Friday, 28 May 2021

Why are Solar PV Panels Getting Bigger?

Power Games

The rated power of solar PV panels has climbed steadily over time.   This has been driven in large part by innovative new processing techniques for the cells themselves, although improvements to the technology of panel assembly has also played a role.  Over the decade from 2010, customers of the panel manufacturers came to expect higher and higher module powers each year.

Because all panels were the same size, the panel power was a good shorthand measure for how advanced the cell technology was.  If a panel was rated at 320Wp then it would generate 14% more energy per square metre of space than a 280Wp module.

Squeezing more power (measured in Watt-peak - or Wp per panel) into the same footprint tended to drive down the cost per installed unit of rated power ($/Wp).  Since the cost of the glass, frame and other components of the module and all the installation materials remained the same for modules of any power.  The inexorable rise in power density of the available cells was a significant factor in helping the industry achieve the amazing feat of cost reductions we have seen over that decade.

In addition, new techniques for squeezing every drop of performance out of the cells in solar modules have also allowed the accumulation of small gains.  Using a greater number of conduction wires (bus bars) with a more slender width on the top face of the cells reduces power lost due to shading of the cell below and also reduces losses from electrical resistance.  Cutting the cells into halves, or thirds, or quarters and wiring all these fragments together into parallel circuits again reduces resistance losses as well as reducing the sensitivity of the module to shading.  The efficiency gap between measuring the cell in isolation compared to the assembled panel has reduced over the years.

However, gains from improving cell powers have reached a plateau.   See my blog on why solar cells are not getting more powerful (coming soon).

So manufacturers reaching for new ways to keep the story of ever more powerful modules at ever lower cost per Wp going have found a simple answer - just make the cells and the panels bigger. 

Breathless excitement from credulous industry commentators as announcement of modules exceeding 400Wp, then 500Wp barriers misses the point.  A not-so sleight of hand is evident as soon as you look at the product behind the headline number.  Panels are not getting better, they're just getting bigger.

You can have any (size), so long as its....

When it came to solar PV panels (modules) we all used to know where we stood.  A solar PV panel was just under 1m wide and around 1.65m long.  It had each of its 60 cells were 156mm square.  A defacto standard for PV panels emerged around 2010 and manufacturers stuck to it.  

Using the Waybackmachine internet archive I accessed historic web pages of industry stalwart Trina Solar.  For each year where I could find the information, I downloaded product information for the highest power module that was being offered for sale at that time.  See the graphic at the top of the page.

You can see that from 2009 to 2018, its module powers increased from 230Wp to 315Wp while their size remained the same.  The corresponding module power density increased from 141 to 192Wp/m2.

Today Trina Solar offers modules with powers ranging from 320Wp to 600Wp by virtue of offering modules of much larger dimensions.

There were many advantages that came from working to a standard size.  

In manufacturing - everything from the sheets of glass to the width of the encapsulant film and backing sheet to the pallets and packaging could all be made to the same size.  Economies of scale for the whole industry drove down costs and raw materials are readily available from many different suppliers.

Downstream, manufacturers of so-called 'balance of materials' equipment could also standardise their products.  

The electrical characteristics of the modules were all in a relatively small range and manufacturers of module level power electronics (MLPE) such as power optimisers and microinverters could easily cover the whole market with a small range of products.

Since the physical dimensions of modules varied so little between models, manufacturers of roof mounting kits also benefitted.  The wind loading per module was the same for all modules, allowing optimised designs for fixings.  The spacing of point loads on rails was consistent for any design.  Roof integrated BIPV products emerged that worked with these standardised modules and relied on their predictable format for wide interoperability and a single wind resistance value (though fire classifications proved less simple).

When designing with standard sized modules, solar installers had confidence that the system they were designing would be available both now and into the future.  This is especially important for new build projects where the design may be priced for construction many months or even years in the future.  Even if the specific module selected with was no longer available, an alternative of the same dimensions could easily be found.

Finally customers benefitted from the knowledge that in the event that a single panel in a system was to fail, that a replacement of the same format would be available in future and could simply drop into the mounting system.

