Nerd’s Eye View: Mercedes AMG M139 Engine with E-turbo and Liquid Immersion Battery Cooling

The Garrett e-turbo has some more special sauce going on at the compressor side that is not mentioned anywhere that I can find easily. You can see this extra actuator on the compressor housing under the compressor discharge. I lightened up the picture a bit, and you can see some moveable feature in the compressor inlet by the inducer of the compressor wheel. Pure speculation on my part, but I am guessing it basically allows air to bypass the compressor wheel. So, my guess is that when the car is coasting down in speed, some moveable part in the compressor stage allows air to essentially bypass the compressor wheel so that no shaft power of the turbo is used to compressor the air. Therefore, the engine is pumping air through the turbine wheel which then extracts energy out of the flow to spin the motor of the e-turbo and charge the battery. Another use case could be when the battery is low and the car is just driving and there is low engine load, this hypothetical bypass feature is used to not waste turbo shaft power on compressing air and maximize the power going to the battery. Again, just a guess on my part.

You will probably hear people say, “442hp out of a 2.0L engine is no big deal. People have been doing it for two decades out of imports!” Yeah, well, try doing that while meeting modern emissions standards and having a warranty. To meet modern emissions standards, there is a massive catalytic converter right after the turbo and I am guessing a gas particulate filter after that. There are quite a few dimples on the exhaust pieces to maintain minimum OEM clearances between parts.

Part of the puzzle for powering the motor of the Garrett e-turbo is this battery pack. Based on the specifications given by AMG of 70 kW continuous power and 150 kW peak power from a pack with only 6.1 kWh of capacity, they are pushing the battery cells really hard. 70 kW corresponds to a C-rate of about 11, give or take assuming I did my math right and using 400V as the pack voltage. For reference, a Tesla supercharging is roughly at a C-rate of 5-6. If you want your cells to last as long as possible, you don’t go over a C-rate of 1. C-rate for batteries is like the equivalent of hp/L for an engine. The higher it is, the more stress on everything and typically shorter life. Running high C-rate means a lot of heat, just like pushing a lot of hp/L. Therefore, serious cooling is required and Mercedes AMG is using direct liquid immersion cooling; so, the battery cells are submerged in a dielectric fluid (non-electrically conductive). This is first-to-market with the technology that I am aware of in a relatively mass-produced vehicle. The McLaren Speedtail also uses direct liquid cooling and Faraday Future developed the technology years ago, but they have not launched their vehicle. We looked at some other battery cooling systems for reference when we dove into the Porsche Taycan a bit.

10 kW of cooling capacity is the stated number from Mercedes AMG, but you now how marketing always likes to inflate numbers and that is likely a maximum number. The question is under what conditions the 10 kW of heat removal occurs. If we assume that 10 kW of heat rejection occurs during 70 kW of continuous power, that’s ~12-15% of energy loss in heat. That efficiency for a battery cell would be relatively poor. I’m going to go with the 10 kW number coming after a burst of the max 150 kW power output.

This diagram shows the basic layout of the guts inside the battery pack. The coolant pump circulates the dielectric fluid through the battery pack. The fluid is likely a type of oil, similar to dielectric oil fluids used for cooling electrical transformers you see on power lines.


  1. Great article!! I take it this battery pack is solely for the turbo? why not tap into an existing hybrid drivetrain to power an “e-turbo” or is the goal specifically about power/capacity/emissions. also, seeing the batteries like that seems obvious now, lol.

    1. The E-turbo is part of an already hybrid drivetrain. The E-turbo needs very little energy, only a couple seconds to spin up the turbo and get the engine into boost and making enough exhaust flow to spin the turbo without electric assist. It would cost too much to only have high voltage electronics and battery for just an E-turbo. But an E-turbo is a small cost add-on to an already electrified hybrid drivetrain that already has a turbo engine. As the European makes have already made a bunch of hybrid cars with turbo gas engines, it was only a matter of time for e-turbo to get implemented. You’re basically paying for an extra inverter and the small motor in the e-turbo, and a few more HV cables and signal wires.

      As for the size of the electric motor powering the wheels, I haven’t seen AMG say power it is. But considering they say the battery can put out 150kW of power, we can make a good guess. Conservatively, 90% of that 150kW turns into electric motor power which equals 135kW. And that equals 181hp. We add 181hp to the 442hp of the gas engine and we have an estimated total output max of 623hp. But typically, the combined output is lower than the individual max outputs combined; the electric motor typically isn’t asked for max power at the higher engine speeds. But… that’s in your everyday boring Prius type hybrid. So… hard to saw what AMG has done.

