I'm a dummy, so explain how compression testing works...

[Jeff S]

Why do you need to crank a compression test many strokes?

Let's say I have a 10:1 compression engine... I replace the spark plug with a pressure gauge and turn thru 1 intake and compression stroke. The pressure in there ought to be 150 PSI-ish, right? 10x ambient.

But that's not what happens, it reads 30 PSI, then next cycle 50, then continues to build until it ends at something that MIGHT be 150, or might be more or less. What's happening here? The exhaust port opens each cycle, so actual pressure inside drops to ambient... so why doesn't the gauge oscillate between 0 and 10x ambient (on an engine with perfect sealing valves, rings, etc)?

OK - so the gauge doesn't show "the pressure in the cylinder NOW"... it must show "max pressure it witnessed".... but still, the same thing is happening in the cylinder every 4 strokes... so what "builds up" over many strokes? What's different about the nth stroke compared to the 1st?

 

[Jeff S]

OK, one thing that is different each subsequent cycle is the pressure WITHIN the gauge. I assumed it was hydraulic from the "plug part" to the "dial part" (measure at the plug)... but that seems to be pneumatic... as evidenced by the PSSSSS when you reset the dial. So, that means the total compression ratio isn't 10:1, it is lower with the added volume within the gauge.

Engine Compression Ratio = (Vol(combustion chamber) + Vol(cylinder) ) / Vol(combustion chamber)
Engine + Gauge Ratio = (Vol(combustion chamber) + Vol(cylinder) + Vol(gauge) ) / (Vol(combustion chamber) + Vol(gauge) )

So, the hysteresis is the pressure that "builds" in the gauge each cycle? If that internal pressure was, say 140PSI on the first crank, the only a TINY volume of air would enter the gauge - so the effective Vol(gauge) would be VERY small - so we're very close to measuring actual engine compression pressure... Each stroke then makes the effective Vol(gauge) closer and closer to zero - as the "volume" that's important isn't the total internal volume, but rather the volume of air that ENTERS the gauge on one compression stroke (which only happens when the pressure outside is higher than the pressure inside).

Maybe? But still, that internal volume seems insufficient to make one compression stroke 50PSI instead of 150.

Assume a 1000cc displacement single at 10:1 ratio. That means the combustion chamber is 1000/9 in volume, or 111.11cc.

(111.11 + 1000) / 111.11 = 10 or 1111.11/111.11 = 10

If we think Engine + Gauge(PSI=0) Ratio = 5:1 (since one crank shows 50PSI instead of 150) then, Vol(gauge) ~= 138.88cc:

(111.11 + 138.88 + 1000) / (111.11 + 138.88) = 1250 / 250 = 5

138cc total internal volume of gauge REALLY doesn't seen possible. So, I did some math but still feel stumped...

 

[julimike54]

Oil pressure builds as it's cranked thus providing better sealing of rings, that's my guess...

 

[FE_Rex]

A 15-3/4” x 1/4” hose (roughly 40 x 0.6 cm) results in unswept volume addition of ~11cc. Your 1L 10:1 motor is now ~ 9-1/4:1. The assumed volume addition does not include the Bourdon tube volume ( assuming an analog gauge).

The static compression ratio does not take into account valve timing events that control cylinder pressure (dynamic compression ratio)

In my car experience, most dead volumes were 90cc or less with swept volumes of ~625 to 1170cc. IIRC, I don’t think a good cylinder ever hit less than 75-80 psi on the first stroke. 180 psi seemed to be the limit for safe pump gas use.

 

[dukey33]

I’m not an expert engine builder but I think you hit on it; the sealing is imperfect.
Rings have an end gap and clearance to the piston lands. When doing compression checks, typically the engine is not at normal operating temp so there are greater clearances all around.
When oil is added to the cylinder, sealing improves and the numbers are higher.

Also, you don’t get 100% fill in the cylinder on every intake stroke. The air enters the cylinder thru a tortured path so the cylinder doesn’t reach atmospheric pressure at BDC.

 

[one Wolf]

 

[Jeff S]

Oil pressure builds as it's cranked thus providing better sealing of rings, that's my guess...

Ahh, indeed. So it might "stroke" thru a 1L displacement, but leaks out a meaningful portion.

most dead volumes were 90cc or less with swept volumes of ~625 to 1170cc

What is "dead volume" ? What I call the combustion chamber above (air volume at TDC)?

