Talk:Leak-down tester

Compression testing needs its own wikipedia page
Why does "Compression test" redirect to this page, when there is no proper discussion of compression testing here? I recommend removing the redirection and making a page to discuss automotive engine compression testing. It is surprising to to see this ubiquitous automotive test does not have its own page. --User:Aigisthos

I have no idea why its like that. I would have thought there would be an article on compression testing by itself. Leak down testing is closely related to compression testing but not close enough to warrent the redirection. I wouldnt know how to go about fixing it.--=Motorhead (talk) 23:43, 15 April 2009 (UTC)

verifying assertions??
I did an experiment to verify the assertion that two identical restrictions in series will produce a 50% leakdown. I used two new motorcycle jets, Mikuni #90, used two matched gauges from an FAA-certified leakdown tester, and checked all my joints for leaks with soap. It was not even close to 50%--I got 23% leakdown. Has anyone else done anything similar?Freddybeamer (talk) 20:40, 5 February 2009 (UTC)

good to experiment
Certainly merits further thought. Does 23% represent a drop from 100? or is it 23 above zero? I dont know what type of gage you have. Are you saying you got 23% across each restriction in series at the same time? The flow into the 2 in series identical restrictions is equal therefore the flow across each is equal, the pressure drop across each must also be equal. If the total input pressure is 100 and the drop across the first is 50 then what remains is 50 as input to the second restriction which exits to the atmosphere at 0. Lets say the pressure dropped 23% across the first restriction in the series. That would leave 77% in the system as input to the second restriction. Therefore the two restrictions cannot be equal, as the pressure drop across #1 is 23 and across #2 it is 77, which MUST be flowing the exact same amount of air as #1. Am I missing something? Id like to hear more about your setup such as where gages were connected and pressures applied to the system. This is ignoring the possible differences that might arise if one was flowing subsonic and the other was choked flow--=Motorhead (talk) 15:26, 8 February 2009 (UTC)

The apparatus consists of 1/4 inch NPT brass fittings, two female tees and two double-ended male connectors. On one end of each male connector I soldered a #10-24 brass nut and then retapped the thread with a M5X0.8 tap to accept the jets. The entire unit was then assembled as follows:  mini-regulator, tee with first pressure gauge, connector with jet, tee with second gauge, connector with jet. The threads of the jets were sealed with pipe dope, the pipe threads with teflon tape. The jets were oriented the same way-so that the air approaches the slot-end of the jet. With the unit pressurized, and my thumb over the end, I could confirm that, at least, the pressure gauges were reading more or less the same pressure, to within 2 or 3 psi over the range of 50 to 80 psi. Whether the absolute pressure readings are accurate or not, I have no way of measuring. The pressure readings were 80 psi, 61 psi. I then swopped the jets and got 80, 60. Then I swopped the gauges and got 80, 58. Then I turned the second jet around, and got 80, 65. Freddybeamer (talk) 01:53, 14 February 2009 (UTC)

ok so...
If I read it correctly you have the same schematic as I put in the article with the exception that instead of an engine there is a second jet attached to the output end to simulate an engine.

Regulator->gage1->jet1->gage2->jet2>out to atmosphere.

That seems to be standard arrangement for a typical leakdown tester except that instead of the typical .040" restriction you have a jet.

Plugging the end gives you the same reading on both gages and indicates the pressure the regulator is set at. So far that all seems totally what Id expect to see.

Then opening the system to the atmosphere you reset the regulator to give 80 psi on gage1 and you see 60 psi on gage2. That gives a pressure difference of 20psi across jet1 and 60psi difference across jet2. Yet both jet1 and jet2 are flowing exactly the same mass of air per min but jet 2 requires 3X the pressure difference that jet1 does to flow exactly the same mass of air.

Is the ratio of 60/80 what you are referring to as 23% in your opening paragraph? 60/80=75/100 or 25 "%" leakage is one way of looking at it.

Jet1 flows air at 80+15(1atm)=95psi abs. Jet2 flows air at 60+15=75psi abs. Absolute pressure ratio of 75/95 = 78.9/100 or 21.1 "%" leakdown ABS.

If thats all correct then seems there is an effect of density difference in the regions of the 2 jets. Jet1 exists in a 95psi abs area and jet2 in a 75psi abs area. The density of the air for jet1 is 21% higher. Im guessing that the extra density allows jet1 to flow more mass for less pressure difference.

The 2 jets flow exactly the same mass of air but since jet1 is flowing denser air it needs less velocity than jet2 which flows the same mass of air but that air is less dense so the velocity has to be much higher. Since the velocity is primarily driven by pressure difference the pressure difference must be higher for jet2. That must be aproximately what explains the results you got.

