Efficiency modules need a downside

[Photoloss]

That might actually work to improve Eff2+3, but you vastly overestimate the effect on T1. Instead of adding modules you can always build more solar panels to achieve the same net effect, and those don't have any diminishing returns. I also vehemently disagree with nerfing vanilla mechanics for the sake of mods, if the power and pollution reduction on refineries and furnaces is considered "acceptable" then it should stay and modders who want their 10MW super-machines should fix their own balance problems.

Sadly your suggestion faces the same consistency problem as the previous one: make power penalties multiplicative and break them completely, or keep them additive and confuse players with unintuitive mechanics. If speed+prod stacking is considered balanced right now multiplicative stacking would render it utterly useless.

[BlakeMW]

That might actually work to improve Eff2+3, but you vastly overestimate the effect on T1. Instead of adding modules you can always build more solar panels to achieve the same net effect, and those don't have any diminishing returns. I also vehemently disagree with nerfing vanilla mechanics for the sake of mods, if the power and pollution reduction on refineries and furnaces is considered "acceptable" then it should stay and modders who want their 10MW super-machines should fix their own balance problems.

Mods are besides the point, it's just more obvious.

The benefit of an eff module is precisely proportional to the kW usage of the machine. An eff1 module is twice as effective in a machine which consumes 420kW than one which consumes 210kW. This is a benefit straightforwardly compared with solar/accu. "Eff1 usage 101" is to always put the eff1 module into an unused slot in the machine with the highest power usage, this is you could say embarrassingly straightforward strategy.

With speed and prod it's more complicated, because while a more expensive machine generally multiplies the power of the module, it also multiplies the downsides. Also in some cases even the bonuses can come with disadvantages, for example when you pile a lot of speed onto an assembler often the belts/inserters can't keep up. Eff modules have no downsides whatever, whether or not to use eff modules simply comes down to comparative advantage - do you get more power out of investing resources in an eff module, or in solar/accu, it's not even really a strategic choice, it's just very simple economics as to whether it is worth using or not.

Sadly your suggestion faces the same consistency problem as the previous one: make power penalties multiplicative and break them completely, or keep them additive and confuse players with unintuitive mechanics. If speed+prod stacking is considered balanced right now multiplicative stacking would render it utterly useless.

I agree it's needlessly complicated.

Another idea, would be to make the eff modules more squarely about pollution reduction than energy consumption reduction. This is less straightforwardly compared with power generation by solar/accu, it involves a completely different set of calculations and a mathematical analysis to say which is better is much harder.

Eff1: -15% Energy Consumption, -20% Pollution
Eff2: -25% Energy Consumption, -30% Pollution
Eff3: -50% Energy Consumption, -40% Pollution

Now the eff1 module would be pretty weak in anything except polluting machinery - it'd still reduce pollution by the same amount in Electric Mining Drills (with those numbers, to 22%) but the energy bonus wouldn't be anything to write home about. At this strength I'd estimate they would be about as attractive as speed1 or prod1 modules - not very, but you may use them in certain circumstances.

The eff3 module would provide a huge pollution reduction to speed and prod/speed setups. You could use Prod3/Speed3 Beacon to maximize production and minimize payoff time, or you could use Prod3/Speed3+Eff3 Beacons to still enjoy the benefits of speed boosted productivity, and also reduce pollution - although because the eff3 modules are competing with speed3 modules for beacon slots, they are having a significant impact on how productive the setup is, you slow the evolution of the aliens, but also slow yourself down. But it gives an actual strategic choice - which really is what this is about, having an actual reason to use eff2/eff3 modules which makes sense from a strategic perspective.

[Photoloss]

Mods are besides the point, it's just more obvious.

The benefit of an eff module is precisely proportional to the kW usage of the machine. An eff1 module is twice as effective in a machine which consumes 420kW than one which consumes 210kW. This is a benefit straightforwardly compared with solar/accu. "Eff1 usage 101" is to always put the eff1 module into an unused slot in the machine with the highest power usage, this is you could say embarrassingly straightforward strategy.

