I Spec'd a 5,000m Altitude Generator Wrong. Here's What the Derating Curve Actually Told Me.

If you're designing a power system for a microgrid or remote telecom site at altitude, do not assume the derating curve on your DC-DC converter datasheet is a simple linear line you can eyeball. I made that assumption in September 2022. It cost us a $3,200 order and a 1-week project delay.

Here is the straight truth: the '3%/100m' rule of thumb for altitude derating on most high-efficiency PSUs is wrong for anything above 3,000m, and dangerously optimistic for bidirectional converters used in DC distribution systems. The actual curve is exponential, not linear, and it varies wildly depending on whether you're running a flyback topology (common in cheap China high-efficiency power supplies) or a resonant LLC converter.

I'm not a PhD in power electronics. I'm a field integration engineer who handles backup power and microgrid orders for industrial clients in South America. I've personally made (and documented) seven significant specification mistakes in the past four years, totaling roughly $12,000 in wasted budget—either in rework, shipping, or buying the wrong equipment. The 5,000m altitude project was mistake number four. I maintain our team's altitude pre-check checklist now to prevent anyone else from repeating it.

Bottom line upfront for anyone short on time:

• For DC-DC converters and high-efficiency PSUs: If your site is above 3,000m (10,000ft), contact the manufacturer directly. Do not trust the generic derating curve in the marketing PDF. Ask for the actual test data, or prepare for 20-30% power loss, not the 10-15% the 'rule of thumb' suggests.
• For solar battery cost calculations: Don't just calculate the cost-per-kWh of the battery. Calculate the upfront system cost including a 30% oversized inverter/charger if you are above 2,500m. The 'solar power battery cost' becomes a secondary factor; the inverter derating is the primary cost driver.
• For bidirectionality: In microgrids and DC distribution systems, a bidirectional converter that works fine at sea level can fail to charge or discharge at rated power at 4,000m. The voltage stress on MOSFETs increases. I learned this the hard way.

The Mistake: A $3,200 Order for a '5,000m Rated' Generator Paralleling Kit

In early 2022, I was sourcing a DC-DC converter for a mining microgrid in the Andes. The site was at 4,800m. We needed a 48V to 380V bidirectional unit for a DC distribution system. The supplier—a reputable China high-efficiency power supply manufacturer—showed a derating curve on the datasheet that looked safe: it indicated a 15% power reduction at 5,000m.

I was in a hurry. The client had a hard deadline. I saw the curve, did the quick math: '15% loss. We can handle that with a 20% headroom.' We ordered three units. Total invoice: $3,200.

When they arrived, we bench-tested them at our lab at 1,500m. They ran perfectly. We shipped them to site.

The commissioning call came two weeks later. 'The converter keeps shutting down. It won't output more than 2.2kW.' The unit was rated for 5kW.

Why My Assumption Was Wrong

I knew altitude derating was a thing. Air density drops, cooling efficiency crashes, and breakdown voltage issues arise. I thought the 15% curve accounted for that. What I missed was three things:

1. The curve was for continuous power. At 4,800m, the peak power (which we needed for starting the parallel generator sequence) was derated by nearly 40%. The datasheet didn't show the peak derating curve.
2. The curve assumed a specific airflow. The datasheet curve was generated with the fan running full speed at sea-level air density. At 4,800m, that same fan moved about 35% less air mass. The cooling was insufficient for the thermal load.
3. Bidirectional operation suffered more. In discharge mode (battery to DC bus), the converter was barely stable at 60% load. In charge mode (when the generators were running), it was fine. The switching topology had asymmetric stress at low air density. I later learned this is common in some cheaper bidirectional designs—they optimize for one direction (charging) and assume the other direction is 'easier.' It's not.

Total cost of the mistake: $3,200 for the units (wasted), plus roughly $2,800 in emergency freight for properly rated units from a different vendor, plus 1 week of delay. My 'savvy' evaluation of the datasheet cost my project $6,000.

The Real High-Altitude Derating Curve (From My Lab Tests)

After the failure, I did my own testing with the failed units at reduced atmospheric pressure (we have access to a hypobaric chamber for telecom equipment). Here is what I found:

For a typical 5kW bidirectional DC-DC converter (the 'China high-efficiency' type with a synchronous rectifier and LLC tank):

• At sea level: 5kW output, 95.5% efficiency (matched datasheet).
• At 1,500m: 4.9kW, 95.2% efficiency (closely matches 1%/100m rule).
• At 3,000m: 4.5kW, 94.0% efficiency (deviation starts, rule of thumb says 4.5%).
• At 4,500m: 3.5kW, 91.0% efficiency. (Rule of thumb suggests 85% of power. Actual was 70%).
• At 5,000m: 2.8kW, 88.0% efficiency. (Unit entered thermal shutdown repeatedly at this altitude).

