Back-pressure Measurement Techniques?

[Wicsteve]

Many small engine manufacturers use exhaust backpressure specifications for emission conformance guidelines. Such specifications are often given to a muffler manufacturer for use in designing an exhaust system.

There does not, however, appear to be any consistancy between engine manufacturers (or suppliers of mufflers) as to the procedure used for measuring back pressure. We've seen measurements made in the first chamber of the muffler, in manifolds at locations immedately downstream of the exhaust ports, in the bends of exhaust manifolds, with different probes, in mid stream, at the tube walls, etc. We've seen drastic differences in measured backpressure depending upon small changes in location of the measurement. In twin cylinder engines there can be large differences in back pressure readings between cylinders (with similar manifolds).

This presents the question for general discussion. Can the forum share some good techniques for taking back pressure measurements? Are there accepted practices and advice for making pressure measurements in manifolds? Since measurement technique appears to affect the results so dramatically, can there be any validity in back pressure specifications for assuring that an engine conforms to emission standards?

 

[dgallup]

When I worked for a major diesel engine manufacture in their engine lab, we used two types of probes. The best probe was inserted into the middle of the exhaust stream with a right angle bend that faced directly into the flow. This kind reads stagnation pressure directly. The second type was just a round hole flush with the exhaust wall. With this type, you have to know the flow rate, temperature and cross sectional area so you can calculate the stagnation pressure. It is generally less accurate but is easier to install and and less disruptive to the flow.

In any case, for best accuracy you want to be in a straight section at least a couple of diameters long. Since these were big engines in test cells, it was pretty easy to get repeatable measurements of backpressure. I expect it may be a lot tougher if you are trying to measure a chainsaw mufflers backpressure.

 

[Wicsteve]

dgallup, your information is very relevent to the discussion.

As clarification, most engines under 10 hp (including chainsaw engines (about 1 hp) will have mufflers that bolt directly to the external engine port (no manifold).Back-pressure measurements are made in the first cavity of the muffler and measurement location is generally loosely defined. There has been little discussion whether the measurement should be made near the cavity walls, in the middle of the cavity, or what influence the flow has on the measurement.

Engines over 12 hp generally have mufflers with pipe manifolds. Such manifolds are typically 1.125 inch diameter tubing, rougly 1 foot long with one or two bends in them. The first bend almost always starts wihin an inch of the engine exhaust port. When the measurements are made in the manifold, the effect of flow is continually a subject of discussion as is the effect of tube bends, and method for measuring pressure.

In the past, small engine manufactures often allowed the end user (for example a lawnmower manufacturer) to retrofit the engine with their own muffler/ manifold design. With tighter emission controls, engine companies needed to tighten their requirements. One general specification is to limit maximum backpressure at WOT conditions. Unfortunately, there is extremely poor correlation between pressure measurments made by various parties.

 

[MikeHalloran]

I've measured backpressure in boats on a fair sample of medium and high speed Diesels from multiple manufacturers.

I haven't run across a requirement for measuring or calculating stagnation pressure. Engine manufacturers seem to universally assume that the backpressure will be measured either at the ports they provide (few do), or at the exit plane of the turbo exhaust flange(s). Some document or imply temporary insertion of a custom thin measurement flange, which they don't specify, supply, or sell.

Marine exhaust system manufacturers traditionally provide a radial pyrometer port in their turbo flange, and it's common practice to measure and report backpressure there, at the bore wall, without correction.

In a typical yacht, there is no room for a straight pipe of any length anywhere near the engine, and often there isn't room for a turbo flange that's thick enough to accept a pyrometer probe, so you will find systems with ports downstream, usually somewhere in a small radius elbow. Since the largest single pressure drop is usually in the transition from the turbo flange bore to the first elbow, and the flow distribution within an elbow is a mess, such systems always catalyze contention about backpressure measurements, unless they are deliberately oversized.

