Twin Seq. Dry sumping for beginners :D

[Turbo-Brown]

I know in my heart of hearts that the PIC thing is the better option, but I think for the £2 it costs that the 555 solution is worth looking into, just incase it does do what I need

The controller (whatever it ends up being) will just need to look at one pressure signal, the pressure at the outlet of the large turbo.

If the air valve were to stay shut, the pressure at the outlet of the small turbo would always be higher than that at the outlet of the large as the small one would be multiplying the large's pressure.

What I think I need is to sense the pressure at the outlet of the large and, when biggy starts boosting, start the air valve opening. That should let the pressures equalise between the two outlets, gradually unloading the small turbo and preventing the small from multiplying the large's boost.

Once the pressures equalise and the large turbo is boosting more and more, the pressure rise will tell the controller to increase the opening of the air valve.

The valve will be normally closed and an increasing 'on' time will cause it to open, so say a 95% duty cycle will mean the air valve's about as open as it can be.

I think that Emerald recommend the frequency be set to somewhere in the region of 15-20Hz for the boost control solenoids.

That sounds like a good idea with the measuring of the adjustable components and replacing them with fixed value jobbies once the development's done too!

I think that the pressure dropping when the throttle's shut and the DV is open should cause the air valve to close and the whole cycle to start again when the boost starts to build. It'll be interesting to see how it does behave as I can see there being some thrashing as pressures equalise.

Think I'm going to go to the parent's in the evenings this week and finish off the butterflies etc, then start to concentrate on the air valve as I think it'll help me visualise things a bit better having the air valve to play with. Can do some experiments with controllers and so forth then too I guess, just to see how it behaves.

Will keep the heater controller idea in reserve for now, not sure how adjustable that would be or how easy to get a pressure signal into.

Thanks so much for the input to everyone who posts, it really is very much appreciated!

[jkarran]

Ok, if you're only wanting to look at the big blower op pressure as the reference signal that simplifies matters. You'll be able to do the core of that that with some op-amps (1 or 2 chip) and a few R's and C's, other than that you'll need a robust power supply and an output driver (MOSFET based). Let me know if you want some circuit ideas.

jk

[frankiej]

I believe that is cash being forced into the engine and bhp being pissed out. Hahaha. That's great Liam

[Turbo-Brown]

I'm very interested in circuit ideas JK! Be really cool to put something together which I can see working, flashing a light or whatever

Mini-update too: Been home to the parents yesterday and today as it avoids having to spend an hour sitting in traffic on the way home from work. Made another 4 butterflies so there's just one left to do. Hopefully get that done tomorrow and that's another component group ticked off the list

[jkarran]

Right, I'll sort you something out, increasing PW with increasing pressure, fully adjustable thresholds.... and a flashing light... should be fun. Do you have a sensor in mind?

I finally figured why your PIC was misbehaving, switching leds off as you switch others on. It's to do with the read-modify-write sequence used to update the ports on PIC16xxxx devices. Not a problem on the 18Fxxx I'm more familiar with. As you may have guessed I have a misbehaving Pic16 sat on my desk currently

jk

[Turbo-Brown]

I really don't stand much chance in the electronics world! There's hope for PICs in the car yet though as I was planning to use them to control things like the noise control, heated windscreen and so forth, only as latches for push-to-make switches though

I was thinking a sensor like this:

235-5841

from www.rswww.com could do the trick. It's not going to see vacuum, and to be honest, one which maxes out at 15psi might do the job just as well, as by that pressure the valve would need to be well and truely open

Thanks so much for all your help mate!

[Turbo-Brown]

Hooray! All the butterflies are done now, next job I guess is to finish off the spindles and once that's out of the way, the throttles are nearly done barring the linkage

Had the day off work today so I've been mocking things up on the spare head I have in the house.

