Pushrod vs DOHC

[dwightlooi]

Essentially, what we have here is two engines which achieves exactly the same power output. One does it with 4.8 liters and 99 hp/liter, the other does so with 6.2 liters and 76 hp/liter. Both are realistic possibilities. The Pushrod engine has better torque across the rev range and lighter engine mass. The DOHC engine has better specific output and lower displacement.

Edited May 7, 2011 by dwightlooi

[loki]

is the 2 valves on the dohc a typo?

[dwightlooi]

is the 2 valves on the dohc a typo?

Yes, I realized that and went on to change the JPEG, but the edit option timed out...

Anyway... one thing of note here is that by adopting the 60 deg bank angle, the engine will not only benefit from the convenience of sharing the production line and components with the high volume V6, it also makes the engine considerably narrower such that there is no longer a width disadvantage compared to a pushrod 90 deg V8

Edited May 7, 2011 by dwightlooi

[balthazar]

InBlockCam sounds nicer, same technical level as 'OHC'. Needs to morph into the popular lexicon.

[Intrepidation]

The 63 ft-lb difficulty of the DOHC V8 is something worth noting.

[loki]

InBlockCam sounds nicer, same technical level as 'OHC'. Needs to morph into the popular lexicon.

mmmmmmmmmmm rootbeer and cream soda haha

I think the car's specs would determine which one. if it's 3200+, lets say, pushrod, under that, DOHC. but since under 3200lbs is rare, generally, pushrod.

Edited May 8, 2011 by loki

[riviera74]

The IBC (i.e. pushrod) V8 seems better, but is it noisier than the DOHC V8? It seems that almost everyone else with a V8 went DOHC years ago since that is all they build.

[dwightlooi]

The IBC (i.e. pushrod) V8 seems better, but is it noisier than the DOHC V8? It seems that almost everyone else with a V8 went DOHC years ago since that is all they build.

Well, yes and no. A properly engineered and put together Pushrod Engine is not noiser or less refined than a DOHC engine because of the valve train per say. A 5 liter DOHC and a 5 liter pushrod engine with the same bank angle and block stiffness has similar vibration characteristics. The problem is that a pushrod engine has lower specific output because of more restrictive airflow in its 2-valve heads, hence a pushrod engine tends to have higher displacement than a DOHC engine with the same horsepower. Bigger displacement means that the pistons are larger, or the stroke is longer, or both. It is the higher reciprocating mass and piston speeds that result in higher vibration levels. This is especially apparent at higher rpms because vibration is a function of kinetic energy and therefore a function of the square of piston speed. Apart from that, poor workmanship or engineering can also led to valve float and rocker clatter. These tend to be worse in sloppy push rod engines than in sloppy DOHC engines because the valve train mass is higher. But, these shouldn't exist in either if the engine is well designed and in good functioning order.

Put simply, a 6.2 liter pushrod is no less refined than a 6.2 liter DOHC, all else being equal. However, it is probably less refined than a 4.8 liter DOHC mill. The same argument can be used against a 6.2 liter design (M-B AMG M156 engine) when comparing it to a 4.4 liter BMW N63 design for instance -- both of which are DOHC.

The migration from pushrods to DOHC powerplants are mainly driven by inline-4s and V6es going DOHC and companies garnering expertise in their design and manufacture.

Edited May 8, 2011 by dwightlooi

[riviera74]

Dwightlooi, has GM ever made a GOOD pushrod 4cyl? I know that the Buick 231 (nowadays the 3800 Series I,II,III) was a fabulous v6. I should know: both my former Intrigue, 1st gen LaCrosse and my Park Avenue Ultra have that engine.

[Robert Hall]

Hasn't this already been debated on here dozens of times? It's a moot point--the world's automakers have long since moved on from pushrod to DOHC..there are only a couple holdouts..

[loki]

Hasn't this already been debated on here dozens of times? It's a moot point--the world's automakers have long since moved on from pushrod to DOHC..there are only a couple holdouts..

and look who's responding to dwight's threads. in this case it's a new member...

[balthazar]

This is less about 'what everyone else is doing' and more about 'what's the best thing to do spec-wise'.

[Drew Dowdell]

I could care less the valve train configuration as long as it gives me the performance I expect or better. I am not convinced that the common mindset of "DOHC is always smoother" is always true. I've driven DOHC V8s that the LT-1 makes sound like a Quad-4 without oil.

I've driven the 3900 for 25 miles in the wrong gear because the engine was so smooth, the only reason I noticed something was up was when the fuel economy average started dropping like a rock.

Conversely, even the same DOHC V6 can have very different personalities depending on the application. The 3.6 VVT in the Camaro Convertible was annoying and buzzy, yet simultaneously lacked feedback.... The 3.6 VVT in the Cadillac CTS was fine.