With speed and prod it's more complicated, because while a more expensive machine generally multiplies the power of the module, it also multiplies the downsides. Also in some cases even the bonuses can come with disadvantages, for example when you pile a lot of speed onto an assembler often the belts/inserters can't keep up. Eff modules have no downsides whatever, whether or not to use eff modules simply comes down to comparative advantage - do you get more power out of investing resources in an eff module, or in solar/accu, it's not even really a strategic choice, it's just very simple economics as to whether it is worth using or not.

While the calculations are much harder you can still ultimately put a cost to things like needing to upgrade inserters, density loss from having to build a second belt and so on. Pollution is rather complex and many of its effects have massive delays, but this can also make a balanced module feel very unrewarding, there's no flashing "time until you get overrun by biters" that goes up as you use more modules.
Efficiency Modules do have some more complexity to them though: since not all of your factory will be running 24/7 you must locate the machines with the highest effective consumption/pollution rather than just reading the base stats. Solar Panels also satisfy any energy demand in your base while the modules only affect the specific machines you place them in, when your production priorities shift you must also move the modules.
Your assessment also neglects the decision-making based on "Eff vs nothing" and "Eff vs other modules". If you have an Efficiency Module of course you should place it in the greatest consumer/polluter, but perhaps you could make even greater gains by producing a Speed Module in the first place. You might also sacrifice efficiency and use the modules over cheaper solar+accu "at a loss" to keep your total pollution in check if a specific production sequence is so critical you want to maximise the raw output.
Efficiency Modules certainly are a simple, boring "fallback" option but their existence adds meaningful complexity to the overall gameplay.

I agree it's needlessly complicated.

Another idea, would be to make the eff modules more squarely about pollution reduction than energy consumption reduction. This is less straightforwardly compared with power generation by solar/accu, it involves a completely different set of calculations and a mathematical analysis to say which is better is much harder.

Eff1: -15% Energy Consumption, -20% Pollution
Eff2: -25% Energy Consumption, -30% Pollution
Eff3: -50% Energy Consumption, -40% Pollution

Now the eff1 module would be pretty weak in anything except polluting machinery - it'd still reduce pollution by the same amount in Electric Mining Drills (with those numbers, to 22%) but the energy bonus wouldn't be anything to write home about. At this strength I'd estimate they would be about as attractive as speed1 or prod1 modules - not very, but you may use them in certain circumstances.

The eff3 module would provide a huge pollution reduction to speed and prod/speed setups. You could use Prod3/Speed3 Beacon to maximize production and minimize payoff time, or you could use Prod3/Speed3+Eff3 Beacons to still enjoy the benefits of speed boosted productivity, and also reduce pollution - although because the eff3 modules are competing with speed3 modules for beacon slots, they are having a significant impact on how productive the setup is, you slow the evolution of the aliens, but also slow yourself down. But it gives an actual strategic choice - which really is what this is about, having an actual reason to use eff2/eff3 modules which makes sense from a strategic perspective.

How would pollution reduction stack (both with itself and with power-draw-scaling) and what's the cap? And while I don't have any hard numbers by the time you can afford T3 beacon arrays you should be able to easily handle the very localised pollution output of these setups.

Now, I'm not saying the current modules are perfect. But in my eyes a full rework of power generation and combat is much more important right now, and since Efficiency Modules critically depend on these two aspects of the game they should be rebalanced afterwards unless someone comes up with a very simple, very meaningful improvement. If biters become smarter pollution reduction might become a trivial "best in slot" upgrade even without the power reduction, if we get an endgame fusion reactor T3 modules might become completely meaningless instead.

[bobucles]

How would pollution reduction stack (both with itself and with power-draw-scaling) and what's the cap? And while I don't have any hard numbers by the time you can afford T3 beacon arrays you should be able to easily handle the very localised pollution output of these setups.

The current system causes pollution to scale directly with the structure's energy demand. Reducing energy use directly reduces pollution, and increasing energy demand obviously increases pollution.

