Vacuum measurement at SU carb

For months I’ve been working on my 67 E-type to get it running again after replacing a rusty gas tank, fuel pump, lines and carbs rebuilt. I now have the car idling and driving but having a problem at 3k RPMs and higher. The car has a new 123 distributor with vacuum advance. I think it is not getting an advance at higher RMP but haven’t had anyone to help me to operate the throttle while checking the timing underneath the car. In the meantime I tried to check the amount of vacuum on top of the front carb that goes to the distributor with a gauge. I don’t get any reading at any RPM. However, there is a good vacuum going to the brake booster from the manifold… So my question is, what kind of vacuum pressure reading should be found at the port on top of the front SU carb? I can blow through the port and know its not blocked. What can’t I get a reading of any value? Thanks in advance.

The ported vacuum on the front carb should be atmospheric pressure at idle and atmospheric pressure anytime the throttle is open more than about halfway. You should have some vacuum at that port if you raise the RPM to about 3000.

Try putting back your old distributor. If it then runs correctly, that suggests your problem is related to timing or your 123.

Are you sure the problem isn’t that you have your carbs set too lean? What you describe is exactly what you would see in that situation.

The XK engine likes to run with an AFR of around 12 or so, or with an exhaust CO at idle of 4 to 5%.

Try pulling on the choke when it starts to break down at 3000rpm and see what happens. If it improves, that suggests it’s lean.


Bill and Andrew. Thank you for the replies. The old distributor is shot, hence that’s why I have the 123. I’ll work my way through the issue but was surprised at how little if any vacuum comes off of the port on the front carb. Seems so low or inconsequential that it would make any difference inside the distributor. I also don’t have an exhaust gas analyser but do have the old fashion color tune. I need to start exploring the high tech 123 on my iphone so see if any old guy can figure it out. Perhaps there is no base advance curve and that has to be programmed. I’ll keep you posted on what I find.

The 123+ distributor is a marvelous device that can positively transform the engine, but you do have to follow the directions.

And one of the most important to install the 123 with the engine set at either TDC or your desired initial timing point. Otherwise you can end up with plug wires not ‘clocked’ correctly on the cap. On that note, once you do the initial install and rotate the distributor body per the instructions to a firing point (for cylinder #1-closest to the firewall, remember!), you likely will be resetting the position of the plug wires vs. the original distributor.

Btw-I connected the blue (ground) wire directly to distributor body via a small screw hole that was already there, as the distributor is well grounded via the clamp.

I set mine up at TDC so I had 0 degrees advance when cranking the engine over, but once running above 500 rpm the advance jumps to 12 degrees.

It is also important to give the distributor a base curve! The instructions that come with the 123 distributor should be adequate, though a bit confusing when it comes to vacuum advance. Just Googling “123+ ignition install” should get you a lot of good information. There are also a number of threads on this forum that give good suggestions for a starting curve. Or just pull out the workshop manual and use the distributor information there to configure the advance set points. Here’s the current setup on my engine, which is mildly modified. On a standard engine you might want to back off the timing a bit.

As for measuring vacuum at the carb, it will be difficult to see a vacuum reading on a typical gauge when standing there with the engine running, but you should be seeing the needle at least flicker when blipping the throttle. Try revving to about 2500 or so and slowly bring the revs down and you should get a reading. Or just trust that the vacuum signal is present and hook it up to the distributor.

1 Like

The ported vacuum is “sucked” through a tiny hole,so it takes a while to draw air out of the gauge’s hose, etc. in order to register the vacuum. If there are any leaks in your gauge, you will never achieve a vacuum. Try sucking some air out of the gauge hose with your finger just over the end until it registers, and see if it “holds” the vacuum when you cap the hose with your finger.

I think the SU vacuum port might be even more restrictive than most. Many cars have a standard vacuum hose connecting the port to the dizzy, but Jaguar fitted 1/8 inch OD copper tubing, which has very little dead space from which to withdraw air before the vacuum is “seen” by the dizzy. IMHO.

