PPU Fundamentals
PPU Efficiency, Mass and Cost: The Engineering Trade-offs
A PPU datasheet leads with a unit price and an efficiency number. Neither tells you what the unit will cost your mission. This walks through how PPU efficiency turns into spacecraft mass, and how to compare suppliers on total cost.
On this page
Every lost watt becomes radiator mass
A power processing unit that is 90 percent efficient at 300 W of output is dissipating about 33 W as heat inside the box. At 80 percent it is dissipating 75 W for the same useful output. In a laboratory that difference is a slightly warmer heatsink. On a spacecraft it is structural.
A satellite can only shed heat by radiating it to space. The waste heat from the PPU has to be conducted to a radiator panel and rejected there, and radiator area is not free: it is panel mass, deployment mechanism mass in some cases, and area that competes with solar arrays and antennas. As a rough planning figure, spacecraft thermal engineers often size passive radiators in the region of tens of watts rejected per kilogram of radiator hardware, so an extra 40 W of PPU loss can add on the order of a kilogram to the platform before margins. The exact figure depends on orbit, pointing and sink temperature, but the direction is always the same: lower PPU efficiency means more spacecraft mass.
That mass then propagates. More radiator and more thermal hardware means less propellant or payload for a given launch mass, or a more expensive launch slot. This is why efficiency is treated as a system parameter, not a component nicety.
The efficiency, mass and cost triangle
PPU design pulls in three directions at once, and improving one usually costs you one of the others.
- Efficiency. Higher efficiency reduces waste heat and therefore radiator mass, and it reduces the demand on the solar array. It usually costs money: better semiconductors, tighter layout, more design iterations.
- Mass. The PPU's own mass matters directly, and it is dominated by magnetics, capacitors and heat-spreading structure. Pushing frequency up shrinks the magnetics but raises switching loss unless the devices and topology are chosen for it.
- Cost. Space-qualified parts, screening and test dominate recurring cost. A cheaper unit often uses a heavier, less efficient topology or a less rigorous parts programme.
The reason this is a genuine trade and not just an optimisation is that the space supply chain has historically offered efficiency and low mass only at the top of the price range. Breaking that link is a device and architecture problem.
How GaN and switch-mode topologies shift it
Gallium nitride transistors change the trade because they switch faster and with lower loss than silicon at the same voltage and current, and they carry roughly three times the power density. That has three effects on the triangle:
- Efficiency up. Lower conduction and switching loss directly raises converter or amplifier efficiency, which is the whole point for radiator mass.
- Mass down. Higher switching frequency becomes practical, which shrinks inductors, transformers and filter capacitors, the heaviest parts of the unit. Higher power density shrinks the semiconductor and its thermal interface.
- Cost pressure. GaN device prices have fallen sharply as industrial and automotive volume has grown, and a smaller, cooler design uses less structure and less qualification effort for the thermal path.
For an RF power processing unit specifically, a resonant switch-mode amplifier such as a Class-E stage on GaN is what makes a conversion efficiency target above 90 percent at 13.56 MHz credible. A linear amplifier at that frequency would waste a large fraction of the input as heat. The space-grade RF generator page covers the amplifier stage, and the adaptive matching network page covers why holding the match is also an efficiency question: reflected power is wasted power.
Total cost of ownership, not unit price
The right comparison between two PPUs is not the quoted unit price. It is the price plus the mass penalty the unit imposes on the spacecraft plus the risk it carries over mission life.
| Cost element | What drives it | Why it is easy to miss |
|---|---|---|
| Unit price | Parts programme, screening, test, volume | It is the only number on the quote |
| Mass penalty | PPU mass plus radiator mass from its losses | It lands on the platform budget, not the propulsion budget |
| Array penalty | Extra solar array to cover conversion loss | Shows up as a power-budget shortfall late in design |
| Integration cost | Bus interface fit, thermal interface, EMC | Non-recurring engineering, often underestimated |
| Schedule and supply risk | Lead times, export licences, single-source parts | Only visible when a part slips or a licence is refused |
| Mission-life risk | Derating, redundancy, fault tolerance | Priced as reliability, paid as lost mission value |
A unit that costs less up front but is five points less efficient and relies on export-controlled parts can easily be the more expensive choice once the radiator mass and the licensing risk are counted. Export exposure in particular is a schedule and availability question, not just a paperwork one: see ITAR-free RF power processing units.
What to ask a PPU supplier
- What is the efficiency at my operating point, not just at full rated power, and how was it measured?
- What is the mass, and how much of it is magnetics and heat-spreading structure?
- How much waste heat must I reject, and at what baseplate temperature?
- What is the bus voltage range it tolerates without an added front-end converter?
- Is the bill of materials free of ITAR and other export-controlled parts? Can you supply the classification?
- What is the supply chain: which parts are single-sourced, and what are the lead times?
- What is the TRL, and what qualification has actually been done versus planned?
- How does the unit behave on a fault: reflected power, arc, over-temperature, bus transient?
The MHz Labs design targets
MHz Labs is developing an RF power processing unit and adaptive impedance matching network for RF and helicon plasma thrusters, sold to propulsion OEMs rather than to satellite operators. The unit is on a TRL 4 to 6 roadmap and has no flight heritage yet. The design is built around the efficiency-mass-cost trade described above:
| Parameter | Design target |
|---|---|
| Conversion efficiency | Not published while unmeasured |
| RF output power | 50 to 500 W |
| Operating frequency | 13.56 MHz ISM |
| Power stage | GaN switch-mode, Class-E family |
| Matching | Real-time adaptive network |
| Unit cost | Targets the DEEP-PPU benchmark, roughly 35 to 50 percent below legacy |
| Supply chain | Fully European, no ITAR parts |
If you are scoping a propulsion programme and weighing PPU options, the market context and the full RF PPU guide give the background, and the team can share the current design targets and integration guidance directly.
Frequently asked questions
How much does PPU efficiency really affect spacecraft mass?
Enough to matter at system level. Every watt the PPU dissipates must be conducted to a radiator and rejected to space, and radiator hardware rejects only tens of watts per kilogram. A few points of efficiency on a several-hundred-watt unit can move the platform mass budget by around a kilogram before margins.
Why is a lower-priced PPU sometimes the more expensive option?
Because the quote only shows the unit price. A less efficient, heavier unit adds radiator and array mass to the spacecraft, and export-controlled parts add schedule and licensing risk. Those costs land on other budgets but they are still your costs.
Does GaN make a PPU more expensive?
Not necessarily any more. GaN device prices have fallen with industrial and automotive volume, and the smaller, cooler design that GaN allows uses less magnetics and less thermal structure, which offsets the device cost.
What efficiency should I expect from an RF PPU?
It depends on frequency, power level and topology. A resonant switch-mode amplifier on GaN at 13.56 MHz can be designed toward a conversion efficiency above 90 percent for the RF stage. Ask any supplier how and where they measured the number.
Is the MHz Labs PPU available to buy now?
No. It is in development along a TRL 4 to 6 roadmap with no flight heritage. Design targets, engineering engagement and integration discussion are open now.
Sources and further reading
- Goebel and Katz, Fundamentals of Electric Propulsion, NASA JPL DESCANSO, for power-budget and thermal context.
- SatNow, What is a power processing unit (PPU) for thrusters.
- EU, Regulation (EU) 2021/821, the dual-use export control regime relevant to supply-chain risk.
Comparing PPU options for a propulsion programme?
MHz Labs builds the RF power processing unit and adaptive matching network to a published set of efficiency, mass and cost targets. Request the current figures and integration guidance.
Get in touch