[siegen]

Are there any good SOHC V8's to compare? SOHC will weigh less and be physically smaller than DOHC. It would lose the intake and exhaust tuning flexibility, which would make it more directly comparable to a pushrod.

[Carguy]

As good as the OHV SB V8 is and could be with additional refinement's I wish GM would do DOHC 4VPC heads with Direct Injection for a premium SB V8. Keep the OHV versions for Chevrolet performance Cars and Truck models. I could see a nice short stroke,low deck height DOHC 5.5L version with 450-470HP as an opt on the next CTS and upcoming large RWD Cadillac model. Then do 5.3L-6.2L versions tuned for torque as GMC truck engines while Chevrolet could use the OHV versions. With the talk of GMC going up market this along with styling and features could help to define GMC more from Chevrolet! This SB V8 based 5.5L could being that its larger be more efficient then the smaller 4.8L HFV6 based V8 above being that it would turn slower to make the same power!

[dwightlooi]

Are there any good SOHC V8's to compare? SOHC will weigh less and be physically smaller than DOHC. It would lose the intake and exhaust tuning flexibility, which would make it more directly comparable to a pushrod.

Mercedes M113 4.3, 5.0, 5.5 liter V8s. SOHC, 3-valve per cylinder, twin spark. Smooth but not particularly high specific output. The advantage of SOHC, includes narrower and lighter heads (not as small and as light as pushrod motors' but smaller and lighter than DOHC designs), plus lower frictional losses by having half as many camshafts and their bearings (again, not as low as pushrod designs but better than DOHC designs). It is for these reasons that Honda 3.0, 3.2, 3.5 and 3.7 V6es in the Accord, TL and RL are SOHC designs.

SOHC designs do not necessarily have to forgo independent VVT. In fact, so can a Pushrod V8. They can use a Concentric Cam setup (ala Dodge Viper's pushrod 8.4 V10) to achieve that.

[siegen]

Mercedes M113 4.3, 5.0, 5.5 liter V8s. SOHC, 3-valve per cylinder, twin spark. Smooth but not particularly high specific output. The advantage of SOHC, includes narrower and lighter heads (not as small and as light as pushrod motors' but smaller and lighter than DOHC designs), plus lower frictional losses by having half as many camshafts and their bearings (again, not as low as pushrod designs but better than DOHC designs). It is for these reasons that Honda 3.0, 3.2, 3.5 and 3.7 V6es in the Accord, TL and RL are SOHC designs.

SOHC designs do not necessarily have to forgo independent VVT. In fact, so can a Pushrod V8. They can use a Concentric Cam setup (ala Dodge Viper's pushrod 8.4 V10) to achieve that.

Indeed. The J37 in the new TL has independent intake and exhaust vtec, using a Y shaped rocker arm for intake valves. I would love to open one of those up and have a look! Prior to 2009, all Honda J-series were intake-vtec only.

I wonder how a 6.2L SOHC V8 would stack up in this comparison. A high-output one should be good for around 80hp/liter, or 500 hp. The torque curve drop off would probably be in between the OHV and DOHC. I think it would strike a good balance between engine complexity, cost, and efficiency.

[dwightlooi]

Indeed. The J37 in the new TL has independent intake and exhaust vtec, using a Y shaped rocker arm for intake valves. I would love to open one of those up and have a look! Prior to 2009, all Honda J-series were intake-vtec only.

I wonder how a 6.2L SOHC V8 would stack up in this comparison. A high-output one should be good for around 80hp/liter, or 500 hp. The torque curve drop off would probably be in between the OHV and DOHC. I think it would strike a good balance between engine complexity, cost, and efficiency.

The Honda SOHC engines do not have continuously variable VVT. They have fixed phasing, but the ability to switch between two sets of cam lobes (some have the ability to switch between three sets). Since it is a completely different grind, this lobe can be of a completely different timing, lift and duration.

In the latest J-series (V6) and the R-series (I4) There is a set of cams optimized for modestly high output -- which at 77~81hp/liter really is NOT that aggressive by Honda's or most standards. The other is essentially an Atkinson Cycle cam used for cruising and low load conditions. The Atkinson cam closes the intake valves VERY late -- well into the compression stroke -- kicking part of the ingested air back out the intake ports. This reduces effective displacement and creates an effectively asymmetrical compression vs power stroke ratio. Both contributes significantly towards fuel economy.