You don’t say which 123 you have, but judging by your reference to your iphone I presume it’s the 123 tune+ which communicates by Bluetooth to your phone. If so, you definitely need to program the 123. The “base curve” is unlikely to do much more than allow you to start the engine. Lots of info in the archives about setting up curves to suit your engine. The 123 Tune+ is not a “drop it in and forget about it” replacement for your old Lucas distributor.

1 Like

Steve, I have had lots of emails back and forth with 123 Sebastian, who has been very helpful. His figures for the vacuum setting are almost exactly as yours. Interestingly, he did make the comment to me as to why I was not setting my ignition advance curve as yours. Mine is set at 10 degrees BTDC. But I can see the logic in your curve, it does make more sense. So, I will try. But one interesting point to discuss. On page 338 of the Bentley book the vacuum numbers say that there is no vacuum below 4 1/2 inches, so it gives the impression that the vacuum start should at nearer 500 rpm than 1500 rpm as your vacuum graph. What do you think ? The 123 is complicated to understand as the vacuum works on absolute pressure, so the 16 degrees of advance looks extreme, but it is the conversion from the numbers on that page.


I don’t exactly recall why I didn’t start the vacuum advance until 1500. If the vacuum source is manifold vacuum then lower would make sense. If it’s ported vacuum (above the throttle plate, which I believe is the case here) then there would be no vacuum to speak of at idle or just above. And I’m not likely to be cruising at less than 1500. But I’ll give it a try next time out.

That’s the beauty of these 123+ things-only takes a few minutes with your phone to try out different settings. That’s how I arrived at the 12 degrees BTDC for idle: by adjusting the timing while it was at fully hot idle (the high tech version of tuning by ear?), that’s where it was happiest.

Hello Steve, I have used your curves and setting. They work very well.
The only change I did was to reduce the maximum to 34@4300 and above on your recommendation. My correspondence with 123 in Holland, did recommend to set timing at TDC with immediate jump to 10-12 at 600. But I was always just setting at 10 BTDC then reducing the other settings to take that into account. But I can see your graph is better. The car feels very smooth with good torque and power.

If you want a reliable vacuum signal, then take it from the manifold. This means the vacuum signal is averaged out and not subject to individual variations.

Manifold vacuum also means you’ll be less retarded at idle, so will be running more efficiently and generating less heat.

The only difference between ported vacuum and manifold vacuum is that the ported vacuum is manifold vacuum but with a missing or much depleted signal depending on throttle position. This also means it cuts in and cuts out more aggressively as throttle is opened or snapped shut. This is why I think people seem to think it is preferable to manifold vacuum.

Below is a link to graph of how much ported vacuum signal you get at various rpm as compared against ordinary manifold vacuum, from which it derives. The air pressure recorded at the ported vacuum port is mostly closer to atmospheric pressure in the airbox so there is less vacuum signal than the manifold vacuum would be, but the table doesn’t doesn’t tell you anything about the rate at which it cuts in or out.

Also appended is a graph of throttle opening versus vacuum (i.e. engine load). You’d expect the very top right hand corner to be 100%WOT. At 1700rpm, under full load, the throttle is 51% open, i.e. the engine is effectively at WOT if only half open at 1700rpm because the engine is getting all of the air it needs at that throttle opening.

These should give you a feel for how much vacuum signal you can expect under what circumstances of throttle opening. Numbers are absolute- 100kPa = 1atmosphere at sea level, so 60kPa = 40% vacuum.

kind regards

If you want to get a handle on how ported vacuum cuts in and drops off, I can send you a datalog and tell you which data fields to compare against each other as the car accelerates or throttle changes.

