Cosworth DFV Cylinder Head - A Closer Look - Ep 21

[@Brit_Toolmaker]

Excellent and Interesting video Brian! I'm a metal pattermaker by trade and was involved in castings during the 1980's and moved across to injection mould toolmaking for the challenges that brings. The technology used back then to make tooling, compared to today, it's like another world. Back then I was using pantographs and hydraulic copy milling machines to make tooling, all from hand drawn dyeline copied blueprints, most of the work I was involved in was formed three dimensional work, that would sometimes require the services of highly skilled model makers to hand create the pattern for copying. Casting wise the work varied from green sand, shell and gravity die castings, because I was at the time in the Thames valley area. it did include some motorsport work. Sand core boxes would vary between CO2 cured, oil baked in ovens with wire supports and also metal heated blown core boxes using very fine shell moulding sand that was coated with a heat sensitive binder. A wide and varied, but always interesting career! If you've not seen it, there's an interesting book called "The story of the Ford Grand Prix Engine" published in 1971 that covers the DFV story in great detail. I recall it mentions that the first racing cylinder heads ran hotter than expected, and upon cutting the heads into sections (exactly like yours shown here!) revealed an error in the coring out by the patternmakers, they'd missed out some of the cooling channels around the exhaust ports. At the back of the book there is an appendix listing every engine made to date listing the engine number, release date, customer and and subsequent changes to the engines specs. Interestingly there are two successive engines built in late 69 with customer listed as " Mitsui Trading Co" who happen to be the European trading arm for the "Yamaha Motor Co Japan" who also happen to be the main engine makers for the "Toyota Motor Co", you will recall that Yamaha entered F1 themselves at some point! The two engines supplied to Mitsui are not surprisingly listed as "Not Seen Since", in the appendix. With regards, John

[@PistonAvatarGuy]

Odd little tangent, there, on the Yamaha engines. Was that meant to be a dig, aimed at Japanese engine manufacturers?

[@mekilljoydammit]

@@PistonAvatarGuy Honestly it's far from the first I've heard of a larger company buying one of Cosworth's customer engines to reverse engineer - not necessarily directly copy but figure out details.

[@PistonAvatarGuy]

@@Brit_Toolmaker In that case, I suppose that it's unfortunate that the Brits didn't think to imitate Honda when they decided to dominate F1 during the eras when power output was at its highest.

[@ralphmillais5237]

@@PistonAvatarGuy Keith Duckworth openly admitted he paid very close attention the the Honda Formula 2 engine that was dominant at that time when he designed the combustion chambers of the FVA and subsequently the DFV.

[@Brit_Toolmaker]

@@PistonAvatarGuy you are picking on the wrong person pal. I happen to own a 1961 Honda 125 Benly, an incredibly advanced design for its tine, mine was one of the very first sold here in the UK. Think you are just trolling me now!!!

[@rafaellastracom6411]

The dimensional accuracy of the mold is very impressive even for modern motorcycle performance heads. Tough to find even the most minute fault with those castings. Great stuff, thanks for sharing. Oh yeah, don´t be such a stranger.

[@camerone397]

lovely stuff. I mean this as a compliment- this is the best kind of content to have on in the background when doing my engineering homework

[@31415926535ism]

The quill assembly was to cure a transient torque spike in the valve train gear assembly that they could never calculate or predict. Amazing brilliant solution by Duckworth, still in use on racing engines. Really enjoyed the video thanks.

[@EngineeredtoWin]

See Ep 2 for torsion damper teardown!

[@31415926535ism]

@@EngineeredtoWin I have watched all your videos and enjoyed them very much thanks. Your channel is sort of unique. Keep it up please.

[@EngineeredtoWin]

@31415926535ism Thanks and I sure will!

[@sparky60ful]

Great video again. Thank you so much for sharing it with us!

[@Fintoman]

This is fascinating series of videos you have put together here Brian. No small feat. As a lover of Formula 1 its engineering and technology, I'm slowly working my way through all of them. A fantastic opportunity to see inside these masterpieces not normally available to us mortals. Many thanks for your considerable efforts and knowledgeable commentary.