I had planned out in my head where all the major components would live and, by and large, that's worked out alright. What I'm left with though is the wastegate for the large turbo, and the air divertor valve fighting for the same space. It's probably a good idea to keep the wastegate as close to the manifold as possible but it looks like that might not be possible

Anyway, I've also been having thoughts on air filtration. I really don't wanna get suckered into the whole cone filter thing as they're an awkward shape and I don't want the engine to be sucking in hot air as is the case on so many setups you see.

What do we reckon to using a pair of E46 M3 filters housed in a shiny ali box behind the driver's head lamp? Trouble is that I've no idea how big they are, anyone know?

[Turbo-Brown]

Oh dear. It seems I've been busy today!

The day began at around 11am after I slept right through my alarm. Eager to get on with things I drove over to the parents and started making bits for the not-quite-finished exhaust manifold.

I've been dreading making the Exhaust Divertor Valve (which doubles as the wastegate for the small turbo) but I thought I'd bite the bullet and just get on with it.

Anyone who's tried to machine stainless steel will tell you that it can be a complete bastard as the metal becomes harder the more you work it. If you try to drill it too slowly, it becomes so hard that you can't drill it any more. Try to drill it too fast and the same happens. Drill it with a slightly blunt bit and it hardens. Drill it with the bit turning too fast and it hardens.

Anyway, after much sweating and the repeated sharpening of drill bits / the making of some special super-stiff drill bits this is what I've come up with:




The channels are for the EDV spindle to run in. I reckon that the step in the larger channel should help prevent too much exhaust gas escaping.

Here's a look down the holes:




I welded a sleeve into one of the halves of the valve to help seal it (all the connections between the various plates are sleeved) so here's a little pic of the welds:




And also made a butterfly to go into the exhaust stream. It's 3mm thick stainless steel which I'm hoping will be resistant to warping given that it will be subject to intense heat when it's open.

It's not quite like the throttle butterflies in that it's just round and it's also quite a loose fit, reason being that I don't want it to expand and jam in the valve assembly.




Plan for tomorrow is to make up the spindle and weld that on

So, after a stressful (but rewarding) day on the tools, I went back home and sat down for a drink with one of the housemates. Of course one drink lead to another, and another....and before I knew it I became the proud owner of a 1993 740i....oops! I've wanted a V8 engined car for ages and it seems that the booze has made me buy one

The depressing side is that Brown is being replaced by Maroon now so the journey is finally at an end

[Turbo-Brown]

Hooray! Finished the butterfly for the EDV today. It was a fairly long and drawn out process which involved making a jig to hold the two spindles in line, maching two spindles up, welding them onto the butterfly, and then making up some stiffening bridges which were then welded in place.

Here's the result anyway with the assembly:




The valve open:




And the component parts. The step in the longer spindle locates nicely and prevents the butterfly from touching anything that it shouldn't, and I reckon it will give a reasonable seal much the same as the spindles used for normal internal wastegates.

[Turbo-Brown]

Some more little bits done today and even a little bit of design work.

Here's what the air divertor valve should end up looking roughly like. Unfortunately it needs to be quite compact which is why the facet angle is quite sharp It's all 3" diameter though so hopefully shouldn't be too restrictive!




Here's the bits to make one of the elbows:




And lastly one elbow all welded up. Just need to weld on the smaller tube at the bottom and the mounting plate plus clean up all the welds inside (and of course polish it all!)






Will have to try and get up at a reasonable hour tomorrow to get these bits finished!

[Yaninnya]

Unfortunately it needs to be quite compact which is why the facet angle is quite sharp It's all 3" diameter though so hopefully shouldn't be too restrictive!

Hi Alex,
To achieve the perfect curve inside you can weld the same shaped parts but from the tube with a lot thicker walls (around 3 mm). Then you can grind it into smooth curve. If you are perfectionist (I know you are!) you can grind it outside too. Some sand blasting and it will looks like factory made.
I hope you will understand what I mean. My english sucks.
Jan

[Turbo-Brown]

Your english doesn't suck at all dude! I'm not that much of a perfectionist though, it'll have to stay as a facet for now

Managed to get up fairly early today and make some more progress. Got both elbows welded together now:




Machined out a flange for one of them:






Welded the flange onto one of the elbows:




And finally welded the outlet from the large turbo to the inlet of the small turbo in place:




Lastly, here's a pic showing where the two elbows will live. Gotta order some material to make the valve section, weld an inlet stub and flange onto the other elbow and finally make up an actuator for it and bob's yer uncle

[oakey]

coming along nicely. Cant wait to see it finished! I wish I could build my own twin turbo m20

[Yaninnya]

Your english doesn't suck at all dude!

