What's new
Van's Air Force

The definitive Van's Aircraft support community! Buying, building or flying an RV? Join our exclusive family of mentors and enthusiasts!

Interesting Observations - Dynamic Compression Testing

L

lr172

Legacy Member
Doing our CI a couple days ago, I got 50/80 on one cylinder. Had some leakage on both valves, so lapped them to resolve. Ended up with 60/80, with all that going past the rings. The other 3 were 72-73. I have always been skeptical of using leak down testing, as it has several flaws built into it and in the auto world, dynamic compression is the standard, with leak down testing used only to find leak sources. While deciding what to do about that cylinder, I decided to do a dynamic compression test, as in the end, all that matters is the amount of pressure produced in the cylinder while the engine is spinning. I tested with four compression strokes per cylinder. Two cylinders came in at 135 and two at 138. The 60/80 cylinder was 138.

I have no intention of pulling the jug for a hone / re-ring, as the performance of that cylinder is well matched with its mates. I thought I would share this experience with others, as 60/80 seems to be the area where most folks would start considering taking steps to address.

If anyone knows. I am curious what a fresh 8.5:1 lyc cylinder produces if anyone has done this. I really wish this was published data.

My engine has 1500 hours.
 
Might be useful:

From https://www.omnicalculator.com/everyday-life/compression-ratio-to-psi ...

Pressure = X / Y × Atmospheric pressure
where X and Y are the compression ratio values (e.g., 9:1 compression ratio to psi, where X = 9 and Y = 1), and Atmospheric pressure is typically 14.7 psi.
For example, with a compression ratio of 10:1, the calculated pressure would be:
Pressure = 10 / 1 × 14.7 = 147 psi

For 8.7:1 the number comes out to 128

Multiple factors:
Density Altitude of test
Gauge Accuracy
Valve timing (if the intake valve isn't closed fully at BDC the psi is reduced), mostly a factor on engine with a 'race' cam.
Formula assumes ZERO ring leakage which isn't real world

I make no guarantee of the accuracy of the formula the guy used.
 
Interesting observation.
If I get a questionable leak down test on a cylinder that did not have any previous issues I don’t get too hung up about it. I’ll move on and test later after running the engine. Almost every time the “issue”’ goes away and doesn’t return at next CI.
If the valves show signs of a problem under the scope, that’s a different story.
Never thought about a dynamic test, but never needed to I guess. I’ll keep this one in the bag.
 
Continental has long taken the position that valve sealing matters, ring seal not so much. Their "test orifice" approach allows a very low leakdown value.
 
Yes, it takes 5-6 "hits" to max out each measurement, starter driven. Plugs out except for the tested cylinder.

It's a precursor to a battery top off and capacity check.
 
DanH said:
Continental has long taken the position that valve sealing matters, ring seal not so much. Their "test orifice" approach allows a very low leakdown value.
Click to expand...
That is logical. Didn't mean to imply leak downs aren't important. They are invaluable in finding leak sources.
 
moosepileit said:
I use the cheapest HF kit. 9:1, 170 to 180 psi. Uncalibrated, but repeatable.
Click to expand...
Yeah, kind of figured that 140 was low. Most cars I work on are in the 160 range. The real question is what should an IO-320-B1A be. The cam makes a big difference and have no idea how similar they are across models.
 
Last edited:
Does anyone happen to know the IVC (intake valve closing) angle on these engines? Also, is it different across different engines models?
 
Last edited:
blaplante said:
Might be useful:

From https://www.omnicalculator.com/everyday-life/compression-ratio-to-psi ...

Pressure = X / Y × Atmospheric pressure
where X and Y are the compression ratio values (e.g., 9:1 compression ratio to psi, where X = 9 and Y = 1), and Atmospheric pressure is typically 14.7 psi.
For example, with a compression ratio of 10:1, the calculated pressure would be:
Pressure = 10 / 1 × 14.7 = 147 psi

For 8.7:1 the number comes out to 128

Multiple factors:
Density Altitude of test
Gauge Accuracy
Valve timing (if the intake valve isn't closed fully at BDC the psi is reduced), mostly a factor on engine with a 'race' cam.
Formula assumes ZERO ring leakage which isn't real world

I make no guarantee of the accuracy of the formula the guy used.
Click to expand...
ICV is typically at least 30-40* after BDC. Given the low RPM range on the Lyc's, I am guessing it is in the 50* range. Late closing tends to help at low rpms, but detrimental at high rpm's, IIRC.
 