Battle of the Formats

A standards war has broken out with TrinaSolar, Risen Energy, Canadian Solar and others lining up against Hanwha Q Cells, REC Group, LG on the one hand and Longi, Jinko and others on the other  - each calling for the industry to focus on cells of the dimensions they prefer.  Will the industry settle on 158.75, 166, 182, or 210mm?  At this point nobody knows.  

(See my guide to different PV cell size formats here)

In any case the cell size no longer quite defines the module size.  Since manufacturers have started slicing up cells into smaller sub-units (half-cut , third cut cells)  there is greater freedom to choose different module sizes.  Dizzy with this new found freedom, module designers are expressing their creativity  with the results of a  a wide range of module sizes now available in the market (see graphic).

Even within single manufacturers a range of many different panel sizes are now on offer.   The graphic shows a sample of products available from four of the larger industry players.  Panel powers range from 320Wp to 800Wp, but as can be seen the power density (Wp/m2) ranges only from 193 to 212 Wp/m2.

This is because the cells are pretty much the same but the packing efficiency is ever so slightly higher in a larger panel (because the edges are a smaller proportion of the whole).

An arms race for 'plus-sized' size panels is in progress, with JA currently in the lead offering the Jumbo Blue at 2.2m x 1.8m and 800Wp.  These panels are completely impractical for rooftop applications, but are aimed instead at utility scale ground mount.

It seems that module manufacturers have decided that the cost and convenience benefits of standardisation of module sizes can go hang.  Instead they are pursuing a product strategy of  scaling up panel dimensions to increase panel power that is leaving the rest of the solar supply chain scrambling to catch up.

Where will this end?  There are clearly practical limits.  How big a module can customers handle?  The answer to this question depends on where you are trying to install it.  If you're on top of a windy roof you don't want to be humping round a 4m2 monster, but maybe that's perfectly fine with mechanised lifting for a ground mount system.

The economics of solar PV has changed.  Modules now dominate the cost structure of a solar PV installation to a much lesser extent.  While it was the case that the module was the principal cost, it made sense to work with a standard format and drive out as much cost as possible.  As panel prices have fallen, maybe it does make sense to design specialist modules for different jobs.  Naturally manufacturers are focussing on their most important markets - solar farms and hence the rush to produce the bigger formats, which reduce other costs in installation and clamps.

Could this trend extend beyond different panel sizes to panels with special design features for different applications?  Maybe in future the panels used for flat roofing installations will be different from those used for fixing to corrugated metal roofs and have special features that make installation easier?  Or maybe the industry will step back from format wars and settle on a few standardised sizes for different applications.  Some folks in the business of making mounting kits for the solar modules made by these manufacturers must certainly be hoping so!

Friday, 5 March 2021

PV Cell Formats and Size Guide


Here's a handy diagram I created to help show the difference between all the new solar PV cell formats in the market right now.

Monocrystalline cells are made by slicing across a cylindrical ingot of silicon.  The least silicon waste is created by having perfectly round cells, but these don't pack very neatly into a solar panel (or module), leaving gaps between the cells which reduce the power output of the panel compared to one that fills the area more effectively.  This trade off between panel power and cell economics has typically resulted in a compromise where mono-crystalline cells have rounded corners.

As silicon prices have fallen, the economics has moved decisively in favour of cutting to make a cell with a more perfect square shape, driving up the panel power by having all of the available area as working photovoltaic material.

After a long period of standardisation on the M2 cell format of 156.75mm, manufacturers cannot agree on a standard size going forward, with each proposing a slightly different format, and of course this means that the finished solar PV modules that the cells are assembled into also differ in size.

In what may come to be seen as a collective failure , and at least for the next few years, the industry has thrown away all the downstream benefits that come from having standard module sizes, a topic I will come back to in a future blog.


Tuesday, 2 March 2021

Do We Need to Talk About Fugitive Emissions from Heat Pumps?

 Could the Benefits of Future Building Standards be Affected?

It’s not a recent study, but its contents may have become newly relevant.  In March 2014, the then Department for Energy and Climate Change (DECC) published the results of a study commissioned from Eunomia Research and London Southbank University - “Impacts of Leakage from Refrigerants in Heat Pumps”.  