[@EngineeredtoWin]

Thanks Russ...if I was starting all over again a script and proper preparation might have been a better idea. However, they are what they are now. Much rambling... Brian,

[@Fintoman]

@@EngineeredtoWin Your "ramblings" are incredibly informative to me. 😄

[@Lou-f]

40:23 reminds me of wreck diving see the valve in there. Framed piece with the drawing sounds fab👍

[@fragdude]

Love the video, very educational. These are great little engines. “Let go in style” it totally did I wonder if they didn’t choose to add oil ports/journals from the center to the exhaust end so that there might be more surface area oil coverage? As you said the center of the head has a fair amount of metal which would all be a heat store.

[@pjay3028]

Truly sensational video. Absolutely absorbing, thank you so much.

[@EngineeredtoWin]

Thanks P!

[@NorthWesternSouthEast]

I appreciate the naming of the early engines. FVA - four valve type A. DFV - double four valve. As you said, the DFV 8 cylinder is very much a double version of the FVA 4 cylinder.

[@Preen59]

As far as I'm aware the acronyms are FVA = Four Valve Assembly. BDA = Belt Drive Assembly. DFV = Dual Four Valve.

[@RichardSFord]

@28:00 Were the patterns made by The Zeus Pattern and Tool Company, Birmingham? They were mentioned in part one of the 1986 Equinox documentary "Turbo" about the development of the Cosworth V6 turbo.

[@camshaftcasting1451]

I stand to be corrected but I believe these pre-COSCAST castings came from Aeroplane and Motor Aluminium Castings, in the Black Country. They had their own pattern shop. I'm not certain A&M made the patterns for these heads. Looking at them, it's clearer see why they decided to start their own foundry using John Campbell's new ideas.

[@st939]

Epic level engineering. I always admired the simplicity and efficiency of the DFV. It didn't dominate for decades by chance.

[@jamesgeorge4874]

Lobbying for restrictions against Offenhauser is what kept them "dominant" at Indy.

[@st939]

@@jamesgeorge4874 Which engine did the lobbying keep out of the competition ?

[@jamesgeorge4874]

@@st939 seriously ? No engine lobbies. _racers who couldn't beat the Offy_ lobbied. For size restrictions. Then boost restrictions, then an outright ban, which they got.

[@st939]

@@jamesgeorge4874 So which engine was left out? Some massive US V8 with a flat plane crank and a central camshaft with pushrods? Which one?

[@jamesgeorge4874]

@@st939 I named them in my I initial reply, and by an abbreviated name in the second. I cannot take this seriously, as you missed the name, context, everything. _Fifth word, first reply_ !!

[@stromlo]

35:14 If you look to the right of the spark-plug area, would that be considered a defect, or is it typical for castings? I realize this would never be detected normally, but the metallurgy is fascinating.

[@stromlo]

And now I just realized that it is probably just the start of a passage that was cut through (i.e. the rest of the passage would be on the now-removed material). Never mind!

[@1PADRAIG]

13:53 These holes in the Head could be holdover from the earliest DFV. They had oiling problems with the heads because the scavenge pumps weren't big enough to handle blow-by. If you look at photos of early DFV, Graham Hill's test at Snetterton and Jim Clark's Dutch GP car, the right valve covers have 2 boxes with hose back to these holes the head. This could been a early fix but didn't need them and easier not to fix molds.

[@newagetemplar6100]

Lovely and much appreciated content. I too can appreciate the art of pattern making and used to spend my evenings when younger visiting my friend who was a pattern maker , he was old school and made some amazing things . I totally get the accuracy of the castings especially when you have to consider 4% shrinkage in 3 Dimensions. There are no voids or porosity in the head either . Off memory lol , going off my block and sump the oil pump and scavenge pump is on one side and the fuel and water pump driven off the opposite side id have to check ….. that’s engine number 35& and dates to the 80’s . Hats off to Brian Hart and especially Roy Millington too . Been wondering where you’d got to , best go and wash the old hands now they looked immaculate at the start of this video. 😉

[@corey_nz]

I've only skimmed some parts of this video so far, but the cross sections of how the coolant passages were routed around e.g. at 19:00 on are fantastic. A long term project of mine is building a 400cc Chevrolet style pushrod V8 - eventually. Just gathering knowledge for now. One of the things I haven't settled on a general design for is a cooling system and how to form that with home casting, lost foam or sand based. This engine will used on some sort of mini hot rod to drive around car shows, so it needs one. There's a lot going on in a head, so that's quite a difficult thing to incorporate, but seeing how the parting lines (?) for the cores are and pressed in spark plug housing/liner have been used have given me a load of ideas already. Thanks for explaining how everything might have been done, it's a wealth of knowledge already on the 5min I watched. I'll finish watching this when it's not 1am 😆 Cheers

[@corey_nz]

Finished watching. Amazing level of detail you go into on so many aspects of this. Thanks for cutting this up and filming it so others can learn. Cheers

[@martin-vv9lf]

you say there were two water pumps, was that a way to cut costs by reusing the existing part from the FVA, or was there a redundancy benefit, by using two isolated coolant systems.