Even though I have been using differential leak-down compression testing myself, I really like your dynamic compression testing approach! I use it on all of my restored 70's Honda CB-Fours. I would expect 8.5:1 compression to read in the neighborhood of 140 to 150 psi. Also, using "old school" techniques of simply comparing the compression to the condition of your spark plugs is something I do, especially on the bottom plugs. I assume, you were testing your compression with a warm engine?
 
RV-9erA said:
Even though I have been using differential leak-down compression testing myself, I really like your dynamic compression testing approach! I use it on all of my restored 70's Honda CB-Fours. I would expect 8.5:1 compression to read in the neighborhood of 140 to 150 psi. Also, using "old school" techniques of simply comparing the compression to the condition of your spark plugs is something I do, especially on the bottom plugs. I assume, you were testing your compression with a warm engine?
Click to expand...
I do all my leak down tests cold. I prefer getting worst case numbers. But i do them the same way all the time for comparative purposes. The dynamic numbers I posted above were also cold. It is just a habit from the auto world where they are typically done cold.

I also read plugs. If you know how to read them, it tells you a lot about how your combustion process is working. I used to set timing on small block chevys by observing the heat signature on the neg electrode. It is a great way to optimize timing without a dyno.
 
lr172 said:
I do all my leak down tests cold. I prefer getting worst case numbers. But i do them the same way all the time for comparative purposes. The dynamic numbers I posted above were also cold. It is just a habit from the auto world where they are typically done cold.

I also read plugs. If you know how to read them, it tells you a lot about how your combustion process is working. I used to set timing on small block chevys by observing the heat signature on the neg electrode. It is a great way to optimize timing without a dyno.
Click to expand...
Back when I was doing owner assisted annual inspections for many years with IA's on certified airplanes, they would almost always want to run the engine before doing a compression test. I've seen numbers go from 60/80 up to 76/80 right after a short engine run. But, I do like you're idea of performing a dynamic compression test--I guess people just don't like that swinging prop out there even with the starter turning the engine over.
 
RV-9erA said:
Back when I was doing owner assisted annual inspections for many years with IA's on certified airplanes, they would almost always want to run the engine before doing a compression test. I've seen numbers go from 60/80 up to 76/80 right after a short engine run. But, I do like you're idea of performing a dynamic compression test--I guess people just don't like that swinging prop out there even with the starter turning the engine over.
Click to expand...
Yes, most folks in aviation are VERY set in their ways. Just ask someone to take off with OT at 90* and let the fun begin. Then watch them start their camry and immediately drive off with the OT at 30*. I guess if it doesn’t have an OT gauge, then it is no problem . I have reasons for doing it the way i do and yes, I occasionally test them hot when the situation calls for it diagnostically. I rarely care when aviation mechanics tell me i am doing things wrong.
 
Intrigued.
Can someone provide a link or description on how to do a dynamic comparison check? Would be particularly interested in do’s and Dont’s.
Also interested in answer to Larry’s question in original post on if anybody knows what psi range in a dynamic compression check would be acceptable. Always like new data but only if I know what it means. Maybe a good question to send to Savvy and Mike Busch.
 
plehrke said:
Intrigued.
Can someone provide a link or description on how to do a dynamic comparison check? Would be particularly interested in do’s and Dont’s.
Also interested in answer to Larry’s question in original post on if anybody knows what psi range in a dynamic compression check would be acceptable. Always like new data but only if I know what it means. Maybe a good question to send to Savvy and Mike Busch.
Click to expand...
A quick google search will help. It is the standard compression test used in automotive, so lots of tips out there. May have to make adapters though. Many of the compression testers only have adapters for 12 / 14mm holes. But if you are using auto plugs, you already have 14 - 18 adapters.

The tester will have a button, like some tire pressure gauges that holds pressure. You will want at least 4 compression strokes in a row when doing this. Again, pretty standard.
 