Also, you don’t get 100% fill in the cylinder on every intake stroke. The air enters the cylinder thru a tortured path so the cylinder doesn’t reach atmospheric pressure at BDC.

Sure, at 10k RPM. But cranked by the starter with the throttle open, I gotta think the intake duration (in seconds, not degrees) is enough to reach full ambient pressure in there. No?


New terminology for me!
Static compression ratio = my math above = (Vol(combustion chamber) + Vol(cylinder) ) / Vol(combustion chamber)
Dynamic compression ratio = take into account valve timing (instead of BDC you care about where in the stroke the intake valve fully close.

Sound right?

In published engine stats, which are they reporting?

 

[FE_Rex]

What is "dead volume" ? What I call the combustion chamber above (air volume at TDC)?

- Combustion chamber
- Head gasket (take into account diameter and compressed height)
- Piston to deck clearance*
- Piston dish/dome*
* may add or subtract volume

 

[one Wolf]

Compression check is only an indicator of engine health, not an absolute measurement. The specs are almost the same for most every engine. Typical procedure is crank a warm motor until the gauge indication stops increasing, < 70psi not very good > 130 psi very good. One cycle of compression is not the way to check and doesn't tell anything. It's not different for a full race motor vs. a lawnmower.

 

[aggie81]

The reason to do a compression check on a cylinder is to find out how fast air is escaping, and from where.
A differential compression tester is the tool for the job.

 

[FE_Rex]

You can use a compression tester for a couple reasons. Most use is limited to multi-cylinder engines help determining the non-contributing cylinder after checking fuel and spark systems

There is an exception for single cylinder 2 cycle small engines - engine runs like donkey and won't idle - needs at least 80 psi to reasonably idle. Its hard to judge compression with a pull cord on these small motors

To run the test the engine should be warm, with a good battery, all plug removed, throttle blocked open and ignition disabled. You screw the compression tester into the spark plug hole and turn the motor over a few times until the gauge stops climbing. All the cylinders should be within ~5-10% of the same reading. If you have a tester that uses a quick connector the the gauge, you can do other quick tests on the bad cylinder by reconfiguring the tester. Remove the shrader valve from the tester, block the motor at TDC with valves closed on the cylinder being tested, remove the radiator and oil fill cap. Hook that quick connector to compressed air - you'll hear minor air escaping past rings into crankcase, but should hear nothing out the exhaust or intake - puking coolant is a bad sign as well as is major leakage into the crankcase. You can do the same pressurized test with the piston at BDC with the valves closed to look for a cylinder pinhole or split.

Most of my experience has been with HI-PO motors and usually these tests only tell you is what bank of cylinder you will be working on and if you will be buying valves, pistons, or got lucky and only need a head gasket. If the motor was over 12:1 and you felt you neeeded a compresion test, you KNEW the motor was gonna come out and apart.

 

[mr72]

Most of my experience has been with HI-PO motors and usually these tests only tell you is what bank of cylinder you will be working on and if you will be buying valves, pistons, or got lucky and only need a head gasket. If the motor was over 12:1 and you felt you neeeded a compresion test, you KNEW the motor was gonna come out and apart.

That's interesting. This entire thread is fascinating to me as I've often had the very same confusion as Jeff. If the compression ratio is 10:1 then the max the compression tester should ever read is 147 psi at sea level, assuming zero volume in the hose. My 11:1 Suzuki should max out at 161 psi in this case but it's well over 180 since the top end rebuild. I'm pretty good with math and this never made sense to me.

But I am curious about the 12:1 question because one mod I'm considering for my Bonneville is a 904cc kit which raises the CR to >12:1. Along with different cams, this mod is said to increase power by 25-30% or so. My instincts tell me that it will increase the stress on the crank bearings and rods by a factor of about 1.8 ( (12/10)^3 ). And in addition it would probably increase the heat substantially, which may not be a great thing for an air-cooled motorcycle operated in Texas. So I want to know more about this 12:1 comment by @FE_Rex but I hate to hijack Jeff's thread. Maybe I'll ping @FE_Rex in a PM.

 

[FE_Rex]

If you’re racing a 12:1, or power adder motor and it shows symptoms needing a compression checker to diagnose, you likely had a detonation problem - blown head gasket (lucky), lifted upper ring land / holed piston (again lucky), fubared rod/split block (not so lucky).

