Making the Weber DFAV 32/36 carburettor work

Discussion in 'How To' started by mikedjames, Dec 18, 2012.

  1. I thought I would start this up as I progress towards making one of these carburettors work.

    The Bouncy Bus has a 1600TP engine. It had a worn out 34PICT-3 carburettor fed from an oil bath air cleaner. The exhaust fed into a 4:1 header and an EMPI dual quiet pack exhaust chosen by the previous owner. One thing that became clear was that this combination was not working well.

    Rather than buying a clone 34PICT3, I bought the DFAV kit from VW Heritage. On the box there is a sticker saying "call this number to find out about correct jetting for your engine" . I did and they told helpfully told me to go and find a rolling road as they are hard to set up.

    Instead of doing that I decided to go for a home brew approach, and do some research.

    But first I slapped the carburettor on and went on a 1000 mile trip. And it turned in about 20MPG and cruised at 67MPH. It idled badly and left black marks all over the rear bumper. I had fitted the pretty chrome filter on the top and left the oil bath behind.

    The exhaust header and tailpipe were replaced just before the trip because I blew up the rusty exhaust with a backfire.

    Research 1: You need a lot of heat to vapourise petrol in this carburettor setup. Heat comes up to the base of the EMPI brand carburettor support riser, from the exhaust through a pipe connected across the outlets of cylinder 2 and 4. Not perfect, it should go from 2 or 4 , up to the carburettor base and then back down to the tail pipe to keep hot gas flow through the inlet manifold. (reference

    As it was, there was an increase in top speed and a lot of lurching and cold running splutter and kangarooing. On the Autobahn no problem.

    Increasing preheat : I decided to reinstate the oil bath air cleaner as I had one. It originally provided a thermostatically controlled source of warm air from the cylinder head, as required to help the carburettor.

    But I needed an air box to adapt the output of the oil bath air cleaner
    to the large carburettor top panel. So first I made a prototype out of pvc pipe and tumble drier vent pipe held together with polyurethane glue. This improved the running and reduced the lurching. Next I made one out of aluminum sheet pop riveted together and sealed with fibreglass car body filler then painted. This bolts onto the top of the carburettor leaving just enough clearance to lift the carburettor with air box attached over the studs in the inlet manifold. Alternatively I have seen air boxes for this carburettor for $149 on ebay which allow you to connect an air hose.

    End result is the carburettor and the inlet manifold do not feel cold to the touch any more when the engine is running.

    I noticed that once the vacuum hoses were all connected up, the engine seemed to run at higher max RPM if it sucked hot air from below through a 50mm hose rather than through the rather tiny nozzle on the air cleaner. So taking notice of the fixing spring for the actuator on the bottom. I cut 20 mm off the air cleaner inlet nozzle
    making the hole bigger looking more like the size of the hole on the later plastic air filter boxes.

    Then I bought a few jets off Eurocarb - changing the out of the box jetting which is pretty poor and where I couldnt set the idle mixture at all.

    I first changed jetting to Primary Idle 50, Primary Main 135 and Secondary Idle 60 and Secondary Main 140. Still smoky and low economy but the lurch when you put your foot down started to reduce.

    Then I finally got round to fitting the PLX devices Inc. air fuel ratio meter. This uses a wideband lambda sensor in the exhaust to gather information about the fuel/air ratio assuming that the engine is working reasonably properly. This cost $200 via ebay. Which by the time tax and duty was paid was about £200. Thats the complete kit including dashboard display, engine bay computer box, and lambda sensor.
    In order to fit this I bought a welder and learnt to weld well enough to get a big lump of metal welded to thin pipe.
    I had to extend the cables from the computer in the engine bay to reach to the dashboard. I decided to use 4 core screened audio cable as this was lying around and there are 4 wires and an earth going from engine bay to dashboard.

    From reading around , I discovered that there is a target Air Fuel Ratio of 14.7 used for best emissions in mass-produced vehicles but this corresponds to maximum exhaust gas temperature, not peak power or efficiency. And the VW engines like to run a bit richer because of their design. But when they are not working hard, then running leaner can be more efficient.

    And on my first drives with the AFR meter, it showed how bad things were. While warming up the engine ran richer than 10:1 and it spat soot and water. Warmed up it was not much better: driving up a slight hill it worked around 11:1 and on the motorway it might flicker to 13:1. Only time you could get it to run lean was stamping on the accelerator, which corresponded to a lurch.

    Now changed jetting to Primary Idle 45, Primary Main 125 and Secondary Idle 60 , Secondary Main 150. This is similar to some recommendations on aircooled. net. While cruising, the AFR sits around 14-16:1 unless wide open throttle where it drops to 12:1.