    1. Yup! Even with the transition to electrification, everything still needs cooling. There is a lot of room for optimization and creativity in the thermal fluid systems.

  2. This makes me wonder where they place this battery in the given vehicles: as 21700 cells (7 rows) alone stack up to 490mm long, and then roughly 252 wide and 168 high as arranged (19.3″x9.9″x6.6″). Do they stuff them where a spare would go, or cut into fuel capacity?
    My dims don’t account for the casing itself nor the pump and control board side of things.

    Definitely a cool breakdown on the system however. Those 6ah batteries, I want em!!

    1. Front what I’ve seen, the plugin hybrids typically put the battery pack in the floor of the rear cargo area. Old school spare tire location. I just looked up a picture and they have it sitting above the rear electric motor. So over the rear axle centerline. Basically behind the rear seat then? On the 6Ah cell… Yeah, I had to double check my math. Caught me by surprise on the capacity too. Granted, I’m not up to speed on the latest battery cell specs, but it’s more than anything I was aware off. Still the possibility my math is wrong! It does depend on assumptions used for cell voltage and I assumed nominal 3.65V

  3. Nice write up Khiem. Regarding the mechanism on the compressor side, Garrett (under Honeywell at the time) published a number of patents on upstream mechanisms aimed at extending the usable compressor map. The common theme one will find if they read these public documents is that something is introduced into the flow stream upstream of the compressor effectively reducing the inlet flow area, and removed from disrupting flow once the operating condition seems it beneficial. Such a system would benefit any turbocharger where both low end torque and high end power are constrained by the compressor map width. My suspicion is that this is what is in this e-turbo and not a compressor bypass. Yours is the first time I’ve seen it mentioned in an article and Garrett has been quiet about the technology which leads me to believe they are waiting for a suitable industry conference to unveil it.

    1. Thanks for the insight! Well, it can’t be too much of a secret or you’d think they would have hidden it better. One of these days, maybe I’ll do a patent search to try to find it.

    2. You know…. way back in the day, I had the idea of a recirculation path to effectively increase map width. Say the compressor wheel would surge at less than 15 lbs/min but the engine was only asking for 12 lbs/min, So the wheel could flow 15 lbs/min with 3 lbs/min getting recirculated so that only engine only see 12 lbs/min. Looking at this a bit more, it looks kinda like it just moves a part of the shroud contour away from the wheel which I guess would effectively reduce the mass flow of the wheel. I guess the information will come out soon enough.

  4. “ There’s also a new and larger turbocharger residing between the engine and the front bulkhead. This unusual location necessitates better cooling techniques to prevent the blower from overheating under sustained high-rpm operation. This is achievable thanks to the new coolant plumbing, but most of it has something to do with the new engine cover that directs cooling air to the turbocharger. Mercedes says the 415-horsepower ‘S’ version of the M139 also utilizes cooling air from the HVAC system when needed.”

    That’s pretty interesting. I think that the HVAC cooling is what you saw on the compressor side of the housing.

    The press release also quotes roller bearings.

    “ Equally, the increased sophistication of the post 2014 power units brought complication to the cooling system with thermal management of the Hydrid system’s; Battery, MGU’s and Control Electronics. These run at much lower temperatures than the ICE, their cooling is achieved by either water-based coolant or dielectric fluid (oil). Running at around +50c they greatly affected by the ambient temperature going above 30c.”

    That was my first thought reading this article. We know that oil has something like half the specific heat capacity of water, so it seems like there would be ambient temperature situations where you could overload the battery cooling system on the Merc.

    The battery is also located above the rear axle, so I don’t know how it gets any airflow over the case.

    1. Got a link to the press release? The language is weird… nothing unusual about the location of the turbo at all. The engine is longitudinal layout and the turbo is hanging off the side, like a bazillion other engines. Marketing though… I get it, they have to make it sound special. In general, turbos don’t need external airflow for cooling them, but you want airflow around the turbo to keep everything around it from baking. Evo X and Nissan GTR have NACA ducts in their hoods for this reason. The WRX diverts some of the air going into the hood scoop over to the turbo area. With this e-turbo, keeping the power electronics cooler is a good thing.