Modern engine control systems, variable valve timing, and chamber design have come a long way from the MSD / Holley-carbed dynosaur big block and FE fords I used to play with

Hence the name FE_Rex

 

[Jeff S]

So I want to know more about this 12:1 comment by @FE_Rex but I hate to hijack Jeff's thread.

Hijack away! I'lll follow along to try to learn about that also.

Oh, and my Triumph is 12.9:1 , so I kinda just think 12:1 is a decent threshold between high compression and REALLY high...

the max the compression tester should ever read is 147 psi at sea level

Perhaps really good exhaust gas scavenging and port tuning etc could yield higher than ATM at BDC?

 

[FE_Rex]

If you’re talking modern sport bikes, my information is like discussing what vacuum tubes to use in your Apple Watch

 

[humanrace]

The formula for max cylinder compression is P2 = P1 × (V1 / V₂), with P-2 actually being PSIG instead of PSI. Most pressure gauges read 0 at atmospheric pressure when they should read 14.7 at sea level. Hence the G added for PSI Gage. There are lots of variables when using a compression tester, one being valve overlap. For instance, most engine designers leave the intake valve open after bottom dead center, while the piston is already rising, to take advantage of the velocity of the air that is still entering the cylinder. At slower engine speeds, this advantage decreases or even becomes a disadvantage. A good test of the condition of a cylinder should use a compression test AND a leak down test. I use a compression test to determine overall cylinder condition, using the value I got on a freshly built and broken in engine as a reference point. The compression tester will give a lower reading if the center of the cylinder is worn, but a leak down might look good. I use a leak down test to determine what is bad at top dead center. The leak down is performed at TDC. Compressed air, usually at 100PSIG is fed into the cylinder through a very small orifice. The tester has two gauges, one for intake pressure and one past the small orifice to test the pressure in the cylinder. If everything is perfect, both gauges read 100 PSIG. They never will. If the cylinder gauge reads 95 while the intake side reads 100 then there is 5% leakage in the cylinder. One can usually tell where the leakage is happening by opening the oil cap and seeing if a piece of tissue will flutter when place over the opening. Same with the intake or exhaust. Each will separately find intake valve, exhaust valve or ring leakage. The combination of the readings of both instruments will determine whether you need $0, $XX or $XXXXXXX. Having a good compression reading is not always an indication of a healthy engine. One of the examples above had an engine with a reading too high for the compression ratio for that engine. My experience shows that there are two likely reasons. Bad gauge and carbon build up. I had a Snap-On leak down tester in my shop and I could plug the output and the gauge would show 3% leakage.

 

[mr72]

OK so anecdotal report here...

My GS500 has reputed compression ratio of 9.0:1. When it wouldn't run and it needed an immediate top end rebuild it measured 120psi in one cylinder and 130psi in the other. On the bench with all valves closed the mechanic poured gas into the exhaust port and it streamed straight out of the bent exhaust valves. They had bent due to carbon buildup on one edge of the valve and would not seal. After the top end rebuild one cylinder reads 160psi and the other 170psi.

But the math. 9.0*14.7 = 132.9 ... so how does this work. I asked Gemini so you don't have to, but about my Bonneville with compression of 9.2:1 and a spec of 170-180psi for "excellent" compression. Here's what it said:

That is a fantastic observation. Your math is logically sound for a static, slow-moving world, but engines deal with a physics phenomenon called adiabatic compression.
Here is the simple breakdown of why the gauge shows a higher number:

1. Air Gets Hot When Squeezed
Imagine you are rubbing your hands together very fast; they get hot, right? When your engine piston flies upward, it squishes the air so fast that the air molecules get "excited" and heat up instantly.

• The Physics: Hot air expands and pushes outward much harder than cold air.
• The Result: This extra heat energy creates extra pressure. If you squeezed the air very, very slowly (allowing the heat to soak into the metal walls), you would see a number closer to your 135 psi. But because the piston moves fast, the heat stays in the air and "boosts" the pressure reading.

It also suggested a few other things but overall the biggest factor is heating.

It said thusly:

That also makes even more sense, and makes that 170-180 make sense and also is why the pressure drops as you leave the gauge connected and watch it... the air cools.