An SOHC design will indeed have a degree of torque linearity and high end torque fall off between DOHC designs and Pushrod designs, but closer to DOHC than Pushrod. However, it also has higher internal friction at a level closer to DOHC than Pushrod. The beauty of the Pushrod engine is that you only have one cam instead of four (or two in case of SOHC). The less cams, the less bearings and the less valves, equals lower parasitic friction from the valvetrain. While the engine does not breathe as well and this hurts output at higher RPMs, this is essentially irrelevant for fuel economy. At cruising speeds and at light throttle, no matter how well the engine can technically breathe, it will be choked by the throttle butterfly to flow only as much air as is needed to produce the amount of power required to sustain cruising speeds. Actually, if this does not happen what you have on your hands will be a case of "unintended acceleration"!

Edited May 10, 2011 by dwightlooi

[regfootball]

checking in to see if this question has indeed been solved. by now.......

[siegen]

The Honda SOHC engines do not have continuously variable VVT.

Apologies if that was how it sounded. I didn't intend to indicate that the J-series had continuously variable phasing, just traditional vtec.

[Robert Hall]

checking in to see if this question has indeed been solved. by now.......

Nope..never..this is one of those things that's going to be argued until the end of time, like FWD vs RWD.

[G. David Felt]

Over all, I have yet to find a DOHC V8 that can beat my Pushrod V8 bored from a 350 to a 402. I will stay with the tried and true means of Torque and power to pull my trailers.

[loki]

checking in to see if this question has indeed been solved. by now.......

Nope..never..this is one of those things that's going to be argued until the end of time, like FWD vs RWD.

least till cams are no more.

[dado]

This again..heh.

Pushrod V8 for trucks,big cheap cars (future big chevy rwd sedan etc) and DOHC V8 (but bigger displacement than this one) for Cadillac models.

Also weight of 220 kg would be too much. Coyote v8 has less than 200 kg. I think GM should be able to come lower..

[siegen]

It would be nice to have individual electronic valve actuators. Think of the weight savings and reduced losses!

[Nick]

It would be nice to have individual electronic valve actuators. Think of the weight savings and reduced losses!

I've been really wondering why we haven't seen something like this come along yet. With all the continuously variable valve timing technology that exists you'd think that they'd finally do away with the cams all together.

[67impss]

Yeah those rotary valves looked promising if they can get the seals to hold up

[siegen]

I've been really wondering why we haven't seen something like this come along yet. With all the continuously variable valve timing technology that exists you'd think that they'd finally do away with the cams all together.

The electricity requirements might be an issue. Not sure how much it would take to operate, with the degree of precision required, the valve actuators. The actuators themselves may be rather spendy. It would also require more initial electricity to start the car, though the reduced friction from no camshafts or corresponding belts/chains may make up for it. It may even result in less electricity being required to start the car. Obviously I have no idea what the trade-offs are. Maybe a higher-output alternator will be required.

[dwightlooi]

It would be nice to have individual electronic valve actuators. Think of the weight savings and reduced losses!

No way! You'll need 16 solenoids, each being the size of a can of soda. That's assuming that you are using one solenoid to operate 2-valves and not counting the electric generation and storage system needed to power them! It'll be bigger and heavier.

This is why even Formula One uses mechanically opened valves and electro-pneumatically assisted closure. The valves are opened with a regular cam lobe, but closed with a combination of a valve spring and pneumatics; the latter being electronically controlled so closing force can be varied according to rpm. The solenoids do not actually open or close valves directly, all they do is meter out a "booster" force in form of air pressure to help close the valves at high RPMs.

[siegen]

Or it could be the software needed to control it.

http://www.launchpnt.com/portfolio/transportation/magnetically-actuated-engine-valve/Here is a company working on magnetic valve actuators.

[dwightlooi]

Or it could be the software needed to control it.

http://www.launchpnt.com/portfolio/transportation/magnetically-actuated-engine-valve/Here is a company working on magnetic valve actuators.

The problem isn't control. The problem is force. A typical valve spring puts out between 120~300 lbs of force. A solenoid with that kind of power is big and power hungry. This causes three problems... (1) It takes up more room that camlobes and valvesprings (2) It needs a lot more electrical power to operate and is available from a typical alternator or 12V battery. (3) It gets pretty hot and needs to be cooled.

And, in the end what do you get? Infinitely variable valve timing and lift? We can already do that via mechanical means. Better efficiency? Perhaps, but it is really a race between reduced frictional losses and increased electric generation losses.

[siegen]

The problem isn't control. The problem is force. A typical valve spring puts out between 120~300 lbs of force. A solenoid with that kind of power is big and power hungry. This causes three problems... (1) It takes up more room that camlobes and valvesprings (2) It needs a lot more electrical power to operate and is available from a typical alternator or 12V battery. (3) It gets pretty hot and needs to be cooled.

And, in the end what do you get? Infinitely variable valve timing and lift? We can already do that via mechanical means. Better efficiency? Perhaps, but it is really a race between reduced frictional losses and increased electric generation losses.