[@EngineeredtoWin]

I would say simplicity in pipe routing and knowing the coolant thermal capture capability from the 4 pot...not much to do with redundancy as if one side failed it was all over either way.

[@pazsion]

to overcome a design flaw they couldnt correct without eebuilding the whole thing. at some point it didnt cycle through properly and was blocked off... so put another on the other side. 🤷

[@martin-vv9lf]

@@pazsion I see. a single water pump would be in a central location on a v8, and a different spot on an inline 4, most likely the same reason as to there being two oil pumps

[@oikkuoek]

The big holes on both ends of the head are for incoming cooling passages. The holes at the exhaust side are exit cooling passages. Depending on your demand for coolant extraction, turbo or non turbo, you can plan your cooling system accordingly.

[@UNTITLED--T]

This is amazing. Thank you!

[@ralphmillais5237]

Fascinating video!

[@Hydrogenblonde]

Very educational video. You don't see one of these heads every day.

[@gafrers]

Great as always, very interesting

[@EngineeredtoWin]

Thanks...

[@michaelgideon8944]

Thanks for this. The DFV port changed how engines are designed. The tumble swirl induced is the secret to make the 4v head work. The previous efforts needed far too much timing to work at high rpm. It's well documented in the books about Coventry Climax. Id like to buy a section of the head, specifically the intake port.

[@EngineeredtoWin]

Mike, yes...a major break through indeed. If you contact me through my email in the about me section on here I'll gladly organise a port part for you. There is already a complete cylinder section gone to UK but plenty left yet. It would be good for you to call it before I do the cuts for the framed displays.

[@andrewwmacfadyen6958]

Yes working on a Rover K the other day thinking how big and influence the DFV was.

[@stevesloan6775]

I love the left right bank replication. Ford has defiantly had a history of good design interchangeability between engines.

[@camshaftcasting1451]

Fascinating and insightful as ever! Thanks!! 2 queries ... Is that right about a "two piece water jacket core, glued together", Brian? It looks to me like it can be a one piece jacket with a core box split line, in the usual place. I think the tappet higher up is used to accommodate a bigger spring pack diameter given the (bore-bore constrained) tappet diameter. If the tappet was over the spring there would be no bore wall between tappets at the top. The spring won't fit in the tappet so the tappet has to move up. Assembly clearance: may not be a big deal. The spring pack and retainer *appear* to fit (just) down the tappet bore, although I'm only looking at an arty x-section drwg. I think the easier assemblability is a side-benefit.

[@paulelephant9521]

New Brian Garvey video, oh yessss! I'm certainly no expert, but I absolutely love these in-depth examinations on F1/high level racing equipment, genuinely fascinating.

[@keithstudly6071]

Was there significant changes in the DFX heads or was it just a short stroke version of the DFV? I would love to see a similar examination of a Offenhauser a it was reported to be extremely difficult to cast and with the head being integral with the cylinders it was complex. Reportedly only one foundry in Oakland California cast all the engines and that was over a period of some 34 years. Other shops tried. with the factories encouragement to produce them but none were ever produced that were usable.

[@Flaviocdr]

Very good work for engine lovers but it lacked an image of the combustion chamber intact. But it was still great to see a Cosworth DFV inside.

[@briankilpatrick6039]

Included valve angle and 4v so pent roof..... Kind of self explanatory

[@jdoe9518]

@@briankilpatrick6039 32 degrees included valve angle for DFV and FVA series. BDA series had 40 degrees included valve angle. Shallow chamber pent roof using finger followers have been as low 12 degrees.

[@carlt570]

Looks like a very thin bottom angle on the inlet seat ? Thank you, great vid very informative

[@MrCrabbing]

Class bit of engineering that head, makes me wonder if people work to the best spec they can to produce the best product and computers work to a minimum spec.

