Troubleshooting a hard start issue

[00Dan]

My O-320 is a hard starter, particularly on the first start of the day (It's been in the low 80s around here recently and I still experience the issue). Respecting the 10 second duty cycle limit on my starter it typically takes two to three attempts to start. I've varied my technique and haven't found anything that works WRT boost pump timing, how many pumps of the accelerator pump while cranking, etc. I'm now annoyed enough with it to dig in and troubleshoot since technique change didn't do the trick.

Some background data: Carbureted O-320 with 10:1 pistons, two slick magnetos (one impulse coupled), Tempest fine wire plugs, no primer system, Skytec NL starter.

Working through the possible causes I've come up with the following, and would like advise on a course of action.

Air: Unlikely. Once the engine is started the EGT values show no signs of an intake leak, and the engine idles normally.

Fuel: The carburetor does not show signs of leaking or flooding, unless I just leave the boost pump on and the engine off, in which case eventually the bowl floods. I did ream the nozzle out for additional fuel flow, and this resulted in a flat band of mixture control when leaning. I don't think this would be the cause given the carburetor should be on the idle circuit at cranking RPMs.

Spark: This is the first place my mind goes. I have two slick magnetos with just over 200 hours on them. Both magnetos were installed with the distributor cap off and the locking pin inserted with the engine set to 25* BTDC. The one impulse coupled mag has a standard 25* lag coupling. After a couple test flights where I suspected detonation was occurring at high power settings, I retimed both magnetos to 20* BTDC. I did so by loosening the retention nuts and bumping them back 5* (I used new lock washers, of course). Assuming this also moved the impulse coupling timing, would the 5* ATDC cause starting issues? All spark plugs had good resistance values when tested at CI back in September.

 

[blaplante]

So you "moved the magnetos back 5 degrees". How? by judging the angle of the magneto? Do you have a buzz box?

What is your cold start procedure... details!

ex:
throttle xx.xx inch open
mixture ... set where?
and crank?

Do you cycle the throttle to get the accelerator pump to squirt?
While cranking?
Before cranking?

 

[dmattmul]

Starter solenoid duty cycle

Respecting the 10 second duty cycle limit on my starter it typically takes two to three attempts to start.

Does anyone also respect the starter relay duty cycle? (20 mins between starts)

Doesn't seem very reasonable.

 

[00Dan]

So you "moved the magnetos back 5 degrees". How? by judging the angle of the magneto? Do you have a buzz box?

What is your cold start procedure... details!

ex:
throttle xx.xx inch open
mixture ... set where?
and crank?

Do you cycle the throttle to get the accelerator pump to squirt?
While cranking?
Before cranking?

I retimed the magnetos by loosening the retention nuts, attaching a buzz box, and adjusting the angle of each slightly to achieve 20* BTDC. Of course, this was based on #1 compression stroke, and the impulse coupling released and the prop backed up from there.

My typical start procedure is to run the boost pump until fuel pressure stabilizes, mixture full rich, start cranking, pump the throttle to full two-three times then back to the idle stop and advance slowly to approximately 1/4-1/2". At 10s I'll let off the starter and let it rest 20s. The engine may sputter as if one cylinder is lighting off in this time, but it doesn't catch.

The second starter cycle I've tried a couple different techniques, but none gave consistent results. One was to repeat what I did the first cycle. Another was to open the throttle to ~1/4", mixture to idle cutoff, and slowly advance mixture while cranking.

If I make it to a third starter cycle it's mixture rich, throttle 1/4" and crank. It typically will light off at this point.

 

[PCHunt]

Three full throttle "pumps" sounds like a lot to me. Your mixture could be too rich to start.
Have you tried starting with either one or no throttle pumps?

 

[Kyle Boatright]

Any chance there's something in your switch configuration that grounds both mags while cranking? That generally shows up as "Starts as soon as I release the key or switch."

 

[00Dan]

Three full throttle "pumps" sounds like a lot to me. Your mixture could be too rich to start.
Have you tried starting with either one or no throttle pumps?

I’ll give it a shot.

Any chance there's something in your switch configuration that grounds both mags while cranking? That generally shows up as "Starts as soon as I release the key or switch."