You can ALSO reduce the pollution multiplier separately, which some mods play around with. Both attributes work together to make structures have nearly 0 pollution output, but it's totally overkill.

[Photoloss]

The current system causes pollution to scale directly with the structure's energy demand. Reducing energy use directly reduces pollution, and increasing energy demand obviously increases pollution.

You can ALSO reduce the pollution multiplier separately, which some mods play around with. Both attributes work together to make structures have nearly 0 pollution output, but it's totally overkill.

Yeah, I meant that pollution-only factor in the context of your suggestion. As in, would 2xEff3+Eff1 reduce a machine to 0 pollution? Would it be capped at 20% base? Would the 20% caps on power and pollution stack? How does the reduction interact with increased power consumption?
I do not expect you to give concrete answers to how all these things should interact, my main point is that the detailed mechanics are not made visible ingame. Ignoring the actual balance it would be incredibly hard to even figure out when using such modules would make objective sense, and you still face the risk of T3 modules being near worthless since the effects of pollution can be combated in other ways.

If the mechanics are made clear and everything is decently balanced I agree this line of thinking makes the game more interesting (still not entirely sure how much better it would be in practice), but as I already said getting there would take a lot of effort and I would like to see combat and power generation get an overhaul. That would essentially make all previous balancing work go to waste.

[BlakeMW]

I was just thinking of using the current mechanisms for pollution modification, which operates on the final energy usage. I don't know if there is a minimum cap or not. Unlike energy use, it might actually be reasonable to have no minimum pollution level. It would cost a lot to zero out your pollution as you'd have to use beacons to zero out the pollution on machines with 2 module slots, it would mean biters don't launch attacks anymore (but you'd still get migrating biters wandering in), the evolution factor would continue to increase just due to the time factor, so it doesn't change the game much over having a factory with very low pollution.

As long as zeroing out your pollution costs more than walling the biters out with impregnable defenses I really don't see a problem, it'd make for an interesting achievement.

[Frightning]

Klonan's right on the money on pg 1 when he said that the downside to Efficiency Modules is opportunity cost. They cost resources to make and a module slot to use, just like any other module, and because of the later, you giving up the option to get Productivity (sometimes) or Speed (always) instead. (and you still spent resources on it too!) I actually don't see much reason to make Efficiency Modules personally: if I'm going to spend the resources to make a module I want a return on that investment, and energy, as it currently stands is easy to come by, with the only significant cost being the fixed cost of creating the Solar panels and Accumulators (and Substations/Power poles) required. So I don't consider the energy savings all that valuable. Likewise, I don't see the Energy cost on Speed modules as all that much of a downside. Productivity modules have a LOT of downsides because their upside is uniquely valuable: Only way to actually get more out of the resources you've mined (you can even use them in Electric Mining Drills to increase the yield of your raw resource deposits), thus they understandably have significant downsides, but also, synergize nicely with Speed modules from Beacons as energy cost per item made actually decreases with speed beacons (despite the beacon energy costs; and the effect is more pronounced the better utilized your beacons are).

[bobingabout]

I think the main reason personally to go with efficiency.... is to reduce pollution. Turrets can't take modules, but don't produce pollution, so that's fine, but nearly everything else with a hefty energy drain specifies pollution as how much per unit of energy. therefore to reduce pollution, you reduce energy consumption.

I personally would like to see the pollution/energy consumption correlation broken, and have pollution specified either as an absolute per second (next to energy, rather than as part of energy), and even options to have it specified by the recipe. (So for example, in my mod, venting oxygen causes no pollution, or even negative pollution, where others like Chlorine and Sulfur Dioxide have a huge pollution value)

[Frightning]

I think the main reason personally to go with efficiency.... is to reduce pollution. Turrets can't take modules, but don't produce pollution, so that's fine, but nearly everything else with a hefty energy drain specifies pollution as how much per unit of energy. therefore to reduce pollution, you reduce energy consumption.