Cheers mate! But I still need to learn a lot.
Jan

[Turbo-Brown]

Been absolutely knackered since yesterday's 500ish mile round trip so haven't made much progress.

Did however manage to order the bits of silicone hose I need during the week and the good folk of Rally Design delivered on Friday so things are starting to take a more defined shape now:




As a low effort fill-in job, I decided on the size of rim spacer I need to have made up for the front wheels. The offset will eventually end up being 23mm which isn't too far removed from what I should be I think, stand to be corrected on that though.

I'm quite chuffed with the look it should give with this being the front wheel:




And this being the rear:




Lastly for today, I tested the old Cossie IC core I was planning to make into the charge cooler for leaks. If you've seen one of these, there are little top-hat pressings pop riveted onto the top or bottom plates. With this particular core, it seems the person who drilled the holes managed to get one a bit wrong as it goes straight through into the core! Presumably this Cossie must have been hissier than most as this hole's definately been there since the IC was made, it's been painted and everything

Guess I'll have to use one of the ones Brown used to wear and hope that's been better made.

[Rik178m]

Wow it looks amazing! Howmuch would you say you have spent on the engine and parts?

[Andy335Touring]

Very nice dude !

[jkarran]

Looking really good

I had to scratch my head for a moment to figure where all the plumbing will go, is there really room in the E30 for all that

One thing that did worry me (and it may not be a problem at all) is what happens to the small turbo when the air divertor valve opens? Basically most of the load will be removed from the compressor and the whole thing will be much easier to rev up right? Now the exhaust divertor will reduce the turbine power available but will it be enough or is there a risk of the small turbo surging and self destructing ?

This is basically the setup used on BMWs twin turbo diesels though so I guess I'm missing something? Sorry, didn't mean to sound negative, I just got to thinking how it works seeing the pictures of it almost together and couldn't fully convince myself that there wasn't a potential problem or at least a risk if the valves stick.

By the way, I had a look at those pressure sensors you're looking at. They could be made to work but they give a really tiny output superimposed on a much bigger voltage, makes them tricky to work with. I had a quick look on Ebay and there was a few automotive 1 and 3 Barg MAP sensors at reasonable/similar money. Something like that with it's preamp and temperature compensation built in is going to be a better bet.

jk

[Turbo-Brown]

Cheers guys

I've not spent all that much so far Rik, most of it's been my own time and so forth, probably up to around a grand so far although I've not been counting. Will keep all my receipts this time though just to see what it does end up costing!

Hey JK, how ya doing dude?

Had a bit of a trial fit of the plumbing today just to be sure, here's the piccies










Also meant I could mark out where the compressed air from the small turbo needed to re-join after the ADV, and even managed to finish off the outlet elbow. Just the valve housing and actuator to do now!






I've been having the same thoughts about the small turbo too, espeially with regards to the exhaust side. I think if I put the wastegate too near to where the exhaust gas from the small re-joins the main stream up to the large, when the WG opens the small will spool again and that might not be such a great thing as you say!

Believe the key to avoiding that will be to move the large WG right down in the engine bay so that it vents pretty much directly from the point where the gas is directed up to the large, thereby maintaining the same pressure(ish) at both inlet and outlet of the small when the WG is open.

I'd also had thoughts of the valves sticking a while back so I decided to make the actuators massive. The two actuators I'm making are a good 3" bore and 50mm stroke. Assuming my vac pump can generate -7psi, that should give a good 220N to shift things about with, and the long stroke of the actuators means the lever arms on the valves can be fairly long too.

Didn't sound negative at all dude, good thought provoking questions!