      In the picture from Mercedes, the bearing is most certainly not a roller/ball bearing, just a good ole journal bearing. So either the cutaway turbo is not representative of the production unit, or the press release is wrong. I’d go with the press release being wrong as the media teams often do not speak to the engineering teams. Also, doing a bearing change from journal to ball bearing would mean redoing all the validation testing, so months and months of work.

      On the battery cooling, battery cells typically like to be around 35C-40C. Give or take. It seems Mercedes likes to keep this pack around 45C. The little bit warmer I think reduces internal cell resistance at potentially a trade-off of life, but 45C is not by any means outrageous. It could be that newer cells are more tolerant of heat too. While the specific heat capacity of the dielectric oils are lower than water, you can just have a higher flow rate to compensate. All high performance EVs have a chiller to sub-cool the coolant below ambient. So the cars run the A/C system and some/all of that cooling capacity goes to cooling the battery cooling fluid below ambient.

      1. I have personally never heard of ‘HVAC cooling’ of a turbo, either. But, it has been quoted in a couple articles, so I am sure they are doing something unique. It’s really hard to figure out the plumbing with the pics available.

        I totally agree with your analysis given the pictures presented, but now, I honestly think that there might be two different turbos for the M139. One for the S model, another for the lower power model:

        From Motortrend:

        “The twin-scroll turbocharger now employs roller bearings to reduce friction and an electronic wastegate actuator for more precise control. “

        “On S engines, the air-conditioning system helps to reduce the intercooler’s temperature under maximum-power conditions.”

        This article shows what looks like a different turbo than the S model, but it has the same M139 model number:

        1. The M139 with 415 hp is from the A45 AMG and does not have an e-turbo. I am 100% sure the e-turbo is journal bearing, and does not have HVAC cooling.

  5. It doesn’t look like my last reply made it up (maybe the links got blocked automatically.)

    But, there must be two different turbos. One must be exclusive to the S model. You can see it in the pics on Motortrend. One has the ‘mystery actuator’ and the other doesn’t. I would still like to know what exactly it does because I have no idea how ‘HVAC cooling’ of a turbo would work.

    “Furthermore, the shafts of the compressor and turbine have roller bearings for the first time ‑ as in the top output variant of the AMG 4.0-litre V8 engine in the AMG GT 4-door Coupé. The roller bearings reduce mechanical friction within the turbocharger to a minimum. The charger therefore responds more readily and reaches its maximum speed of up to 169,000 rpm more rapidly.”

    “ Fresh air is used in addition to oil and water to cool the turbocharger. This is directed specifically to the charger from the radiator grille, via the engine cover designed as an air deflector and ducts beneath the bonnet.”

    “ The turbocharger resides between the engine and the front bulkhead. This unusual location necessitates better cooling techniques to prevent the blower from overheating under sustained high-rpm operation. Mercedes says the 415-horsepower ‘S’ version of the M139 also utilizes cooling air from the HVAC system when needed.”

    Sorry, that it’s not letting me post links, but those are direct quotes from the press releases from AMG. I didn’t write them.

    1. Ah… the Motortrend article is for the previous generation I wrote about too.

      As for the HVAC cooling of the intercooler, that’s neat and I didn’t know they were doing it. That M139 uses an air to water IC. If I were doing it, I’d have a chiller that uses the refrigerant from the HVAC system to sub-cool the water going the the IC. Just like how the high power EV battery packs do.

      The photo in the carmag link without the mystery actuator, I saw that picture a year ago too. So I’m assuming an earlier development variant.

      1. I wonder if the HVAC cooling is solely responsible for the 34hp increase on the S-model.

        Your idea for the HVAC cooling sounds a lot better. I was thinking that they were just blowing de-humidified cool air over the IC, which would seem totally pointless in a hot, cramped engine bay.

        I missed that first article, thanks for the link. Love these articles, BTW.

  6. So how does it compare with hybrid turbos from TTE or MACHGRADE? The hybrids deliver more power than this electric unit, is there any more downside than increased turbo lag? I mean, the lag has some old-school vibe and if the car is faster, then … it’s faster : )

    1. There’s nothing special about hybrid turbos; they just have bigger wheels stuffed in the original castings. E-turbos took quite a while to develop and bring to market due to advances required in high speed motors, the high voltage electronics surviving the heat and vibrations of the turbo, and the necessary controls. Of course, they’ve been in Formula 1 since the introduction of the current turbo V6 hybrid era. E-turbos are being developed in many sizes.

Leave a Reply

Your email address will not be published. Required fields are marked *