I personally would like to see the pollution/energy consumption correlation broken, and have pollution specified either as an absolute per second (next to energy, rather than as part of energy), and even options to have it specified by the recipe. (So for example, in my mod, venting oxygen causes no pollution, or even negative pollution, where others like Chlorine and Sulfur Dioxide have a huge pollution value)

By the time I reach the point in the game where I am actually using modules much, I am past the point of worrying about the biters. I already have that problem solved with plenty of defenses.

[bobingabout]

Who said anything about biters? I like to leave the information overlay on without having a minimap that basically just turns red and unreadable.
This also have improvements in other areas too.

[Qon]

I like to leave the information overlay on without having a minimap that basically just turns red and unreadable.

Check graphics settings.

[aeros1]

Personaly don't like current module system. Though my idea is much more complex and don't think it is as fun. But in short if modules would modify output or input of recipe as well as give bonuses based on what ingredients used, if it is to say in short. Would allow also more different kinds of intermidiates used on serial production, as well as make each module unique(1 per factory or beacon, other same have no effect) So if any one interested I might write down design document for this idea, but I don't think such boost in complexity would attract players, but would allow good sensible variety of modules.

[dragontamer5788]

If anything, Efficiency Modules 3 modules need a massive buff.

The math has been crunched: Speed Modules allow an Assembly Machine with PM3 to use less power per item crafted. To keep up with Speed Modules for the alleged "Efficiency" boost, Efficiency3 modules need ~150% energy reduction or something huge (+Speed reduces energy per item). For endgame, Productivity3 is the best module in the game, bar none. Solar energy means converting energy into extra items is the braindead choice to make. The additive effect of Speed Modules means that Speed Beacons greatly improve speed. (+50% speed really is 2.25x speed on a 4x PM3 Assembly Machine. Increases speed from 40% to 90%, that's 2.25x more items).

I dunno, maybe buff Efficiency Module 3 to -250% energy usage. That's how high it needs to be before I'd personally start considering them in a PM3 / Speed3 build. At least, with the current mechanics. The 20% minimum energy usage thing can remain there: that's fine. A bonus point of having Efficiency Module 3 have -250% energy usage is that it'd communicate the point of "Oh, these percentages are additive" to the players. So players will more naturally reach the correct conclusions when they see such an absurd number.

At 250%, Efficiency 3 modules in an Assembly3 machine still won't be able to drop PM3x3 down to 20%. (-250% drops PM3x3 from 340% to 90%). But at least Efficiency3 modules will do the job, and actually reduce energy usage better than Speed3. (PM3x3 and -250% Electricity Reduction == 125% Watts per unit)

Right now:

* PM3x3 + Speed3 Drops PM3x3 from 475% per unit to 300% Watts per unit
* PM3x3 + Efficiency3 Drops PM3x3 from 475% per unit to 405% Watts per unit.

At a minimum, Efficiency3 needs to be -130% just to TIE Speed3 in Watts per unit. And Efficiency3 can't just tie Speed3: Speed3 was cheaper to make, and also has a fancy speed bonus on top of it.

So yeah, Buff Efficiency3 to -250%, and then maybe I'll use that thing. Even then... probably not. Beacons use 480kW anyway, and no amount of Efficiency Modules seems to make sense inside of them, as there aren't any assembly machines or chemical plants in the game that use anywhere close to 480kW. The Beacon basically eats all of the efficiency gains you'd hypothetically use (especially since Beacons have no "idle" time)

[Frightning]

If anything, Efficiency Modules 3 modules need a massive buff.

The math has been crunched: Speed Modules allow an Assembly Machine with PM3 to use less power per item crafted. To keep up with Speed Modules for the alleged "Efficiency" boost, Efficiency3 modules need ~150% energy reduction or something huge (+Speed reduces energy per item). For endgame, Productivity3 is the best module in the game, bar none. Solar energy means converting energy into extra items is the braindead choice to make. The additive effect of Speed Modules means that Speed Beacons greatly improve speed. (+50% speed really is 2.25x speed on a 4x PM3 Assembly Machine. Increases speed from 40% to 90%, that's 2.25x more items).