Would you need valve springs? Why not a pair of electromagnets to control movement each way? I understand that size may be large, but that's how all new technologies start.

[dwightlooi]

The problem isn't control. The problem is force. A typical valve spring puts out between 120~300 lbs of force. A solenoid with that kind of power is big and power hungry. This causes three problems... (1) It takes up more room that camlobes and valvesprings (2) It needs a lot more electrical power to operate and is available from a typical alternator or 12V battery. (3) It gets pretty hot and needs to be cooled.

And, in the end what do you get? Infinitely variable valve timing and lift? We can already do that via mechanical means. Better efficiency? Perhaps, but it is really a race between reduced frictional losses and increased electric generation losses.

Would you need valve springs? Why not a pair of electromagnets to control movement each way? I understand that size may be large, but that's how all new technologies start.

No you wouldn't need springs at all. But, electromagnets that can move the valves fast enough dictate coke can sized solenoid actuators fed with very high voltages and current. It takes something about ~20 times the mass and much larger than springs to provide the same motivating force as the springs. To work these you also need to supply lots of volts and lots of amps, which meas a larger generator, bulky power control circuitry and big fat wires.

Edited May 13, 2011 by dwightlooi

[Drew Dowdell]

Dwight, can the range of cam lob timing in VVT ever be made so broad that it goes passed the line of improving engine performance and into causing misfires, exhaust lock, etc?

Basically, would adding solenoids fix anything that VVT can't already do?

Sure with a solenoid valve, you could open the exhaust valve at the beginning of the power stroke.... but why on earth would you want to?

[Nick]

I don't think it is about doing something that VVT can't already do, but rather, doing it differently. It would also bring a pretty drastic redesign to the ICE in general...No more timing chains. No more cams. Infinitely variable valve timing. I'm not sure how loud those solenoids would be, or how heavy they are, but it seems reasonable to suspect that you would at least need a bit more electric power to run the engine, and the engine may be a bit taller than a typical DOHC engine. You'd have far less moving parts, far less drivetrain loss, the possibility of a quieter, smoother (and perhaps lighter?) engine. It just seems like it would be a natural progression.

[dwightlooi]

Dwight, can the range of cam lob timing in VVT ever be made so broad that it goes passed the line of improving engine performance and into causing misfires, exhaust lock, etc?

Basically, would adding solenoids fix anything that VVT can't already do?

Sure with a solenoid valve, you could open the exhaust valve at the beginning of the power stroke.... but why on earth would you want to?

Well, most VVT gears of today have variance of around 60 deg. Going beyond that is not difficult, it is just pointless. VVT is used to achieve the following under the appropriate conditions:-

• Shifting the torque peak to a higher or lower rpm (flattening the curve) -- accomplished by advancing and retarding
• both cams in unison
  
• Creating an Exhaust Gas Recirculation (EGR) effect -- accomplished by retarding exhaust cam while advancing intake cam
  
• Reducing effective displacement and compression -- accomplished by retarding the intake cam into the compression stroke
  

All of the above are limited by the lift and duration of the actual cam lobe grinds. For instance, to get enough intake duration to effect an Atkinson Cycle operation (keeping the intake valves open for first 20~40% of the compression stroke) the required duration is too long for maximizing power output. In another scenario, where you want very light valve lift and overlap for high specific output, the low rpm drivability may be beyond the ability of VVT to tame regardless of how you change the overlap or timing -- there is just too much duration and too much lift.

It is for these reasons that Honda prefers switching between two sets of cam lobes with different lift and duration ground into them. It is also for this reason that BMW and Nissan opted to have an intermediate rocker system to allow the amount of effective valve lift to be adjusted. The Honda system cannot provide lift levels in between the two steps, hence it cannot avoid the "sudden" getting on cam effect VTEC engines exhibit when the system switches cams. The BMW/Nissan systems cannot change duration, hence they are not as effective at turning in 120~125bhp/liter solutions that will still drive fine from idle on up. There are system concepts which allow lift and duration to be varied, but these generally require that lift and duration be increased or decreased in unison, and the valvetrain mass of these designs are quite high limiting maximum rpm.

Electromagnetically actuated valves will provide complete freedom to have any lift, duration, timing, overlap and ramp profile at any time or even different ones for different cylinders (shutting down some cylinders completely for instance). Engines have infinitely variable effective compression and displacement. They will also be able to have meet emission requirements or fuel requirements for different markets simply with a software flash.

The problem is that Electromagnetic valve actuation is not proven to be reliable and requires heavy and bulky hardware that sucks down electric current like its free. This may mean that the energy costs for operating it may be higher than the energy savings garnered from having the best possible cam grind and timing in ALL situations, ALL the time.