much less expensive than others.  For example, this 1N3336B Zener, 68 V, 50 W, DO-5 is $5 from a distribution.  

https://www.newark.com/solid-state/1n3336b/zener-diode-50w-68v-do-5/dp/10P4952

The zener's in series don't have to all be the same voltage.  I imagine having one that is the rated voltage for a DC fan to cool the heatsink they would be mounted on.  Another problem is I don't have a power supply that operates up to 600V to test this.  Any idea to produce a relatively high wattage 600-700V power supply may be with a variable transformer feeding a step up transformer and bridge rectifier/capacitor.    Since the 600V is well below the breakdown voltage of typical small transformers, it should be possible to use an ordinary 120VAC to 24VAC step-down transformer in reverse.    To get 650VDC, I would need 460VAC so a ratio of about 4 operating from 120VAC utility voltage.  

[kripton2035]

>Splitting the array is not an option because of how it is wired.
then it has not been wired the right way !
good luck !

[zac]

>Splitting the array is not an option because of how it is wired.
then it has not been wired the right way !
good luck !

The current array is wired in series and I don't want to dig another trench to run additional wiring. 

[kripton2035]

I more or less understood something like that...
you must match the array and the possibilities of the inverter before digging and installing...
the easy move would be to change the inverter now to match the array then.
I don't know if your zener mods will work or not, but it's worth a try.

[pickit2]

how about a comparator circuit with a relay to add a load if the voltage is above the say 540V.
you could even add it to the panel groups, so if a group is below the upper limit they still supply the full power.   

[Old_but_Alive]

@pickit2,  good thought. dont like relays though

i would add a variable current source (load), big mosfet ( probably several in parallel, with equalising reistor in the Source, so the G-S voltages are the same), big heatsink and fan, and control loop to keep the voltage at 540V.

so the current is variable.

and fuses of course

aditionally,Vgs falls as the temperature increases, so you need the resistor to mitigate and equalise  this.

Aquick look, found a mcrochip part APT6011B2VRG, APT6011LVRG MOSFET MOS5 600 V ..., I am sure there are many others.

[pickit2]

Relay was a quick response. as you could use the excess voltage, for whatever ie: charge a smaller battery. 
I would use a powered Test load, don't everyone have one in their tool box.

[Old_but_Alive]

@pickit2,

I was in no way critisizing you darling.

lets converge on a good solution

Mike

[zac]

I more or less understood something like that...
you must match the array and the possibilities of the inverter before digging and installing...
the easy move would be to change the inverter now to match the array then.
I don't know if your zener mods will work or not, but it's worth a try.

When I started the ground mount solar array in 2008, I did plan ahead to some degree, but did not imagine I would purchase equipment in 2025 that can only withstand 550V maximum open circuit voltage instead of the more typical 600V or 750V of the SMA TL series grid tie inverters I installed.  I had forgotten the details of my wiring and thought the wiring to the array was direct burial UF cable, but it turned out that I used 3/4" liquidtight metallic flexconduit.  Oddly, the wire in the conduit is only 14 gauge.  Presumably, that is what I had on hand or was available at a good price at the time, but I found 3 unused wires in the conduit so could split the array without pulling additional wiring. 

I got a Bluetti EP900 on/off grid battery system with 2 mppt solar inputs that are limited to 550V.  Since I only got the EP900 a couple weeks ago, I'm just setting it up now.  I'm planning to stay with the old SMA TL grid tie inverters when utility power is available, but want to configure the system so that I can easily switch the solar array to feed the EP900 during extended power failure. 

I suggested the zener voltage limiters to a friend many years ago who used it in his small hydropower system and that has worked well for nearly 25 years.  I haven't used them myself though.  After finding that I did run conduit and have  spare wires in the conduit, I may just split the array into 2.  Combining the 2 with diodes probably won't be the most efficient, but may suffice since it would be used in that configuration except during extended power failure which is not a frequent event. 

[zac]

I still haven't decided what to do, but found this 1000VDC variable power supply on ebay for $50:

https://www.ebay.com/itm/334503593016

The specs say it can put out 1000V @ 0.4A which should be adequate for testing this. 

https://assets.ctfassets.net/ntf98c1thwoe/1AojVLJke9WasHhTvkaMS2/184f9df025c552959ed4360400e8e0be/labconcopowersuppliesmanual1992-2007.pdf

[sadman]

A comprehensive article series on subjective topic

https://www.electronicdesign.com/technologies/power/article/55286293/texas-instruments-power-electronics-evolve-for-the-era-of-high-voltage-dc-hvdc-distribution?o_eid=5207B4070834D1J&oly_enc_id=5207B4070834D1J&rdx.ident[pull]=omeda|5207B4070834D1J&utm_campaign=CPS250423223&utm_medium=email&utm_source=ED+Power+%26+Analog


https://www.electronicdesign.com/12345/whitepaper/21233720/electronic-design-factor-pfc-into-your-power-supply-design?o_eid=5207B4070834D1J&oly_enc_id=5207B4070834D1J&rdx.ident[pull]=omeda|5207B4070834D1J&utm_campaign=CPS250423223&utm_medium=email&utm_source=ED+Power+%26+Analog

https://www.electronicdesign.com/technologies/power/whitepaper/21162012/electronic-design-how-to-improve-llc-converter-power-density?o_eid=5207B4070834D1J&oly_enc_id=5207B4070834D1J&rdx.ident[pull]=omeda|5207B4070834D1J&utm_campaign=CPS250423223&utm_medium=email&utm_source=ED+Power+%26+Analog

https://www.electronicdesign.com/technologies/power/whitepaper/21183762/texas-instruments-understanding-common-isolated-power-ics-for-digital-isolation-systems?o_eid=5207B4070834D1J&oly_enc_id=5207B4070834D1J&rdx.ident[pull]=omeda|5207B4070834D1J&utm_campaign=CPS250423223&utm_medium=email&utm_source=ED+Power+%26+Analog

https://www.electronicdesign.com/technologies/power/article/55241448/electronic-design-overcoming-the-ai-power-crunch-requires-more-robust-isolation?o_eid=5207B4070834D1J&oly_enc_id=5207B4070834D1J&rdx.ident[pull]=omeda|5207B4070834D1J&utm_campaign=CPS250423223&utm_medium=email&utm_source=ED+Power+%26+Analog


sadman

[kreutz]

You can also install a buck converter at the array's output and hook the inverter's PV input cables to the output of the buck converter. That will probably require a PV output voltage monitoring system and a bypass.