I dunno, maybe buff Efficiency Module 3 to -250% energy usage. That's how high it needs to be before I'd personally start considering them in a PM3 / Speed3 build. At least, with the current mechanics. The 20% minimum energy usage thing can remain there: that's fine. A bonus point of having Efficiency Module 3 have -250% energy usage is that it'd communicate the point of "Oh, these percentages are additive" to the players. So players will more naturally reach the correct conclusions when they see such an absurd number.

At 250%, Efficiency 3 modules in an Assembly3 machine still won't be able to drop PM3x3 down to 20%. (-250% drops PM3x3 from 340% to 90%). But at least Efficiency3 modules will do the job, and actually reduce energy usage better than Speed3. (PM3x3 and -250% Electricity Reduction == 125% Watts per unit)

Right now:

* PM3x3 + Speed3 Drops PM3x3 from 475% per unit to 300% Watts per unit
* PM3x3 + Efficiency3 Drops PM3x3 from 475% per unit to 405% Watts per unit.

At a minimum, Efficiency3 needs to be -130% just to TIE Speed3 in Watts per unit. And Efficiency3 can't just tie Speed3: Speed3 was cheaper to make, and also has a fancy speed bonus on top of it.

So yeah, Buff Efficiency3 to -250%, and then maybe I'll use that thing. Even then... probably not. Beacons use 480kW anyway, and no amount of Efficiency Modules seems to make sense inside of them, as there aren't any assembly machines or chemical plants in the game that use anywhere close to 480kW. The Beacon basically eats all of the efficiency gains you'd hypothetically use (especially since Beacons have no "idle" time)

I think you misunderstand what Efficiency modules are useful for, their not valuable for things you need to mass produce, they're valuable for everything else you've chosen to automate. You can cut quite a bit of energy costs by running 1xspeed3 and 3xeff3 in things like your assembly machine assembler or substation assembler etc. (That setup gives +50% speed but still results in -80% energy cost, which makes it very energy efficient). The point is, multiple speed modules are a waste for things where you don't need the throughput, and most of those things are not intermediate products, so prod modules are also out of the question.

[dragontamer5788]

I think you misunderstand what Efficiency modules are useful for, their not valuable for things you need to mass produce, they're valuable for everything else you've chosen to automate. You can cut quite a bit of energy costs by running 1xspeed3 and 3xeff3 in things like your assembly machine assembler or substation assembler etc. (That setup gives +50% speed but still results in -80% energy cost, which makes it very energy efficient). The point is, multiple speed modules are a waste for things where you don't need the throughput, and most of those things are not intermediate products, so prod modules are also out of the question.

But its mass-production of components that actually takes up the hundreds of MW (or in some cases: GW) of power.

Saving the few hundreds-of-kilowatts by making Assembly Machine3 assemblers use 1/7th the energy per item is kinda meh. You're literally saving no more than ~200kW on a base that is otherwise consuming 1,000,000+kW (aka: a GW) at a time.

Saving a few hundred kilowatts is a drop in the bucket and changes almost nothing. I mean, yeah, its simply better to build out the tens of thousands of Solar Panels at this stage of the game. But I'd like to imagine that Efficiency 3 modules would have a better use.

[BlakeMW]

Eff2 module is so sad.

The Oil Refinery is handy because 10% of 420kW is 42kW, which happens to be the average output of a solar panel. So an eff1 module is equal to 3 solar panels, and an eff2 module is equal to 4 solar panels.
So you can spend 1 eff1 module to save energy equal to 3 solar panels. Not bad.

The big problem is, when you think about upgrading that to an eff2 module:
You spend 5 Advanced Circuits, 5 Processing Units and 3 Eff1 modules to save additional energy equal to 1 solar panel.