Using the formula posted earlier, which I've seen before in these discussions,

8.5:1 = 125 psi
9:1 = 132 psi

If you're getting 140+ psi doing a dynamic check, where are you finding your "good" number to compare to?

140 psi = 9.5:1
170 psi = 11.5:1
 
Snowflake said:
Using the formula posted earlier, which I've seen before in these discussions,

8.5:1 = 125 psi
9:1 = 132 psi

If you're getting 140+ psi doing a dynamic check, where are you finding your "good" number to compare to?

140 psi = 9.5:1
170 psi = 11.5:1
Click to expand...
I don't have good numbers to compare to, which is why I asked.

Expexted pressures are dependant upon cam timing. Because the intake valve closes after BDC, you cannot calculate this with compression ratio alone. You have to know the closing point and do some math with the effective swept volume.

It is a whole bunch easier to have a range gathered from known good engines of similar setup. This data is very hard to find because this type of testing is rarely done in aviation.

Many times, we are looking for individual cyl problems, so we are just looking for one that is not like the others. While 140 is a bit low compared to what I am used to, it is not surprising, as the lyc is a low RPM engine and speculate the IVC is much later than traditional engines because of this and explains the low test results.
 
Thought I would add some info. Today I spoke with a local A&P that I know and trust. He indicated that he uses dynamic compression testing ocassionally. In his experience, he sees 120-150 PSI on lycomings, depending upon how worn they are. Standard CRs. Unsure how reliable this info is, but thought I would share it.
 
Snowflake said:
Using the formula posted earlier, which I've seen before in these discussions,

8.5:1 = 125 psi
9:1 = 132 psi

If you're getting 140+ psi doing a dynamic check, where are you finding your "good" number to compare to?

140 psi = 9.5:1
170 psi = 11.5:1
Click to expand...
The formula numbers above seem low to me when comparing them to the many other air-cooled engines I've tested over the years with similar compression ratios. For example, I would expect a 9:1 compression ratio to read 150 psi and 8.5:1 to read 140 psi, but not much lower.
 
RV-9erA said:
The formula numbers above seem low to me when comparing them to the many other air-cooled engines I've tested over the years with similar compression ratios. For example, I would expect a 9:1 compression ratio to read 150 psi and 8.5:1 to read 140 psi, but not much lower.
Click to expand...
Agreed. Those seem low based upon my experience as well. While not a physicist or mathmetician, that formula seems far to simple.
 
Snowflake said:
Using the formula posted earlier, which I've seen before in these discussions,

8.5:1 = 125 psi
9:1 = 132 psi

If you're getting 140+ psi doing a dynamic check, where are you finding your "good" number to compare to?

140 psi = 9.5:1
170 psi = 11.5:1
Click to expand...
With my 390 with 10.76:1 compression, and after 6 cranking strokes, I was getting over 220 psi. dynamic cylinder pressure and the comressed air in the cylinder was 150 degrees F.
 
Yes, temp will change the pressure change as well, and if we want to get really into it, non- ideal gas assumptions, heat loss, etc.. A rough estimate is the temp change would maybe increase the pressure by about 30%, so about 160 psi on an 8.7:1. PV=nRT ideal gas law.
 
Good idea, mine are all mid to high 70s but will try this next time.

Before doing my annual Ill go out and do at least a 30 to 40 min run at 75% plus power and then do my leak down hot.

Otherwise I find that without the hard run that occasionally the odd cylinder will be lower than expected and Ill have to re do it later.
 
lipper03 said:
Yes, temp will change the pressure change as well, and if we want to get really into it, non- ideal gas assumptions, heat loss, etc.. A rough estimate is the temp change would maybe increase the pressure by about 30%, so about 160 psi on an 8.7:1. PV=nRT ideal gas law.
Click to expand...
Wow. Never would’ve guessed the temperature impact would be that large. Really wished i had taken more science in college.
 
Heat of compression is a real thing. Positive displacement blowers routinely have a temperature rise approaching 200F when running with a pressure rinse of 10 psi of more. That's peanuts compared to the pressure rise in an engine. Look at diesels. Hot enough to ignite. . . granted with an even higher compression ratio.