    I am not finished here but already I can state :
    The engine runs smoother, accelleration is not all confined to a burst around low throttle. Surging and lurching has mostly gone despite the AFR still becoming quite high when acellerating. The engine runs quieter and cooler now , about 10 degrees C as far as I can tell.
    bluerustybucket and paradox like this.
  2. Since doing this, I have run through a few tanks of fuel and got a feel for how it runs at high speed along rhe M28. ​
    I bought a bigger secondary idle, size 70 and set the primary air as 160 and secondary air as 180, according to recommendations on Still rich at 70mph but better. Around 60mph on a level road the AFR hits about 14:1 which is sort of OK.

    It will go up the Guildford bypass at 50mph from the north in 4th gear. But it runs 2 degrees C hotter so now I am focussing on the missing small bits of tinware -under cylinder deflectors and the right hand side tin that goes below the main rear tinware.

    The lean spot only happens now if you really stamp on the throttle so the engine cant keep up.
    paradox likes this.
  3. After a lot more driving I now have a 137 secondary main and a 80 secondary idle fitted.
    This runs well with a Just Kampers stainless single quiet pack.

    P main 125
    P idle 45
    P air 160
    S main 137
    S idle 80
    S air 180
    paradox likes this.
  4. Changed secondary main to 140. Slightly lean mid-throttle.
    paradox likes this.
  5. Changed Primary main to 130, secondary main to 150. This flattens off the mid-throttle lean tendency even more. Runs a little cooler.

    Tendency to flat spot almost completely gone - only when cold and I floor it.
  6. This would be useful if I understood any of it :-(
  7. excellent write up, usefulinfo as we have one of these carbs!

  8. I was thinking just the same, but very good write up.
  9. Good write up.:)
    I used to just tune mine up the same as my old Cortina. :confused: It (Bay) used to break down with sooted up plugs which would result in a ride home on the back of a recovery flat back truck, I was getting about 6mpg if I drove it hard along the M57/ Knowsley Expressway!!! ( I worked 6 miles from home and if I was late for my shift it would happily guzzle a fivers worth of petrol!) . Then I bought a once used, second hand carb and manifold which cured the sooty plugs problem.:cool: 1+1/2 turns out with the mixture screw then a fiddle with the tickover and slight adjustments with the mixture again to even it all out. :)
  10. So if you were offered twin dells would you bin this carb off ? You still have flat spots and with such long runs from the carb it's never really going to be great with all the best will in the world .
  11. Having got this far with this carb I would not change it and start messing around again. It makes the bus go fast enough and accelerate better than with the original 34PICT-3
    I dont have to mess with balancing and I only have two sets of jets to mess with..

    I have made a quick video showing the completed installation.

    There is actually enough space (and 'give' in the sound deadening ) above the black air box to unbolt the carburettor and lift off the box and carburettor together. The prototype was about 1.5 cm lower and was a lot easier to remove as you could unbolt the box from the top of the carb.

    I used the PLX AFR meter to keep tabs on the mixture, and used the graph of the last 30 seconds shown on the DM-6 to review e.g. what happened as I pulled out of a turning , and also used a daily commute to repeatedly drive a route consisting of motorway and hilly back roads so I could check both the top end and the mid range performance as I made changes.

    The original out of the box experience with the 32/36 carburettor was that it was better at speed than the original but the flatspots and lurching were terrible. But it was good enough to go 1000 miles like that without a problem. The real problem was the trail of oil drops looping across the low countries to Cologne and back.

    The flat spot is now really only while the preheat air is cold. Once that is warmed up it pulls well enough to get out on to busy roundabouts without fear of flatspot . I dont see it as a problem anymore.

    If I was going to change from this setup on a 1641 I think I would go to FI.
    Last edited: Oct 4, 2013
  12. Great write up @mikedjames ....I've fitted one of these carbs myself, after faffing around with IDFs (and a borrow of @Dicky's dells). Did a fair bit of research on the vac signal and accelerator pump side of things , which are the only flies in the ointment as far as the Progressive's concerned. You might be interested to know that Pegasus do a twin spray bar (so the pump jet injects metered fuel into both barrels. as opposed to just the primary (which reduces stumble when you floor it). Also, came across a very interesting article about tweaking the vac signal to suit the 034 (005 etc.) dizzy. The prog's vac signal is a below that specified for 034, 005 etc. , but it can be "derestricted" by opening up the aperture leading to the port...the actual port is the little hole just below the butterfly, people mentioned opening up the hole where the brass spigot small increments...yet to do it on mine but will be high on my list of jobs to sort the bus out. I'd like to get hold of a good vac gauge so you could measure the draw at idle....I'll keep you posted...
  13. Interesting about the dual spray bar - I am thinking also of getting a larger secondary idle jet than the 90 that is in there at the moment.

    A followup - I was playing with setting the idle mix back to 14.7:1 after I had tweaked it to 13.5:1 because of some exhaust leaks on a trip round Cornwall a while back causing backfiring and distress to the lambda sensor (it got too hot and went offline going down long hills).