Now I want to do a little comparison, if you could get a solar panel which provides 42kW at the same cost:energy ratio, how much that solar panel would cost:

Actual Solar Panel:
27.5 Copper Plate
40 Iron Plate

"Eff2 equivalent" Solar Panel (Refinery):
322.5 Copper Plate
175 Iron Plate
93.75 Petroleum

"Eff2 equivalent" Solar Panel (Chemical Plant):
645 Copper Plate
350 Iron Plate
187.5 Petroleum

"Eff2 equivalent" Solar Panel (Pumpjack)
1505 Copper Plates
817 Iron Plates
437.5 Petroleum

Anyone willing to put an eff2 module in a pumpjack, has to ask themselves if they'd also be willing to pay 1500 copper, 800 iron and 450 petroleum for a solar panel. Because that would be an equally good investment.

[dragontamer5788]

Anyone willing to put an eff2 module in a pumpjack, has to ask themselves if they'd also be willing to pay 1500 copper, 800 iron and 450 petroleum for a solar panel. Because that's would be an equally good investment.

Good numbers BlakeMW.

Strangely enough, because of the super-cheap prices of Efficiency1, I think they're at a good spot. Its just that Efficiency2 and Efficiency3 modules are so horribly subpar, that its clear to me that they need a buff.

Maybe if Efficiency2 were -100% and if Efficiency3 were -250% (both subject to the 20% minimum), both would be usable. Any thoughts BlakeMW?

[Frightning]

The other thing not mentioned so far is that lower energy costs means lower pollution, which can be hugely valuable against aggressive biters, especially on death worlds (I've started one of these myself recently). So what you're really paying for, going eff module route instead of just more solar panels, is the lower pollution from your assemblers and other production buildings. Prod modules+speed beacons may be amazing for mass production, but that comes with absolutely MASSIVE pollution values (50+ per ore miners, 20+ per assem3, etc.) which will cause you to get attacked TONS by biters, or else have to clear out a large area and keep it clear (either by patrolling it regularly, or by walling it off, which is expensive and if you use laser turrets, energy intensive).

[Frightning]

As a followup on my previous post, lemme actually calculate pollution/unit values for production with various module types:

Before I begin, let me explain the math here a bit:
The basic idea here is that we want to measure how much pollution was generated actually making a given item, so that means we need to know both the actual speed of the assembler, accounting for yield bonus w/ prod modules, as well as final pollution/sec value, which is determined from the relevant machine's base pollution/sec value by adjusting it according to energy consumption modifier, and then adjusting that value via the pollution modifier, if any. So we have the following formulas:

final pollution/sec=(base pollution/sec)*(1+energy modifier)*(1+pollution modifier)

actual production speed=(crafting speed)*(1+speed modifier)*(1+productivity modifier)

pollution/unit=(final pollution/sec)/(actual production speed)

Where all modifiers are %s in decimal form.

For Assembly machine 3s, we have a base pollution/sec of 1.8, and a crafting speed of 1.25, hence its base pollution/unit is 1.8/1.25=1.44, using these numbers we obtain final numbers, for 4x of given module, of:
Productivity modules:
Tier 1: 10.96/0.58=18.90
Tier 2: 15.912/0.62=25.66
Tier 3: 22.68/0.7=32.4
Speed modules:
Tier 1: 5.4/2.25=2.4
Tier 2: 6.12/2.75=2.23
Tier 3: 6.84/3.75=1.824