My magnetos are connected to ground via two lever switches, and the starter has an independent pushbutton. I considered making a cutout for the non-impulse mag when I installed this (replaced a bad key switch assembly) but opted for the simpler method of a “R mag off when starting” placard. Short of something having come loose this shouldn’t be possible. At CI in September I validated the P-lead grounding paths during the ground runup by shutting the engine down with the magneto switches.

 

[PCHunt]

Quote:

Three full throttle "pumps" sounds like a lot to me. Your mixture could be too rich to start.
Have you tried starting with either one or no throttle pumps?

I’ll give it a shot.

I assume "No Pun Intended" ??

 

[rmarshall234]

I’m thinking along the same lines as Captain Hunt…too rich to fire. Since you have an accelerator pump you could also try this…key off, no boost pump, open and close the throttle twice. Wait about 30-45 seconds for that fuel from the accelerator pump to vaporize, throttle cracked and crank. If it lights off right away you’ve solved the problem. Good Luck.

 

[BobTurner]

My magnetos are connected to ground via two lever switches, ....

Any chance the left and right wires are crossed? e.g., you're actually trying to start on the right mag?

I've always found "throttle cracked", e.g., 1/8" max, seems to work best on carb'd engines. But, I've never encountered a carb'd engine that did not have a primer line going to at least one cylinder. There must be a reason so many have primer lines.

 

[00Dan]

I’m thinking along the same lines as Captain Hunt…too rich to fire. Since you have an accelerator pump you could also try this…key off, no boost pump, open and close the throttle twice. Wait about 30-45 seconds for that fuel from the accelerator pump to vaporize, throttle cracked and crank. If it lights off right away you’ve solved the problem. Good Luck.

I suspect this would just result in that fuel leaking out my airbox before it could evaporate, not to mention the fire risk.

Any chance the left and right wires are crossed? e.g., you're actually trying to start on the right mag?

I've always found "throttle cracked", e.g., 1/8" max, seems to work best on carb'd engines. But, I've never encountered a carb'd engine that did not have a primer line going to at least one cylinder. There must be a reason so many have primer lines.

Easy enough to idiot check myself on this; I triple checked just that when I replaced the key starter but ya know how it is…

 

[Draker]

Does anyone also respect the starter relay duty cycle? (20 mins between starts)

Doesn't seem very reasonable.

Hmm, doesn't this depend entirely on the starter? My starter placard specifies the following:

Starter cranking is limited to six 10 second cranking cycles with 20 second cool down between cranking attempts, then 30 minutes cooling.​

So with mine I can go (10 seconds, 20 second break) maximum six times before I get out and give it a 30 minute rest.

 

[rmarshall234]

Only suggested it as a _troubleshooting procedure_, good luck.

 

[lr172]

Putting a pin in the hole with the flywheel at 25* DOES NOT get you 25* timing, If you then repeated that same process at 20* that still isn't correct. You must always use a buzz box to get to the correct timing. What you described can leave you WAY off the mark, as there is a LOT of slop in that arrangement. It is only there to get you in the neighborhood.

You are priming via the accl pump correct? Are you doing that with the engine cranking? If not, it is not priming anything but the air filter. This must be done WHILE the engine is cranking.

Stater duty cycle varies with the starter, but usually 30-60 is allowed before a cooling period is required.

 

[lr172]

Does anyone also respect the starter relay duty cycle? (20 mins between starts)

Doesn't seem very reasonable.

NOOOO. It is the starter motor with the duty cycle, NOT the relay. They are usually 20-30% duty cycle. Usually they can handle 60 seconds of cranking, but some will say 30, then another 30 and then a 2 minute cooling period and then 30 start followed by 120 cooling thereafter. This varies by starter, but usually the first cooling cycle can be skipped or shortened on the first start attempt (no heat soak).

20 minutes is so far off the map that it can't even be one of those urban legend things like the OT being 100*

 

[lr172]

Quote:

Three full throttle "pumps" sounds like a lot to me. Your mixture could be too rich to start.
Have you tried starting with either one or no throttle pumps?

I’ll give it a shot.

I assume "No Pun Intended" ??