    At low throttle this improved response. The leaner running remains up to around 50-60mph on the flat.

    I also increased the main jet to 160 to see what happened. Seems like there was an improvement but the traffic on the M27 in the rush hour got in the way around 75mph..When running like this the AFR drops to 12.5:1 which is a good idea as it is cooler.

    I do not see a lack of vacuum advance with this setup - it makes it to 30 degrees at 3000 RPM. Possibly the fact the air intake comes through the oil bath and a 50mm hose rather than straight from an aircleaner increases the signal.

    I looked inside the air cleaner intake with the engine running but cold and the hot air flap was closed.

    Turns out the peg on the vacuum capsule had lost its top, so as the vacuum built , the flap went up and then dropped back down. Discovered the manifold vacuum I had connected to was pretty fierce. So I made a scissors spring out of 0.8mm stainless welding wire, fixed through a 1mm hole drilled across the peg just below the cup where the flap would have engaged, and then looped through the little loop in the end of the flap.
    So now the vacuum capsule thumps down and the flap opens fully and the spring extends about 5mm.
  14. I have worked out the downhill backfire - it is because the fuel pools forwards in the carburettor because the float chamber's long axis is along the length of the bus. As the fuel pools forwards, it drops in the primary part of the float chamber.
    So the fuel feed to the primary idle jet vanishes and it goes lean. Oh well. Just need to tap the throttle from time to time.
    Probably also a problem with other two barrel carburettors. If you read the Weber book they tell you to fit two barrel carbs with the float chamber across the width of the car not along the length.

    Since the above, a thousand miles or so. I changed the secondary main to 150 again as it felt a bit dead in the midrange - basically all of the secondary throttle operation just felt a bit 'meh'.
    I also cleaned the distributor and replaced the missing slider bearing in the vacuum advance mechanism with a brass panel pin with a flattened head, so that now works.
    Feels like the secondary barrel does something - at one point as I started the hill from Portsmouth to the West on the M27 tonight accelerating from 40 in 4th gear, I did a double take because it felt like it was in third when I put my foot down at 40 mph.
    Usually I need to go to 50 in third and then very gently accelerate on that hill in 4th. It still maxed out at about 65mph though as the vacuum advance gives up with wide open throttle.
    Last edited: Jan 2, 2014
  15. interesting about the primary idle jet..if it's located at the back of the carb, wouldn't it be swamped rather than starved when you were going downhil?
  16. Just a note on my findings with this carb, assuming it's the EMPI copy rather than an original Weber carb. Firstly, there will always be a delay / flat spot to a certain extent because of the length of the inlet manifolds and the time taken for the extra fuel-air mixture to reach the cylinders. Secondly, the accelerator pump only works on the smaller primary venturi. This means that when you floor the throttle and open up both butterflies, the squirt of fuel needed to enrich the mixture before the airflow gets going only works on one of the (now) 2 open venturis, leaning out the mixture as you are trying to accelerate. The best you can do is to change out the accelerator pump for an original Weber item, which gives a better squirt of fuel than the EMPI version and so reduces the leaning effect when accelerating.
  17. Flat spot can be reduced altogether with proper pre heat and good jetting/ignition timing....also, by attending to the pre-heat side of things using stock air cleaner (with vac signal and tinwear) plus a well designed plenum (instead of the crappy pancake filter) you can actually improve torque just where the flat spot would normally occur
  18. It's not so much about flat spot (as in a dip in the torque curve), rather that the throttle response is poor as an inherent part of the carb / inlet tract design. Pre-heat is addressing the issue of condensation in the long inlet tract when it's cold, but is limiting performance by feeding the engine warmer, less dense air. Twin carbs (single or twin choke) is an inherently better solution, rather than spending time working around basic design problems when using this arrangement.
  19. I agree in as much as for out and out performance and fuel efficiency, one carb per cylinder is almost always going to be better than a centre mount, but then if that was the only deciding factor, we'd all be running EFI. Twins also have their inherent issues in terms of initial cost, 4 sets of jets, and chokes, synchronicity, set-up, and maintenance. Both designs require dialing-in on a Dyno to get the most out of them. As I see it, the disadvantage of one design tends to be the advantage of the other, and vice versa.
    I believe the reason many people dismiss progressives is simply because they don't spend the time initially, or use components best suited to it's characteristics.
    If you stuck a pair of dells on, fresh out of the box, chances are they'd be out....why should a progressive be any different? I'm not saying one approach is better than the other, but it depends on your budget, what sort of performance you're after and also your driving style. I'll be happy if I see good fuel returns and drivability...if this can be achieved with a single or progressive then why have the hassle of twins? As it stands, a stock Pict would be undersized for my engine requirements, otherwise, I'd be using one....if a major power increase was my main objective, I wouldn't be buggering about with an old aircooled....
    You're absolutely right in saying that the performance potential of a centre mount is limited by design, but there are other factors to consider....

Share This Page