Cheers for looking into that for me, will have to buy a couple of MAP sensors I guess, pair of 3bar ones should be fine I guess, one for the ECU and the other for the ADV controller.

[jkarran]

Your ali welding's looking really trick, as is the whole thing for that matter! All polished up it's going to look really

I guess the thing to do with the small turbo is try it! Might be worth installing a mesh screen somewhere near its outlet for the first few runs though just in case

Big actuators seem like a very good idea, if the valves were to open in the wrong sequence it would be quite easy to over-rev the little blower.

If you can find OP Voltage vs Pressure curves for the MAP sensors you pick I'll draw up a simple PWM generator to get you started. Is the MAP sensor that feeds the ecu looking at the same part of the air circuit (give or take p dropped across the cooler core) as the one for the ADV controller? If so then you can get away with one sensor.

jk

[Gunni]

So have you finalized the way the system works?

I was thinking that you might want to have it controlled in regards to actual air input flow?
so knowing the flow limits of the HX27 you can move the flapper so that the bigger turbo takes over from that point before the small one reaches it´s peak efficiency using a MAF to measure the flow?

[Turbo-Brown]

Cheers JK, think the camera makes the welding look better than it is though All this practice is definately making a difference though which is good

Think the thing to do is veeeery slowly and progressively build up the load, or possibly even peg the Exhaust valve open and just monitor the output of the small turbo to start with. With the constant vacuum supply it should be possible to do that from the software too

Will buy a couple of sensors as soon as I can and do some measuring of the outputs The ADV controller is going to need to take it's signal from the outlet of the large turbo so that it's only detecting when that one's boosting. Then the ECU will get it's signal from the vacuum side of the throttles so that the ECU sees the same boost as the engine.

[Gunni]

So have you finalized the way the system works?

I was thinking that you might want to have it controlled in regards to actual air input flow?
so knowing the flow limits of the HX27 you can move the flapper so that the bigger turbo takes over from that point before the small one reaches it´s peak efficiency using a MAF to measure the flow?

Did you see my little comment?

[Turbo-Brown]

Aah sorry dude, must've been typing while you were pressing reply

My plan it to map the operation of the two wastegates against throttle position and crank speed, and have a seperate PWM controller take care of the ADV control.

It certainly sounds like a good idea to use a MAF though, but I'm going to see how we get on with the plan as it is as it's fairly simple and relies on the bare minumum of external electronics.

I think the response of the two turbos should be quite consistent so it should be possible to map the turbo functions over a period of time to run open loop.

[Gunni]

Are you going to effectively block the smaller one and not use anything from it when the bigger starts to be able to boost. Wouldn´t blocking the entry into the turbine of the small one effectively slow down the compressor as well?
So it wouldn´t matter if the large one was blowing through it as well as into the intake?


Using the MAF it would have to have a Flapper location point vs. RPM vs. Airflow to effectively work all the time.
so that absolute minimum lag is feeled and the true sequental turbo would be accomplished.

Again I must apploud you for the effort, the results will be astounding.

I think this will be my target when I do turbo on my S50

[Turbo-Brown]

I'm relying on there being a minimal pressure differential across the small turbo's turbine and compressor housings to let it just sit there doing nothing, so we can see that with the ADV open, the pressure from the H1C at the inlet of the HX27 will be the same as at the HX27's outlet.

Think the only thing which might upset that balance is the position of the large wastegate which is going to need to vent from the collector in the manifold, before the EDV.

Hope that's all going to work anyway as there's not much room for any additional valves

You'd really try this on an S50 Gunni?! Top man!

[Gunni]

You'd really try this on an S50 Gunni?! Top man!

We´ll I have you to prove the "theory" then I can just "copy"
the operations to the S50 obviously making more power everywhere,
And hopefully more drivable power at part throttle, making for quite the ride

When your car starts and is mapped, I´m coming down for a sit in.
I´ll come steal it if your not home ;)
but of course return it when the ride is over

[Turbo-Brown]

You've gotta find room for it all too!