As you can see, productivity modules are awful for pollution/unit, though those numbers can be improved quite a bit with speed beacons. Part of why they are so bad, is the speed penalty, which can be offset by speed beacons (in my kilobase, I have 4 beacons/assembler, so with tier 3 modules I could achieve: 37.8/4.2=13.2 pollution/unit), however, they also have a pollution modifier, which amplifies the energy cost to pollution conversion rate; this also means Eff modules provide greater pollution reduction for assemblers w/ prod modules in them, even if you use beacons for the eff modules, which isn't energy efficient, it IS still very helpful for pollution reduction. Speed modules also increase pollution/unit, but not nearly as badly as productivity modules, with tier 3 not being all that much worse than the base pollution/unit, and like with energy efficiency, using speed beacons w/ prod modules reduces the pollution/unit relative to just prod modules and no beacons. Efficiency modules can dramatically reduce pollution/unit values, for instance, achieving 20% energy cost (min) with no speed/productivity modifiers means that pollution/unit is now 1/5 of what it would otherwise be (1.44/5=0.288 for assem3 w/ no other modules effects). At tier 3, it's possible to do even better by running 1 speed module 3 with 3 efficiency module 3s, which gives +50% speed while still achieving exactly -80% to energy cost. Hence pollution per unit is down to an astonishing 0.36/1.875=0.192 pollution/unit, or a mere 13 and 1/3 percent of what it would normally be.

[BlakeMW]

The other thing not mentioned so far is that lower energy costs means lower pollution, which can be hugely valuable against aggressive biters, especially on death worlds (I've started one of these myself recently). So what you're really paying for, going eff module route instead of just more solar panels, is the lower pollution from your assemblers and other production buildings.

True, but the problem remains that you need to compare the eff2 module with an eff1 module. Taken alone, an eff2 module can have a pollution payback time (relative to the pollution it took to make it) in as little as about 2 hours, IIRC. The problem is, if you can use an eff2 module, you can definitely use an eff1 module, so we need to consider the payback time of upgrading the eff1 module to an eff2 module for an extra 10% pollution reduction. And that's something like 8 hours. So it takes 8 hours to save pollution by making an eff2 module.
(Obviously these calculations are very nebulous because so many variables go into pollution, from electricity source, to smelting setup, to module usage, to what the module is actually used in, I remember I calculated it to be 2 hours under fairly favorable conditions)

Maybe if Efficiency2 were -100% and if Efficiency3 were -250% (both subject to the 20% minimum), both would be usable. Any thoughts BlakeMW?

I have a personal mod for use on top of Bob's Assemblers - primarily it nerfs all module slot counts in Bob's Assemblers, but because these assemblers have much higher energy usage and crafting speed it makes modules far more potent. As such I decided to double all module costs.
That was fine for eff1, but made eff2 and eff3 really expensive - also I didn't want to make Power Armor MK2 (too) much more expensive.

After thinking about it, the economy of eff2/eff3 is so bad, I decided the best solution was to dramatically slash their cost. The recipe for each higher tier remains the same, except instead of requiring 4 or 5 modules of the lower tier it requires only a single module, you take one eff module of a lower tier, add some ingredients, and upgrade that one module to a higher tier:

Level 1: 5 Electronic Circuits, 5 Advanced Circuits: -30% Consumption
Level2: +5 Advanced Circuits, +5 Processing Units: Additional -10% Consumption (-40% total)
Level3: +5 Advanced Circuits, +5 Processing Units, +1 Alien Artifact: Additional -10% Consumption (-50% total)

When you actually look at those numbers, the first level is still by far and away the best value for money, while spending 5 advanced circuits and 5 processing units for a measly 10% reduction is a dubious investment.
You know there's something messed up with the balance when you can cut the cost by 95% and still think "eh, why bother?"

Without changing pricing, another scheme would be to make them more about pollution reduction:
lvl1: -20% Energy Consumption, -10% Pollution.
lvl2: -40% Energy Consumption, -20% Pollution.
lvl3: -80% Energy Consumption, -30% Pollution

Pollution is a multiplicative effect which is applied after speed/consumption effects, if you apply a -30% pollution effect to a speeded up assembler the pollution reduction is quite potent. With this scheme if you are using alternating rows Prod3+Speed3 beacons where each assembler benefits from 8 beacons, and outfitted 2 of those 8 beacons with eff3 modules you'd basically halve the pollution while also reducing energy consumption by 1.3MW per beacon - this would be at the expense of not having speed3 modules in those 2 beacons so it wouldn't be a clear winning move, but it'd at least be a semi-viable option.