For this instance, heat transfer to the cylinder will decrease temperature and pressure a bit. Us engineers would calculate peak theoretical pressure and temperature assuming no heat transfer and no gas leakage. My books are at the office, so it'll have to wait or someone else could ask AI to calculate adiabatic temperature (and pressure) after compression.
 
Yes, this test is not a combustion event either, only air. ~30% is based on the 150F measured from post #25. One could try to get much closer with more specific data. I was just trying to show that the ranges are in the ballpark of previous posts but need to include some temperature rise.
 
Full Throttle said:
With my 390 with 10.76:1 compression, and after 6 cranking strokes, I was getting over 220 psi. dynamic cylinder pressure and the comressed air in the cylinder was 150 degrees F.
Click to expand...
How are you measuring temperature in the cylinder while doing this test?
 
Okay, just did some more math.

Based on ideal gas law, PV = nRT, you can work out that the pressure reading should be the compression ratio times the temperature ratio times the ambient pressure. Temperatures referenced to absolute zero, so in Kelvin (Celsius plus 273) or Rankine (Fahrenheit plus 460).

Based on that, at STP (0C/32F and 14.7psi/29.92mmhg) the additional temperature contribution would be about 24%. That drops to about 15% at room temp (20C/70F) so the temperature effect is both significant, and critical to the measurement. So to do accurate compression checks with a gauge on the spark plug hole you'll also need a peak recording thermometer in the cylinder alongside it.
 
Makes me long for the simple method of “Armstrong compression testing”….pull the blades through and see if there are any weak strokes. Back when I had an ancient J-3, that was good enough…as long as we had three out of four cylinders with compression, we went flying around the patch…..🤣
 
Yeah, for starting at 77F and 1 atmosphere, without heat transfer Google says compression at 9:1 yields about 318 psi and over 830F. Absolute numbers don't mean much in cases like this. As Larry said, comparative data is much more valuable. Note that end values depend on starting conditions, so you shouldn't expect identical results across the seasons, even if your engine is unchanged.
 
Snowflake said:
Okay, just did some more math.

Based on ideal gas law, PV = nRT, you can work out that the pressure reading should be the compression ratio times the temperature ratio times the ambient pressure. Temperatures referenced to absolute zero, so in Kelvin (Celsius plus 273) or Rankine (Fahrenheit plus 460).

Based on that, at STP (0C/32F and 14.7psi/29.92mmhg) the additional temperature contribution would be about 24%. That drops to about 15% at room temp (20C/70F) so the temperature effect is both significant, and critical to the measurement. So to do accurate compression checks with a gauge on the spark plug hole you'll also need a peak recording thermometer in the cylinder alongside it.
Click to expand...
As I mentioned before, the compression ratio is not relevant. CR is a comparative number based, in part, upon swept volume - the entire stroke from BDC to TDC. It also requires you to know the chamber volume at TDC. In reality, the cylinder is not compressing anything untill the intake valve is mostly closed. This is typically 30-40* ABDC, but often greater on low revving engines. The point of valve closing is based upon cam timing and lobe profile and is highly variable across different engines. There are several dynamic CR calculators out there that will provide an effective CR if you input the valve closing angle. However Lyc doesn't publish this and have been unable to find it.

For example, I built a big block chevy back in the day with 12:1 CR and ran it on pump gas. The key was it had a VERY agressive cam. The effective CR, after factoring in IVC, was down in the area of 9:1 or less.
 
Last edited:
Snowflake said:
How are you measuring temperature in the cylinder while doing this test?
Click to expand...
I put a temp sensor in # 4 exhaust pipe and cranked the engine. That's as hot as it got after 6 strokes. I have 4 individual exhaust pipes. The exhaust pipe is about 24" long and 1.67 ID. Ambient outside temp. was 71 F.
 
GROK does all the calculating for me. It's just interesting to see the difference between the technically correct mathematical answer and the real world empirical data. I use the information I gather for reference and trend analysis. And, ya have to ask the right question to get the rightr answer from GROK.
 
Full Throttle said:
GROK does all the calculating for me. It's just interesting to see the difference between the technically correct mathematical answer and the real world empirical data. I use the information I gather for reference and trend analysis. And, ya have to ask the right question to get the rightr answer from GROK.
Click to expand...
To be fair, nobody has shown a technically correct mathematical answer yet. I've just put together the math I know, there are clearly other factors so my math is incomplete. There is a mathematical solution here, we just don't have it yet.
 