Sorry I hadn't read the whole thread before posting above. Three full shots on an 80* day sounds like way too much. I think I used three shots when it was below freezing when I had a carb. Try one full shot and bring the throttle back to 1/4" and keep cranking.

 

[blaplante]

Also possible that your accelerator pump is not working, or not working very well.

 

[00Dan]

Putting a pin in the hole with the flywheel at 25* DOES NOT get you 25* timing, If you then repeated that same process at 20* that still isn't correct. You must always use a buzz box to get to the correct timing. What you described can leave you WAY off the mark, as there is a LOT of slop in that arrangement. It is only there to get you in the neighborhood.

You are priming via the accl pump correct? Are you doing that with the engine cranking? If not, it is not priming anything but the air filter. This must be done WHILE the engine is cranking.

Stater duty cycle varies with the starter, but usually 30-60 is allowed before a cooling period is required.

Timing was with a buzz box, the original timing was with the pin in the magneto (distributor cap off). Any accelerator pump priming is done with the engine cranking.

Sorry I hadn't read the whole thread before posting above. Three full shots on an 80* day sounds like way too much. I think I used three shots when it was below freezing when I had a carb. Try one full shot and bring the throttle back to 1/4" and keep cranking.

It's entirely possible I've been flooding it. I've tried one shot, and seen the engine sputter as if it was almost there, and given it one or two more to try. Sometimes it works, not reliably of course.

Also possible that your accelerator pump is not working, or not working very well.

When I had the carburetor off earlier this year to open up the main jet it appeared to work, insofar as actuating the throttle with fuel in the bowl caused a squirt of fuel to come out the accelerator nozzle. I didn't measure the amount metered.

 

[dmattmul]

See attached data sheet

NOOOO. It is the starter motor with the duty cycle, NOT the relay. They are usually 20-30% duty cycle. Usually they can handle 60 seconds of cranking, but some will say 30, then another 30 and then a 2 minute cooling period and then 30 start followed by 120 cooling thereafter. This varies by starter, but usually the first cooling cycle can be skipped or shortened on the first start attempt (no heat soak).

20 minutes is so far off the map that it can't even be one of those urban legend things like the OT being 100*

See attached urban legend

 

[jrs14855]

Starting

Same setup except gravity fuel system. In warmer weather it starts on second or third blade with no prime, throttle at idle stop. In colder weather throttle at idle stop, engage starter a couple of very short strokes of the throttle from idle to maybe 1/8 travel. If engine tries to quit repeat the short pumps of the throttle as necessary until the r/m stabilizes.
Your lack of success with flooded start procedure would seem to eliminate fuel. More likely issues with the left mag.
22 degree timing was recommended to me for the 10/1 pistons.

 

[lr172]

See attached urban legend

I have never used a cole hersee intermittent solenoid, only automotive varieties that are not made by them. 10 sec on and 30 minutes off is an insanely outrageous duty cycle and something that would NEVER survive in the auto world. The fact that C/H has this duty cycle does not imply that every auto starter solenoid also has the same the same duty cycle. I have seen people crank away cars for literally minutes over a very short period with absolutely no damage to the solenoid. Just think about the UPS driver. He restarts his truck every 30-60 seconds. Could you imagine him waiting for 30 minutes after two starts? Maybe these are special units for massive diesels that draw 700 amps.

 

[dmattmul]

ES 24021

I have never used a cole hersee intermittent solenoid, only automotive varieties that are not made by them. 10 sec on and 30 minutes off is an insanely outrageous duty cycle and something that would NEVER survive in the auto world. The fact that C/H has this duty cycle does not imply that every auto starter solenoid also has the same the same duty cycle. I have seen people crank away cars for literally minutes over a very short period with absolutely no damage to the solenoid. Just think about the UPS driver. He restarts his truck every 30-60 seconds. Could you imagine him waiting for 30 minutes after two starts? Maybe these are special units for massive diesels that draw 700 amps.

Sold by Vans (Data sheet added in previous post) part number above. C/H part number 24021. Stop cranking.

 

[YellerDaisy]

They all seem to prefer a recipe of some sort. Both my previous carb'd, no primer O-320 and now a carb'd, no primer O-360 start very reliably with the following... which may not work for any other engine on the planet.