Packaging is becoming a headache, partly because of the size of the components I want to use like these giant actuators, and partly because things like the wastegate are big anyway.

Think there's going to be an element of disassembly required to change the plugs too

Gotta say though that it's immensely satisfying getting things to sit in the engine bay as well as they do currently.

Think the next big thing will be getting the two down-pipes into the engine bay and past the steering column as I want to run a full 3" system for the main system and a parallel full 2" system from the wastegate so that I've got a really good and free flowing exhaust system which doesn't make too much noise

[Gunni]

Mod the firewall if you need.

I think it´s worth it

[e30_Turbo]

Alex,

If you can adjust the drivers side bulkhead a little, widen the gap between the block the bulkhead, will def help.

I did mine a little but as I didn't want to remove the brake master cylinder I only did what I could.

OR....... run the downpipe straight out via the battery tray through the wing, it's been done, I've seen one in action, although you wouldn't want to go cruising with the windows down

That would just leave the screamer pipe to route, which at 2 inch will be real easy.

My screamer exits under the engine near the sump, no issues so far but I've been assured I get some flames under there, save on neons I suppose

[Turbo-Brown]

I'll get it down there somehow!

If you can get a pair of 2"ers down there then a 3 and a 2 shouldn't be TOO much problem, even if the 2" needs to take a different route.

Don't really want a screamer to be honest, even with a silencer it should still be pretty loud!

[Scoupe]

Think this may have already been covered, but if not then you may be able to work out the range of rates you need for the PWM and then choose the size of capacitor that you need, and then use the size of the variable resistor to change the frequency of the circuit. Using a 555 Astable f=1.44(R2/R1+2R2)C

where:

f=frequency
R1=resistor 1 (between +v and R2)
R2=resistor 2 (between R1 and C)
C=capacitor (between R2 and -v)

R2 would be the one changed to a variable resistor if you go down that route

I can put a simple diagram of this up if you are interested in it, its really easy.

I spose another route you could go down is to use 555 Monostable circuits, or possibly even an RS flip-flop (or maybe JK flip-flop)

Again, I can put some more up about that if you are interested....if not, I'll go back to my box!



I also have a bit of experience with PIC chips and they're related elements. Although from reading the posts above, it seems that JK? (I think) has a better experience and knowledge of the things than me!

Saying that, glad to try and help if I can.



Other than all of that....looking awesome mate. Wish I had the machiening and welding skills to undertake a project like that! Keep going!

[Turbo-Brown]

Hi Scoupe, how's it going?

I'm very interested in seeing any diagrams which I could put together on strip board or something similar. Think I need around about a 15Hz ish as the base frequency and then variable PWM based on our MAP sensor.

Actually, I must make a note to test the MAP sensor I've got already.

It's a 2.5bar one but by the time the large turbo has reached a few psi the valve needs to be well on it's way to being open so limited sensor range shouldn't be a problem.

Cheers for your input dude, I very much appreciate all the help and encouragement that everyone's been giving!

[Scoupe]

Good thanks mate, and yourself?

I just think your projects brill, so would very much like to see the finished results...and as for effort....well I think you've won that round!

What do you actually have and need then for inputs and outputs, and do you have some ideas for voltages? Then I can be a lot more helpful!

You have an input from the MAP sensor, and presumably from a CAS sensor. And from that you need an output for? (I presume injector timings, but despite reading the whole topic, I'm still struggeling a little! Sorry, just want to make sure I don't end up telling you the wrong thing!) Which would leave the output as a square wave, which you would obviously need the variable pulse widths for.

So all that rubbish ^^^ meant was...

What inputs do you have?
What outputs do you have?
And what needs to vary inamongst them?

Cheers mate...good luck

[Gunni]

he´s running the engine management separately from the turbo controller.
I´d say that the pic should take it´s rpm signal from the tacho one, it´s square wave.

As a side note. when the small turbo kicks in the big one will follow right away as alot of air flow will be going trough the engine now, so the time frame isn´t that big if punched at low rpm,