FYI I'm also talking to Gemini about the UTP 904cc kit with 12:1 pistons and ol' Gem says this is kinda dumb for a TX motorcyclist who values durability and reliability.

 

[tejasbusa]

A good test of the condition of a cylinder should use a compression test AND a leak down test. I use a compression test to determine overall cylinder condition, using the value I got on a freshly built and broken in engine as a reference point. The compression tester will give a lower reading if the center of the cylinder is worn, but a leak down might look good. I use a leak down test to determine what is bad at top dead center. The leak down is performed at TDC. Compressed air, usually at 100PSIG is fed into the cylinder through a very small orifice. The tester has two gauges, one for intake pressure and one past the small orifice to test the pressure in the cylinder. If everything is perfect, both gauges read 100 PSIG. They never will. If the cylinder gauge reads 95 while the intake side reads 100 then there is 5% leakage in the cylinder. One can usually tell where the leakage is happening by opening the oil cap and seeing if a piece of tissue will flutter when place over the opening. Same with the intake or exhaust. Each will separately find intake valve, exhaust valve or ring leakage. The combination of the readings of both instruments will determine whether you need $0, $XX or $XXXXXXX. Having a good compression reading is not always an indication of a healthy engine. One of the examples above had an engine with a reading too high for the compression ratio for that engine. My experience shows that there are two likely reasons. Bad gauge and carbon build up. I had a Snap-On leak down tester in my shop and I could plug the output and the gauge would show 3% leakage.

During my career with one of the major Japanese motor companies, I found that the above statement is very true, with one small change. Using 100 PSI as your input pressure would force low tension piston rings out against an oval or other out of spec cylinder and actually give a leakdown number that was acceptable. I found that lowering the pressure to 50 PSI gave a much more accurate reading than 100 PSI.

 

[Cagiva 549]

Havent used a compression tester in 30 years , anytime I need to diagnose a weak motor I use the non contact thermometer , start the motor cold and shoot exhaust manifold temps at the same spot on all cylinders and look for a cold cylinder . I seldom work on motors with spark plugs these days .

If the motor has a valve problem you will here the popping in the exhaust or intake , if a piston or rings have failed crankcase blowby will be evident in either case a overhaul is warranted , I dont patch up a multi cylinder motor , if I have to open up a motor it will go back together with a compleat rebuild .

 

[Ben Eberle]

OK so anecdotal report here...

My GS500 has reputed compression ratio of 9.0:1. When it wouldn't run and it needed an immediate top end rebuild it measured 120psi in one cylinder and 130psi in the other. On the bench with all valves closed the mechanic poured gas into the exhaust port and it streamed straight out of the bent exhaust valves. They had bent due to carbon buildup on one edge of the valve and would not seal. After the top end rebuild one cylinder reads 160psi and the other 170psi.

But the math. 9.0*14.7 = 132.9 ... so how does this work. I asked Gemini so you don't have to, but about my Bonneville with compression of 9.2:1 and a spec of 170-180psi for "excellent" compression. Here's what it said:



It also suggested a few other things but overall the biggest factor is heating.

It said thusly:

That also makes even more sense, and makes that 170-180 make sense and also is why the pressure drops as you leave the gauge connected and watch it... the air cools.

FYI I'm also talking to Gemini about the UTP 904cc kit with 12:1 pistons and ol' Gem says this is kinda dumb for a TX motorcyclist who values durability and reliability.

So, a 9.2:1 compression ratio takes the pressure from one atmosphere to 23.35 atmospheres. The decrease in volume accounts for 9.2 atmospheres so the rest of the increase is due to the air becoming hot. Since the pressure is 2.43 x 9.2 atmospheres, it's obvious that the temperature is 2.43 times hotter. If the air started at 100 degrees F, add 459.67 to 100 to convert F to Rankine temperature. T1 is 559.67 Rankine so T2 should be 559.67 x 2.43 or 1359.7 R. Subtract 459.67 from 1359.7 R to convert back to F and you get 900 degrees F.

A lot of pilots think their turbojets have a 50:1 "compression ratio". Not really, they have a 50:1 pressure ratio which is a 16.35:1 compression ratio. Air enters the compressor at minus 40 F and exits the compressor at 823.7 F. (theoretically, the fact that the compressor is not 100% efficient means there is also heat added due to airflow friction)

 

[Irishcoffee]

If it fires and starts, call it good and go ride.