Engine hot (immediately after to about 4 hours after shutdown): Mixture to full rich, crack throttle slightly open, start.

Engine cold and temps above ~60F: Mixture to full rich, one pump of throttle, crack throttle slightly open, start.

Engine cold and temps in the 30-60F range: Mixture to full rich, two pumps of throttle, crack throttle slightly open, start. Sometimes three pumps if at the colder end of the range.

Engine cold and temps very cold: I dunno - very rarely do that without pre-heat.

 

[RVDan]

Fuel: The carburetor does not show signs of leaking or flooding, unless I just leave the boost pump on and the engine off, in which case eventually the bowl floods.

My thoughts on this statement: The fact that the fuel pump can flood the carb is a problem. The float and needle valve should be able to control the 2-5 psi fuel pressure without the engine running.

Either the float height is wrong, or the needle valve is not seating correctly. When you start to crank the engine, the mechanical pump will supply pressure, even if the electric pump is off, thereby flooding the engine.

 

[rmarshall234]

My thoughts on this statement: The fact that the fuel pump can flood the carb is a problem. The float and needle valve should be able to control the 2-5 psi fuel pressure without the engine running.

Either the float height is wrong, or the needle valve is not seating correctly. When you start to crank the engine, the mechanical pump will supply pressure, even if the electric pump is off, thereby flooding the engine.

^^^
Good one. The OP may be starting from a flooded condition all the time.

 

[BillL]

If it is starting at the ultra rich condition, an observer or GoPro should show black puffs for a bit after starting. Otherwise I got Nothing.

 

[jabarr]

Totally agree with Dan and Robert. Also mag timing procedure also seems sketchy.
Has it always been hard to start or just recently?

 

[dougweil]

Assume your ignition, ignition, and air systems are normal. I have seen this issue on Piper PA28s in a flight school that taught at and also with my RV-4. Try this technique: throttle closed, boost pump on, advance throttle briefly to full and then close throttle (just up and back in over about a 2 second period), immediately turn boost pump off, throttle stays closed, engage starter. No pumping the throttle or anything like that. I have seem folks crank away with the throttle being pumped or advanced and simply closing the throttle results in starting. Worth a try!!!

 

[lr172]

My thoughts on this statement: The fact that the fuel pump can flood the carb is a problem. The float and needle valve should be able to control the 2-5 psi fuel pressure without the engine running.

Either the float height is wrong, or the needle valve is not seating correctly. When you start to crank the engine, the mechanical pump will supply pressure, even if the electric pump is off, thereby flooding the engine.

If the needle / seat has failed and is allowing the bowl to overfill the engine would CONSTANTLY run VERY rich and VERY poorly; that is NOT a problem limited only to start up. If you are seeing liquid ful on top of the carb after a start attempt, the more likely cause is a failed accelerator pump seal. If this seal is bad, each time you pump the throttle, fuel comes past that seal and flows out of the shaft seal (located on top near bowl vent) on the up stroke instead of going into the throat. This might also explain why you can't get it started with one pump on an 80* day. Try this - at the end of the runway, quickly apply about half throttle from idle. If it stumbles, that points to a problem in the accelerator pump. An engine transitioning quickly to higher power needs a quick shot of extra fuel to get there smoothly. That is what the accel pump does.

If you can sit there with the eng off and boost pump on and observe fuel flowing out of the carb, it needs to come off NOW! This assumes that you have confirmed that the boost pump has not failed and is delivering a pressure reading beyond the max allowed, which will also cause these symptoms. The needle / seat can only hold back about 7 psi before letting go. If the float is failing and getting heavier, this number starts to drop.

 

[lr172]

Assume your ignition, ignition, and air systems are normal. I have seen this issue on Piper PA28s in a flight school that taught at and also with my RV-4. Try this technique: throttle closed, boost pump on, advance throttle briefly to full and then close throttle (just up and back in over about a 2 second period), immediately turn boost pump off, throttle stays closed, engage starter. No pumping the throttle or anything like that. I have seem folks crank away with the throttle being pumped or advanced and simply closing the throttle results in starting. Worth a try!!!