One variable that has not been mentioned is cranking speed. There is a huge variance here that will have a direct effect on the readings you get.
Your “math” needs to assume a perfect no leak scenario or an assumption off of that.
A leak down test provides a level of control and consistency that a dynamic test can’t provide. It also allows us to identify paths of leaks, its primary purpose.
So, as the title of the post says “Interesting Observation” and I agree. But as a tool, I’m not seeing any real value here. There are many ways to investigate why a leak down test failed or is marginal and if corrective action is necessary.

Interesting discussion but I’m not digging out my old auto gage anytime soon.
 
At around the 25 hour mark I did both leak down and automotive compressions.

The results COLD were:

Leak down: 78/80 on all 4
Automotive style: 1: 170 2: 160 3: 157 4: 150

So there was some variation which didn’t surprise me as when I pulled the prop through when cold, one cylinder felt a bit softer. I’d assume that’s #4 from the results above. I think they feel about the same now. The engine now has 65 hours and seems fully bedded in - less than 0.5qt burn in the last 25 hours now that I’ve switched to multigrade and runs very smoothly.

This is 8.5:1 O320 with Lycoming cylinders.

At the next 50hr interval I’ll do the same and update my results.
 
JonJay said:
One variable that has not been mentioned is cranking speed. There is a huge variance here that will have a direct effect on the readings you get.
Click to expand...
Please explain how rpm affects combustion pressure and why it makes a huge difference.

No one was asking you to change your testing method. I mentioned an alternative and the reasons I choose to use it. We are all welcome to follow our own path.
 
IO390 said:
At around the 25 hour mark I did both leak down and automotive compressions.

The results COLD were:

Leak down: 78/80 on all 4
Automotive style: 1: 170 2: 160 3: 157 4: 150

So there was some variation which didn’t surprise me as when I pulled the prop through when cold, one cylinder felt a bit softer. I’d assume that’s #4 from the results above. I think they feel about the same now. The engine now has 65 hours and seems fully bedded in - less than 0.5qt burn in the last 25 hours now that I’ve switched to multigrade and runs very smoothly.

This is 8.5:1 O320 with Lycoming cylinders.

At the next 50hr interval I’ll do the same and update my results.
Click to expand...
Thanks for sharing your details!
 
I should add that I just found that the oring on my tester was cracked, so my numbers are likely off from actual, though the relative comparison should be solid. Plan to re run the test.
 
Snowflake said:
To be fair, nobody has shown a technically correct mathematical answer yet. I've just put together the math I know, there are clearly other factors so my math is incomplete. There is a mathematical solution here, we just don't have it yet.
Click to expand...
The math is in the google result: https://www.google.com/search?q=what+is+the+temperature+and+pressure+after+compressing+air+nine+to+one&sca_esv=219388647f983b16&sxsrf=AE3TifNXSKDkPV3kqyQkRDTvAtThUHAlug:1759176025025&source=hp&ei=WOXaaM-RPI-3wN4PgO3w2QI&iflsig=AOw8s4IAAAAAaNrzaZozpT1tdJRtC9ZNSewhkZL75iMU&oq=what+is+&gs_lp=Egdnd3Mtd2l6Igh3aGF0IGlzICoCCAAyBBAjGCcyBBAjGCcyChAjGIAEGCcYigUyDhAAGIAEGLEDGIMBGIoFMg4QABiABBixAxiDARiKBTIIEAAYgAQYsQMyDhAAGIAEGLEDGIMBGIoFMgUQABiABDIOEAAYgAQYsQMYgwEYigUyDhAAGIAEGLEDGIMBGIoFSJYZUABY1AZwAHgAkAEBmAGmAaAB5QaqAQM0LjS4AQHIAQD4AQGYAgegAuYFwgILEAAYgAQYsQMYgwHCAgsQLhiABBixAxiDAcICDhAuGIAEGLEDGNEDGMcBwgIOEC4YgAQYsQMYgwEYigXCAgUQLhiABJgDAJIHAzQuM6AH3V6yBwM0LjO4B-YFwgcFMC42LjHIBw4&sclient=gws-wiz

I just told it 9:1 compression, and it assumed standard temperature and pressure for the values I posted. As Larry explained, true compression ratio is lower than theoretical based on valve timing, not to mention the leaks that we're trying to identify. Temperature and time also play into the results, as that heat of compression immediately begins to transfer to the surroundings, not to mention that air movement by valves also varies with engine speed. Heck, even humidity will impact the results a bit.