Thanks for pointing that out. Im going to remove my illustrative comment. Although it doesnt change the fact that leakdown "%" dosent mean anything and does not relate to anything meaningfull but I made a boo boo in assuming that the flows and pressures would be the same. Id be interested to know what is the sonic speed at the two different densities to see if that has an effect also. It takes aprx. 15psi to achieve sonic speed so I think both jets are in that region but jet1 is closer to subsonic than jet2. Could be an effect there also. If I get time Ill fool around with it. For my part I dont use this style tester. I use an actual flowmeter (rotameter) at a standard pressure in place of a restriction which I can then relate to engine size rather than to an arbitrary restriction plug. Thanks--=Motorhead (talk) 23:01, 14 February 2009 (UTC)

In defense of leakdown testing
There is no denying that the leakdown tester has its problems, but it is simple and convenient--it’s the interpretation of the results that is lacking. In all the books and articles I have read on engine building, I have yet to see any more than vague hand-waving about the results of a leakdown test.

With that in mind, allow me to present, without proof, a distilled version of the analysis of pneumatic flow through two restrictions is series, done by B.W. Andersen in his book “The Analysis and Design of Pneumatic Systems”. The purpose of this is to hopefully make leakdown testing more useful and less mysterious. After all, the aim of leakdown testing is to try and measure the size of some tiny holes buried deep in the engine.

The analysis goes like this
 * Pressure P1 is the pressure measured upstream of the orifice--the constant test pressure whether it’s 100 psi or 80 psi,
 * Area A1 is the area of the leakdown tester orifice,
 * Pressure P2 is the pressure downstream of the orifice and is also the pressure in the combustion chamber,
 * Area A2 is the aggregate area of all the leaks in the combustion chamber,


 * Q is the flow (mass flow or volume flow) of air through the system--orifice and combustion chamber, at the constant test pressure P1,
 * Qmax is the maximum flow through the orifice with the test pressure at P1 and the pressure P2 low enough that the flow through the orifice is “critical” or at the speed of sound.

The analysis boils down to two equations:

$$ \frac{Q}{Q_{max}}= \sqrt{ (14.929) \left [ \left( \frac{P_2}{P_1}\right)^{1.42857} - \left( \frac{P_2}{P_1}\right)^{1.71428} \right]  }$$

$$ \frac{A_2}{A_1}= \sqrt{ (14.929) \left \{  \left [ \left( \frac{P_1}{P_2}\right)^{0.2857} -   \frac{1}{2}  \right]^2 - \frac{1}{4} \right \}  }$$

I've crunched some numbers

So, how well does this model fit reality? The first and second columns can be verified by attaching a really small valve to any leakdown tester and a flowmeter. Any test pressure above about 30 psi gauge can be used but please note that all pressures used in the equations and the table are absolute pressures (gauge pressure + 14.7 psi). Also note that the analysis breaks down for pressure ratios below 0.525 because the flow regime changes to sonic flow through the orifice. Column three, the area ratio A2/A1, is harder to verify because the areas in the equations are actually so-called “effective areas”. However, notice that I set up some of the rows so that the area ratios are double other rows. If A1 is substantially constant, then doubling A2 should give the corresponding pressure ratio, flow ratio and leakdown. This is done by using two really small valves in parallel. Separately set one valve, then the other, to read the same leakdown, then test the valves together. I’d be interested to know if anyone else can reproduce these numbers.Freddybeamer (talk) 02:33, 8 March 2009 (UTC)

Great stuff
Shouldn’t there be a flow coefficient applied to the area formula though? Or am I thinking too sub sonically? I wonder also what would be the flow coefficient and flow regime of very tiny leaks such as through a good but not perfect valve seat. The scale effect should certainly affect the flow regime at such tiny openings. Do shocks and choked flow exist in passages only few atoms wide? Another thing I’ve wondered is the variation of gas viscosity between a leakdown test and a running engine. We test at low temperature but the leak actually occurs at very high temperature and very much higher viscosity. It could be that a leak at testing temperature is very much reduced at engine running temperature due to that viscosity increase. It would be useful to carefully correlate leakdown tests with blowby measurements on the running engine. That would shed light on all this. I wish I had thought of it when I had access to all that equipment.--=Motorhead (talk) 18:20, 8 March 2009 (UTC)

Relationship To Commercial Leakdown And Home Built Testers Should Be Clarified
Typical commercial leakdown testers sold in auto parts stores have two gauges but the downstream one is calibrated in percentage instead of pressure. It would be useful for the article to include a description how this type of tester works. The initial adjustment of such a tester requires adjusting the upstream pressure till the downstream gauge reads zero percent, so I suspect that all that is involved is the mathematics of ratios. Nevertheless, it would be useful to have this explained.

Many built-it-yourself guides on the internet show the construction of a leakdown tester with only a downstream pressure gauge. The upstream pressure is set by using the gauge on the air supply. It is convenient to have both gauges and the regulator handy, but is there any gain in precision from having these devices close together?

Tashiro (talk) 16:39, 5 October 2011 (UTC)