Engines with multiple exhaust outlets typically give different backpressure readings at each outlet. Wildly different for staged turbos that are shut off.

One big contributor to uncertainty is the dynamics of the measurement. In a boat, the only backpressure measurement that anyone really pays attention to is at WOT. Absent a dead calm sea, and sometimes even then, the boat is pounding up and down, so the gauges and the technicians reading them are being tossed around. If the boat's exhaust is underwater, the backpressure also changes as the boat rolls laterally, and that happens with every steering correction. All of that means that the gage needles move around, a lot. I have seen people record max/min values and interpolate, report just the max or just the min, or try to estimate where the needle spends most of its time. I personally have found it helpful to insert a valve that can be throttled to damp the gage movement. That gives a stable, if slow responding, value. I have also seen people use Magnehelics with all of the unused ports plugged, which gives a wonderfully low, stable, erroneous reading.



Mike Halloran
Pembroke Pines, FL, USA

 

[Warpspeed]

dgallup has the right idea. A pitot static tube is the correct engineering way to measure static pressure in a high velocity airstream, but on really small engines it can still sometimes be rather a challenge.

What you need is some really small diameter copper capillary tube, (hint, this is used in the refrigerant line of domestic refrigerators and airconditioners). Seal one end with a nice spherical blob of silver solder, and drill four pin holes around the circumference of the tube about ten diameters back from the spherical nose.

Just beyond the four holes, bend the tube at right-angles. Polish up the outside of the tube with fine emery or steel wool so there are no burrs or sharp edges.

This pitot static probe can then be inserted through a small drilled hole in the exhaust pipe and positioned to point directly into the flow along the pipe centerline. It will give wonderfully stable and repeatable readings of static pressure. The hole in the exhaust pipe can be easily welded up afterwards, or sealed with a small screw.

This method also works well when measuring boost pressure drop in turbocharger pipework.

 

[MikeHalloran]

It probably works like a charm on the cold side of the turbo. On the hot side, you might want to position and stabilize the probe with something other than your fingers, because hot gas will be flowing out of the port around it.





Mike Halloran
Pembroke Pines, FL, USA

 

[Warpspeed]

Yes indeed, even more so in a very hot exhaust pipe!

But a small probe closely fitted through a small access hole does not require much mechanical clearance, and the total leakage will be minute compared to total flow.

It will require the probed to be clamped somehow in position to aim it correctly along the pipe centerline. Bare fingers are definitely not recommended for this.

Commercial pitot probes almost always have a pointer and a depth scale attached to the outside pipe, because the probe once inserted into a metal duct cannot obviously be seen.

 

[NickE]

Just to continue this discussion....

If I used a static pitiot tube to measure exaust back pressure in an automotive application... And my data logger likes 0-5v or 0-10v signals... What sensor would you reccommend?

Or alternately what max min etc pressures am I looking for?

(I've already studied how to deal with the heat, I just dont know what max pressure I need to be able to measure, and other such parameters)

 

[Warpspeed]

Be a bit careful if using electronic data acquisition with something like this. The problem is that the exhaust pressure will be pulsing, especially very close to the engine. Pressure transducers invariably have a fairly good frequency response, and the electrical output will also be pulsing along with the pressure.

The data acquisition will instantaneously and repetitively sample the incoming electrical waveform. You will most likely end up recording a series of instantaneous pressure points scattered all over the place from one sample to the next.

Some sort of averaging or low pass filtering will most likely be needed. That could be pneumatic, electrical, or perhaps in the data logger software.

A heavily damped mechanical pressure gauge is a lot easier to use, but the needle will probably still flutter. A mercury filled manometer will be both highly accurate and highly damped and be even better.

 

[turbocohen]

The best source I use for damn near any sensor is Kavlico.

http://www.kavlico.com/pages/industry_apps/automotive/powertrain2.htm#egrflowrate

You will need to determine if the mean or real time pressure is what you need and place the sensor accordingly. Better yet, ask a sales engineer at Kavlico.