In the end, conduct your testing with as close to identical conditions as you're able, and trend your results over the long haul to establish norms and deviations.
 
lr172 said:
Please explain how rpm affects combustion pressure and why it makes a huge difference.

No one was asking you to change your testing method. I mentioned an alternative and the reasons I choose to use it. We are all welcome to follow our own path.
Click to expand...
Sorry, over stated.
Not my path. It’s what they teach and what every mechanic I have ever seen uses,…. for decades, but you know that.
If others find value in a dynamic test, great.
As I stated, interesting conversation…
 
A leak down test is the method specified in AC 43.13 and is especially helpful in determining where the pressure loss is occurring. Pulling a cylinder if it isn't above 60/80 as Lycoming long recommended (Service Instruction No. 1191A) may be unnecessary, expensive, and potentially cause bearing problems if the torque isn't done correctly on re-install. Continental published a service bulletin (SB03-3) addressing low results in compression tests. They have data showing 100% of rated horsepower produced by test engines with leak down readings in the low 40s. The main thrust of their SB is to identify the source of the leakage via sound and verify valve condition via borescope. If it is valves, something needs to be done, but moderate ring leakage doesn't keep an engine from making power, though the oil consumption is likely to be high. A dynamic or pump up test comes closer to showing if a cylinder is unusable because of ring leakage. Compression, combustion, and power will happen fast enough that leakage you would see in a differential pressure test may be pretty much moot. If it will pump up to some reasonable value, and if it's similar to the opposing cylinder, operating the engine is probably safe.

Ed
 
Bicyclops said:
A leak down test is the method specified in AC 43.13 and is especially helpful in determining where the pressure loss is occurring. Pulling a cylinder if it isn't above 60/80 as Lycoming long recommended (Service Instruction No. 1191A) may be unnecessary, expensive, and potentially cause bearing problems if the torque isn't done correctly on re-install. Continental published a service bulletin (SB03-3) addressing low results in compression tests. They have data showing 100% of rated horsepower produced by test engines with leak down readings in the low 40s. The main thrust of their SB is to identify the source of the leakage via sound and verify valve condition via borescope. If it is valves, something needs to be done, but moderate ring leakage doesn't keep an engine from making power, though the oil consumption is likely to be high. A dynamic or pump up test comes closer to showing if a cylinder is unusable because of ring leakage. Compression, combustion, and power will happen fast enough that leakage you would see in a differential pressure test may be pretty much moot. If it will pump up to some reasonable value, and if it's similar to the opposing cylinder, operating the engine is probably safe.

Ed
Click to expand...
Good info. Thanks for posting that.
 
Bicyclops said:
A leak down test is the method specified in AC 43.13 and is especially helpful in determining where the pressure loss is occurring. Pulling a cylinder if it isn't above 60/80 as Lycoming long recommended (Service Instruction No. 1191A) may be unnecessary, expensive, and potentially cause bearing problems if the torque isn't done correctly on re-install. Continental published a service bulletin (SB03-3) addressing low results in compression tests. They have data showing 100% of rated horsepower produced by test engines with leak down readings in the low 40s. The main thrust of their SB is to identify the source of the leakage via sound and verify valve condition via borescope. If it is valves, something needs to be done, but moderate ring leakage doesn't keep an engine from making power, though the oil consumption is likely to be high. A dynamic or pump up test comes closer to showing if a cylinder is unusable because of ring leakage. Compression, combustion, and power will happen fast enough that leakage you would see in a differential pressure test may be pretty much moot. If it will pump up to some reasonable value, and if it's similar to the opposing cylinder, operating the engine is probably safe.
Click to expand...
Now that makes some sense. Nice write up.
 
You must log in or register to reply here.
Back
Top