Governments are looking for a huge increase in the number of heat pump installations in an attempt to decarbonize the heating of buildings.  Leading the charge will be developers of new homes, since these are the easiest to legislate for.  The UK government has unveiled its Future Homes Standard and Scottish Government is consulting on a New Build Heat Standard.  Neither of these regulations would force developers to install heat pumps in the homes they build, but the standards will be set in such a way that fossil-fuel powered boilers will not comply - the clear intention is to use regulation of the new-build sector to scale-up the heat pump industry.

The approach being considered by Scottish Government is to require heating systems to emit no carbon dioxide at the point of use.  Technologies considered to meet this requirement are direct electric heating (panel heaters, underfloor electric, electric boilers), heat pumps and also heat networks (where you could be burning coal at the far end of the pipe but this doesn't count as point of use).

But since the goal of all of this is to reduce global warming, then surely the emissions of other gases that cause global warming should also be considered.  It turns out that the refrigerant gases used inside some heat pumps are greenhouse gases that are thousands of times more powerful than carbon dioxide.  So long as they stay safely inside the heat pump during its life and are collected up at end of life, there is no problem, but the question the report set out to answer was how much of this refrigerant gas is lost to leaks, and how big an issue could that be?

The researchers examined the logbooks of heat pumps to see how often they had to be re-charged with refrigerant, and also conducted experiments where heat pumps were emptied and then recharged with refrigerant gases and the amount lost in the process was measured.

How Much Refrigerant is Lost?

The report found that:

  • 10% of domestic heat pumps experience a leak of refrigerant in any given year
  • The median amount lost was 35% of  charge (the total refrigerant gas in the system)
  • So the equivalent annual leakage rate was 3.48% of the charge
  • The mean refrigerant charge for a domestic heat pump was 3.3kg
  • So the mean annual leakage for a domestic heat pump is 0.114kg of refrigerant
  • Decommissioning losses at end of life were 15% of charge 
  • Commissioning and recharging losses were around 0.06kg
So if we assume a 20 year life for the heat pump, a charge at commissioning, one recharge and decommissioning at the end of life then the total losses are:

Filling: 0.06kg 
Recharge: 0.06kg
Leakage: 10% x 35% x 3.3kg x 20 years = 2.31kg
Decommissioning: 15% x 3.3kg = 0.495kg

Total Refrigerant Losses over 20 year life: 2.925kg

Refrigerant loss per year: 0.146kg

Does it Matter?

At the time of the report, R410A was the most used refrigerant.   It has a Global Warming Potential (or GWP) of 2,088, which means that 1kg of R410A released into the atmosphere has the same effect on global warming as if 2.088 tonnes of carbon dioxide were emitted.

So, a refrigerant leakage of 0.146kg per year from a heat pump is like it emitted 305kg of CO2 per year.  This is expressed as 305kgCO2e (CO2 equivalent)

The authors of the report compared the impact of refrigerant leakage to the benefits of the low-carbon energy delivered by the heat pump.  The annual heating load for a domestic property was assumed to be between 10 and 25 thousand kWh and it was concluded that the impact of leakages was small. 
However, the leakage losses from the heat pump does not vary with the amount of heat it supplies.  The more heat the heat pump delivers, the less of a penalty the refrigerant leakage is per unit of heat delivered.  Conversely the more energy efficient is the building and the lower its heat demand, the more of a penalty the leakage becomes per unit of heat.

Since the regulations are aimed at new homes, perhaps we should instead consider refrigerant emissions in relation to the heat demand of these.

Homes that are built today have very much higher levels of airtightness and insulation than the average building stock.  Based on energy calculations we see in our design team, an average sized new build house (say 85m2 total floor area) will have space heating demand around 3,300kWh per year to which you can add 1,600kWh per year for domestic hot water.  The proposed standards envisage tightening rules on insulation too, so by then buildings will have an even higher level of thermal efficiency than this.

Many houses are now being built to ‘passivhaus’ standards which remain at a comfortable temperature most of the year without heating, so the annual heating demand approaches a minimum which is that needed for hot water for showers and baths.

If we divide our 305kgCO2e figure by these lower figures for annual energy demand, the contribution to global warming per unit of heat delivered looks very different.