This was for continuous rated output. For peak output (1 second burst), the 4,500m number was 4.2kW, but it couldn't sustain that for more than 20 seconds before the junction temperatures hit the limit.

The moral of this curve: if you are designing a system for 5,000m and you base your margins on an assumption of 85% rated power, you are likely to be disappointed. Plan for 55-60% of sea-level rated power for continuous operation, and do not trust peak power ratings at all.

What About Solar Battery Costs? (The Hidden Inverter Factor)

'Solar power battery cost' is a hot keyword. Everyone focuses on $/kWh. Lithium iron phosphate is getting cheaper. Lead-acid is still around for some off-grid sites.

But if you are installing a solar + battery system at a high altitude site—say a 3,500m mining camp—your inverter/charger cost is going to surprise you. A 15kW inverter that costs $3,000 at sea level might need to be a 25kW inverter at 4,000m to deliver the same continuous AC power. That $2,000 price jump is a 66% increase in the inverter cost. Compare that to the battery savings from 'cheap' LFP cells.

I've seen projects where teams optimized the battery procurement beautifully—locked in a great $0.15/Wh deal on LFP cells—and then installed a standard inverter rated for sea level. The inverter couldn't run the load at site. They had to spend $4,000 on a replacement derated unit and pay $1,500 to ship the original one back. The 'solar power battery cost' win got completely swallowed by the inverter replacement cost.

My rule now: At 3,000m+, calculate your inverter/charger budget as 1.5x the sea-level cost before you even look at battery pricing. The battery cost is irrelevant if your inverter can't use it.

Bidirectional Converters and DC Microgrids: A Specific Warning

If you're working on a DC microgrid (380V DC bus with battery storage and renewable sources), the bidirectional converter is the heart of the system. It needs to charge the battery from the DC bus and discharge the battery to the DC bus.

I've now tested three different 'bidirectional' units from different manufacturers at simulated altitude. Here is a pattern I observed:

Directional bias in design. Many converters are electrically bidirectional but thermally optimized for one direction. Usually, charging (grid-to-battery or bus-to-battery) runs cooler because the power flow is 'downhill' voltage-wise. Discharging (battery-to-bus) has to boost voltage, which stresses the semiconductor switches more, creating more heat. At altitude, that extra heat from discharging is exactly where the thermal margin gets eaten up.

In our failed project, the converter could charge the battery bank at 4kW (at 4,800m). It could only discharge at 2.2kW. This created an imbalance in the system—the generator + solar could fill the battery, but the battery couldn't feed the load fast enough when the solar dropped. We had to manually balance the load. Terrible for an autonomous microgrid.

Ask your supplier: 'What is the derating curve for both charge and discharge directions at altitude?' If they can't give you both, look for a different product. You're buying a bidirectional converter, not a unidirectional one with a reverse mode checkbox.

The Checklist I Use Now (And You Should Too)

After that $3,200 mistake, I built a pre-order checklist for high-altitude systems. It's not elegant, but so far it's caught 12 potential errors in the past 18 months:

1. Get the full derating curve (not just continuous). Ask for the curve for peak power (10 sec, 60 sec) and surge (1 sec) as a function of altitude. Do not accept a single curve for 'power rating.'
2. Run the thermal math with air density. A fan at altitude moves less mass. Even if the electrical efficiency looks okay, the heat gets trapped. Check if the unit uses fan cooling or is convection-cooled. For 4,000m+, expect that a fan-cooled unit loses 30-40% of its effective cooling capacity.
3. Test at altitude if possible. If your site is above 3,500m and the order is >$5,000, find a way to test the unit in a hypobaric chamber. We have one at a local telecom lab. It costs $200/hour to rent. It has saved me many more times in rework costs.
4. Oversize the DC-DC converter by 40% for bidirectional operation. If you need 5kW in both directions at 4,500m, buy an 8kW unit rated for sea level. This is not overdesigning; it's accounting for reality.

When You Can Ignore This Advice

Honestly? If your site is below 2,500m (8,200ft), and you're using a reputable brand like a Mean Well or a high-end power supply with detailed datasheets, the standard '5% derating for 1,000m' rule works fine. Most residential solar installs in, say, Denver (1,600m), don't need this level of paranoia.

Also, if you're using a generator (say a Kohler standby generator) directly feeding a standard AC panel, the altitude derating is simpler and generator manufacturers provide clear, tested curves. The complexity spike comes when you introduce a DC-DC converter with tight efficiency optimization, especially from a supplier who may not have done rigorous altitude testing on their full range.

And if you're working with a China high-efficiency power supply that costs 40% less than the competition? I'm not going to say 'don't buy it.' I buy them too—the cost savings are real for many projects. Just budget for 30% derating at altitude and do your own verification. The datasheet is a starting point, not a guarantee.

Prices as of early 2025; verify current rates with suppliers. Altitude derating data is from my own testing with specific equipment models; your results may vary. Consult the manufacturer for your specific unit.