Sorry, but you got bad advice. Pumping the throttle with the engine not spinning just sends a bunch of raw fuel into your air cleaner assy. Does little to prime the engine (though it does help) and creates a serious fire risk if you happen to experience a backfire, which is not that uncommon on our lyc's during start up. If you are using throttle pumping to prime the engine, IT MUST be done with the engine cranking. The advice works on a car because the carb is ON TOP of the engine and gravity brings it to the cyl's. Our engines have it on the bottom. Even in the auto world, they don't pump to prime fuel. The one pump recommended by the auto maker was not to pump fuel in, but to allow the choke to close. It would not go full without the throttle being depressed.

Sadly many instructors were taught this and pass it on to their students. Mine literally yelled at me when I told him that I refused to do it.

 

[rmarshall234]

If the needle / seat has failed and is allowing the bowl to overfill the engine would CONSTANTLY run VERY rich and VERY poorly;

I had a float stick open on the start to one of my formula one races. My crew tried lifting and dropping the tail several times to free it but it still ran like **** at idle. At race power it ran fine - presumably because I was drawing fuel from the bowl at a quick enough rate. When the race was over and I returned to idle it ran like **** again.

 

[00Dan]

If the needle / seat has failed and is allowing the bowl to overfill the engine would CONSTANTLY run VERY rich and VERY poorly; that is NOT a problem limited only to start up. If you are seeing liquid ful on top of the carb after a start attempt, the more likely cause is a failed accelerator pump seal. If this seal is bad, each time you pump the throttle, fuel comes past that seal and flows out of the shaft seal (located on top near bowl vent) on the up stroke instead of going into the throat. This might also explain why you can't get it started with one pump on an 80* day. Try this - at the end of the runway, quickly apply about half throttle from idle. If it stumbles, that points to a problem in the accelerator pump. An engine transitioning quickly to higher power needs a quick shot of extra fuel to get there smoothly. That is what the accel pump does.

If you can sit there with the eng off and boost pump on and observe fuel flowing out of the carb, it needs to come off NOW! This assumes that you have confirmed that the boost pump has not failed and is delivering a pressure reading beyond the max allowed, which will also cause these symptoms. The needle / seat can only hold back about 7 psi before letting go. If the float is failing and getting heavier, this number starts to drop.

I replaced the needle, seat, and float this spring when I had the carb off to ream the main jet. When I first removed the carb from the engine I did not drain the bowl, and I actuated the throttle shaft by hand - I observed the fuel from the accelerator pump emitting from the accelerator nozzle, and did not observe any leakage.

I don't think I've tested leaving the boost pump on since I replaced the needle and seat, previously the overflow condition would occur if I left the pump on for several minutes. I was told this was normal. For reference, my electric boost pump will stabilize at approximately 4.5 psi on a mechanical gage with the engine off.

 

[lr172]

I previously the overflow condition would occur if I left the pump on for several minutes. I was told this was normal. For reference, my electric boost pump will stabilize at approximately 4.5 psi on a mechanical gage with the engine off.

a proper needle/seat and float will hold back 4.5 PSI for months without overflowing (referring to modern needles with a rubberized tip); That is it's job and if not doing it, will create all sorts of mixture problems. A fuel level change of just 1/64" in the bowl will have a significant affect on mixture. What you are seeing is NOT normal. If you get overflow out of the carb after a few minutes with the boost pump on, something is wrong. As the previous poster mentions, not all needle leaks are equal. Some only bad at lower flow rates and some all the time. All a balance between leak rate, fuel flow rate and head pressure. The last two generally being inversely related with mechanical pumps.

 

[00Dan]

Regarding the needle, when I replaced it I found the old rubber tip to be quite worn. I tested the new one today and it held the boost pump pressure.

I also tried some of the techniques in this thread and found that full rich, throttle just barely cracked, and no priming pump worked well to start the motor on the first cycle. I did find the engine sensitive to how far open the throttle was, any more than just barely off the closed stop and it didn’t like it. This supports the theory that I was inadvertently flooding the engine on my start attempts.

 

[msires]

How many hours on the mags? Have you checked the resistance of the spark plugs?