Table 1: Release of refrigerant gases per  kWh heat delivered (CO2e)

Annual Heating Budget (kWh)






Gas Heating CO2 emissions for comparison

Annual Emissions from Refrigerant Leakage per kWh of heat delivered (kgCO2e)








The table shows the global warming impact of refrigerant leakage for each unit of heat provided by a heat pump running R410A for different levels of annual heat demand.

It can be seen that for buildings with high heating demand (such as the levels assumed by the report authors), the CO2e figure is very low compared to delivering the same heat with gas heating.  However for very highly energy efficient properties the penalty from the refrigerant leakage rates is very significant – approaching that of gas heating.

Current Refrigerants in Use

Of course, we would expect heat pump technology to have moved on in the last six years and with regulations aiming to phase out the use of the refrigerants with the highest GWP, maybe R410A is no longer in use?

I did a quick straw-poll of the products available today from a range of heat pump manufacturers.  As you can see from the table below, there are clearly now models available that work with refrigerant with much lower GWP than the R410A used in the calculation above.  Notable leaders are products from Vaillant using R290 (which is propane gas and has a GWP of 3) and from Mitsubishi using R744 (which is carbon dioxide and so has GWP 1) - although both these manufacturers still offer models with the much higher GWP refrigerant for some reason, perhaps cost or efficiency?

It is also clear that there are also many, many models still being sold today that are using refrigerants with GWP above 1,000 and that R410A is still very popular.









Altherma 3 H HT - EPRA014-018DW




Altherma 3 R - ERGA04-08EVA




Daikin Altherma - EDLQ-CV3




Altherma R - ERLQ-CV3












Ecodan R744




Ecodan R32




Ecodan R410A




Aquarea HT Bi-bloc F Generation




Aquarea T-CAP Bi-bloc H Generation




Aquarea High Performance All in One Compact J Generation




Arotherm plus




Arotherm Split









Based on the report, it appears to me that the level of “Fugitive emissions” from heat pumps is significant enough that heat pumps with refrigerant of high GWP  should not be considered as zero carbon (equivalent) at the point of use.  

This is especially the case when the heat pump is providing heat to a highly energy efficient new building and the emissions ‘cost’ is spread over a much smaller heat ‘benefit’.

Perhaps legislators should consider the global warming potential of refrigerant in heating systems when they create new building standards?  Since heat pump models are available with improved refrigerants, perhaps they should consider applying a ceiling GWP value to the refrigerant in use?

I would be interested to hear from colleagues in the heat pump industry about whether this issue has been much discussed, and what the future direction of travel in heat pump technology is to address the likelihood or impact of leakage of refrigerant - please comment below!

Friday, 22 January 2021

Comparing Building Regulations for Energy - Scotland and England


Development of new homes in Glasgow by CCG Homes, Installation by Arc-Tech

New regulations have been published for Part L building regulations in England.

Let's take a quick look at how they compare to Scotland, where almost all new homes now come with solar panels.

The building regulations work by defining a specification for a 'Notional Dwelling' - read this article for more detail of how it works

In the table below I have shown a selection of the more significant requirements for each.  It is evident that not only do both Notional Houses include solar PV, but that the England requirement is much higher.

Housebuilders do not have to build the Notional House specification, but they must equal or exceed its performance.  The scope for reducing the solar provision by increasing the insulation performance of building elements such as walls, roofs and windows (collectively often referred to as fabric measures)  is limited in the new English regulations, which equal or exceed the fabric requirements in Scotland.

‘Notional Dwelling’ Summary Specification

Selected Requirements


Scotland 2015 (Gas Package)

England 2021

U Values in W/m2K



External Walls















Air Permeability

7 m3/h.m2 at 50Pa

5 m3/h.m2 at 50Pa


Gas boiler 89% efficiency

Gas boiler 89.5% efficiency

Heat Emitter Type

Regular radiators

Low temperature heating 55°C flow

Heat Recovery

Instantaneous Waste Water 45% efficient

Instantaneous Waste Water 36% efficient

Solar PV in kWp

Smaller of

(a) Total floor area (m2) x 0.01


(b) 0.3 x roof area based on 30 degree roof pitch x 0.12



40% of dwelling floor area/6.5

Example solar PV for 85m2 house with two floors