kind regards

Marek, Steve and John. Thank you for your comments. This has quickly gotten over my head. I have successfully programmed the 123 with the centrifugal advance curve starting with the base 10 degree BTC as stated in the manual and peaking at 32 degrees at 4k RPM. (My car is a 67 coupe) I’ve only had the car running in the garage but it is revving much better at higher RPMs than it did before so clearly the lack of an advance was the problem. Now my confusion is with the MAP curve programming. Since I couldn’t get any reading with a vacuum gauge at the carb port, I don’t see how I can program a MAP curve that relies on that vacuum. I certainly have a strong vacuum coming off of the manifold and could tap into it if that is desirable. One thing that I don’t understand is, if the engine is getting all the advance it needs from the centrifugal advance, then what is the purpose of the MAP curve and is it needed? A distributor without the vacuum port relies totally on the centrifugal advance so can’t I do this with my distributor and not program the MAP and just block the carb and distributor ports off? I’m also still confused about setting the static timing before programming the curve. Are you gentlemen in agreement that the timing should be set at 0 TDC and not 10 degrees advanced and then program all the advance from that base? Finally, if the initial setting is 10 degrees advanced and the maximum curve is 32 degrees, does that mean the real number is 42? Thanks again.

The reason why there is a vacuum related component to the total advance you would want to command is so that you can have more advance at medium to low engine loads, thus running the engine more efficiently, yielding more power and economy at anything below wide open throttle. If you have no vacuum related advance variation (the amount of vacuum is a measure of engine load), then the only safe option is to set it up conservatively for maximum engine load. This means low-ish amount of advance at full load (as it is dangerous to over advance an engine under heavy load) and this setting carries over to all easier loads, even though a higher amount of advance is tolerated at low to moderate engine loads.

In short, you will spend 99% of your time running more retarded than necessary, lowering fuel economy and needlessly generating more heat.

kind regards

I kind of understand this. But if one can program the centrifugal advance curve for any RPM why can’t this be done adequately without the vacuum advance? Is the engine load variation so dramatic that the only way to get it correctly dialed in is to have the vacuum advance provide the fine tuning that is not possible with only the centrifugal advance? If this is the case, then I understand it. Fortunately, my E-type Jag is not driven that much - like 2,000 miles in eight years. So the perfect tuning is not a major deal, but as car enthusiasts we want it to be good. What is that saying, “don’t let the good be the enemy of the perfect?” Thanks for the discussion.

See if the port in the carburetor is plugged up.
Can you pull a vacuum on the port? That’s one easy way to find out if it’s plugged.

W. I couldn’t get a vacuum from the carb port. This was discussed earlier in the thread. It’s a very small diameter hole. I can blow through it with the engine off so I think it’s not clogged. I have a strong vacuum off of the manifold. It was suggested to tape into the manifold vacuum line that goes to the vacuum tank, but I’m afraid this would be too much vacuum. I don’t know. Again, I’m not sure why I even need the vacuum advance since the engine seems to be running fine with only the centrifugal advance.

A couple of things to consider, if you suck on your vacuum gauge does it give you a reading? That proves the gauge is OK. Are your carburetors synchronized so they all begin to open at exactly the same time? If the front carb opens late then the engine will rev up and you won’t have any vacuum.


Don’t overthink this and try to engineer a better solution when you don’t fully understand what’s going on. The small hole on the front carb, if not blocked, WILL provide a sufficient vacuum signal to any distributor. Do not modify or enlarge the hole.

If you want maximum smoothness and fuel efficiency at part throttle cruise (which is probably 90% of the driving we do) then hook up the vacuum port on the distributor to the factory vacuum line on the front carb and run the curve I sent you. If you switch to straight manifold vacuum (which really is preferred, but forget about that for now) then you will need a different curve, which will take some experimentation to get it dialed in.

Here’s a useful read on manifold vs ported vacuum and the desirability of vacuum control on top of mechanical control of advance. But again, don’t jack around with things until you fully and completely understand the dynamics or you undo all of the very competent engineering that the factory and 123 have provided.

Steve. Very good. Thanks for sending the write up on vacuum advance. The author says there is very little written about this subject and I find that to be true. I shall program your curves and test drive the car this weekend. Thanks for your time and help. Etch