[@MrCrabbing]

Great to see you, have to watch later on though.

[@gerryellis3847]

The FVA 1600 was based on a Ford block. But there was an 1800 sport car version of the engine, the FVC. However, Cosworth built one short stroke engine of 1500 cc. The FVB put it in an F2 car (possibly a Lotus 48), entered it in a small race for Jim Clark. He won. The FVA gave circa 220 BHP. If the FVB gave 200 BHP and was drivable, the DFV got the final go-ahead.

[@michaelthompson1110]

I have read in a book that the Cosworth DFV was cast in Ford’s own Dagenham engine plant. Is this so?

[@markshepler7525]

Lots of good insight

[@uffebrentsen744]

Really like how you pinpoint and explain all of the features of the head. Just wondering why you got the head cut and split using wire EDM. By doing that you would not have lost that much material off of the head. Did you had any thoughts about using EDM or not?

[@EngineeredtoWin]

@uffebrentsen744 EDM is expensive and I don't own one...I can live with 2mm waste...

[@Павел-у4р4ч]

Great video

[@tturi2]

Im studying mechanical engineering, any insights on how your first few years went?

[@Preen59]

This is a fantastic video. Your name is familiar.... Did you have a forum thread on casting on Club GTI 10 or 15 years ago? I remember spending hours upon hours reading it.

[@EngineeredtoWin]

Thats me!

[@Preen59]

@@EngineeredtoWin that's unreal! It's so good to see you still diving into this stuff. Being a toolmaker, I have loved all the information you've put on the web over the years. Thanks for sharing it.

[@shoominati23]

Hey Brain, do you reckon you can get access to the 1994 Mercedes-Ilmor 500i Pushrod Indy engine? Just on the off chance.. 😂

[@jkent9915]

“Mightn’t” not a word you hear in America every day.

[@tiitsaul9036]

Very interesting. Thanks

[@qivarebil2149]

Making a common core for the intake and the water, sounds a bit strange...?

[@r8oscar]

Came from Facebook 👍👍👍

[@pazsion]

i was wondering why anyone would cut this up, and then you show the combustion chamber 😳 wow.

[@321-Gone]

Is it an Irish thing to spend more time explaining why you would or won't talking about something than just talking about it? My Latin wife wishes I had that power. Great video.

[@EngineeredtoWin]

Yes...definitely an Irish thing...also applies to giving tourists directions!

[@hartmutschubert7355]

Seen Toyota F1 head casting cuts with even worse wall thicknes differences. The guy just said it does not have to be pretty but it has to be within the specs. The symmetry of the DFV design will help with anisotropic shrinking. It is just a different approach to quality.

[@rolandotillit2867]

Straight ports are still converging ports because of the boundary layer that builds up on the walls, the thickness of the BL is dependent on wall texture as well. Cross section area is simple physics, the smaller the duct, the faster the air goes to a point, and the bigger it is, the slower the air goes. Air flow into the engine is at its most efficient around 350ft/sec, above that and the flow speed starts working against you(turbulence, separation, internal friction from compressibility). Below that and the flow doesn't have enough energy to fill the cylinder properly. As far as dimples, I'm against them, it's a time consuming permanent change to the port, and you can make similar gains for much less money time effort and without permanently modifying the cylinder head. That said, they do work, for example most Hondas have oversized intake ports, so adding dimples will make the port effectively smaller and you'll gain flow velocity as a result. You can't apply that on an RB head, where the ports aren't as oversized from the factory, adding dimples there is worthless.

[@finlaymcdiarmid5832]

I heard the highest port velocity that is efficient for high rpm is 680fps. Not 12k rpm high, closer to 20k rpm.

[@rolandotillit2867]

@@finlaymcdiarmid5832 That's about as high as you can realistically make intake air speed to be honest. Compressibility effects at that speed means the diminishing returns keep piling up. At that flow velocity you're basically running 3-4psi of boost just from the induction pressure waves. To get that to work takes MEGA budget. It's little wonder those engines didn't last that long.

[@barenekid9695]

Interesting mostly...vignette in to 60's State of the art.

[@deanstevenson6527]

🥝✔️

[@SharonDavis-v6n]

Taylor Ruth Robinson Timothy Jones Brenda

[@markshepler7525]

Ok....American...I can barely understand your brough.....Irishman