 

[tbuelna]

warpspeed is correct. Pressure in an exhaust system is a tricky thing to measure because it is a very dynamic environment.

If your limited on budget, you can get a qualitative measurement by connecting the exhaust pipe with a long length of tube (to eliminate heat transfer) to a mechanical pressure gauge, with a small diameter orifice in line to dampen the pressure fluxuations. This set-up won't give you quantitative results, but it will give you qualitative results from one test condition to the next.

If you have lots of budget, you can get a water cooled pressure transducer with a very high sampling rate and a data acquisition system to record it. That system will give you more data than you ever hoped for, or can use.

 

[NickE]

Warpspeed, tbuelna - Yep, small tubing to kill the heat, adn a damper were being designed in. (As well as data averaging.) Also this is a qualitative project. I just need to see variation btw test setups and possibly over rpm/time curves to see chnges at flow rate/temperature.

Turbocohen- I was going to use an omega sensor that outputs directly to 0-5V or 0-10V... I'm not concerned with production capability. (And I'm only buying one.) Thanks though.

 

[turbocohen]

On the dyno I have used a section of 1/4 stainless tubing connected to a welded on swagelock connector with a thin orifice disk inside the connector held in place by the pipe. You can also add a needle valve to dampen out remaining pulses if an avg pressure is all you need. The section of pipe that is inside the exhaust stream can be welded shut and have a few 1/16" holes cross drilled inline with the flow stream.

T

 

[NickE]

I was going to use 1/16" 316 tubing, cause I have compression fittings in that size already for the thermocouples.

I'm still looking for max pressure on the sensor... I've heard rumor that 12" H2O is a good nominal... I just dont know, and it seems that you folks have done this before...

What max pressure and what range should I look for in a sensor?

 

[Warpspeed]

Peak pressures could conceivably be a lot higher than that. Cylinder pressures at the instant of exhaust valve opening may be in the region of 100psi. An explosive pressure wave will empty into the exhaust port and quickly dissipate along the exhaust pipe volume. If a hypothetical probe was located right at the exhaust vale seat it would see some fairly high peak pressures.

Personally I would say a pressure transducer in the 10psi to 30psi range would be appropriate, depending on the type of measurements planned.

Typical road vehicles at full power might see average pressures in the 2-10 psi along the whole exhaust length.

 

[Wicsteve]

Exhaust manifold back pressure as measured on small engines (under 25 hp) would typically be under 40 inches H2O dependent upon the exhaust system.

 

[NickE]

This is a road going vehicle and a 2.0L FI @14psi. The pressure measurement location would be after turbocharger, after catalytic converter, after resonator, before backbox.

So around 30-40psi max with a resolution ~2psi?

 

[Rob45]

One big point in measuring exhaust backpressure:
unless you reference your measurements to an exhaust gas temperature, you'll have no repeatability whatsoever.
The exhaust system must be up to normal operating temperature - but not more; this is one reason there is wide disagreement between exhaust backpressure as measured on an engine dyno, and backpressure measured on the road, where is significant cooling of the exhaust system.
Exhaust backpressure is dependent on the gas temperature, so you need to simultaneously measure temps.

A good little transducer for this is made by PCB, their model 1501B01, that is "rated" 10 psig but good to ~20 psig, with a sensitivity around .5V/psi. Filtering the output gives very good results.

 

[carnage1]

before you buty the omron I would suggest hooking up something cheeper. it will get you ballpark readings to make sure you don't have to purchase more than one.
motorola
mpx4250 4.7v output 36 psi scale 55psi overpressure <20$

 

[NickE]

Ive already speced what seems like a decent transducer from Omega...

http://www.omega.com/ppt/pptsc.asp?ref=PX209_PX219&Nav=preb03

Cost is an issue, but reliability and ruggedness are important to.