Testing fuels and blends by connecting to gen set and load

Single Tank WVO systems and blending SVO WVO to thin it.

Moderators: SunWizard, coachgeo

Testing fuels and blends by connecting to gen set and load

Postby David » Mon Dec 08, 2008 7:04 am

As I mentioned in another post, the other day I was mucking around with an industrial diesel generator testing it on various fuels. I was actually just wanting to see how this engine would perform on SVO as I would like to use it for a grid feed setup which would require long runs.

I was using a multimeter measuring the voltage of a couple of batteries I was charging in the process and noted a difference in the output on different fuels. With the idea that this may give some indication as to different power outputs and possibly economy on various fuels, today I did some more runs on the engine to verify my initial findings in this regard. .

Before I go any further, let me just say I am only mentioning my observations for interest value and do not propose them to be definitive scientific testing in any way. They are backyard tests and should be treated that way so those that feel inclined to point out the no doubt many flaws in what I was doing, please save your bandwith. No one is trying to say what I am doing proves a thing, it is posted for " entertainment" value only!

The engine I have is a Hatz 1B20 which is a direct injection, 297cc, 3600rpm direct injection, fixed timing, air cooled diesel. It is directly coupled to a Bosch 80A 12V alternator. I attached the output of the alternator which is also coupled to the engines inbuilt alternator to a 90AH Automotive battery.
To this I attached a 240V, 300W inverter which was running a 300W incandescent light bulb. I don't have a tacho to measure RPM but I guess I was running the engine around 2000 RPM only. I have noted at this load, running the engine at full speed increases output less than 1/10th of a volt so in the interests of less wear and tear on both the motor and my hearing, I chose to tun it well under max output.

I monitored the voltage only because I do not have a clamp meter capable of measuring the amperage. Given the changes that measuring 1/100th of a volt can detect, I think this is sufficient as an indicator for my backyard experiments.

Today I fired up the engine and ran it for about 15 min on bio. I let the alternator charge the battery and observed that it stabilized at 14.18V.
When changing fuels and adding or removing the test load, I saw that that the voltage stabilized at this point every time which was obviously the inbuilt regulators cut off point so it was a good base to know when the battery was fully charged.
I was unsure if the battery may mask some of the changes in the alternator output but connecting even a tail lamp of less that 1amp showed a more significant voltage drop than any of the fuel tests so I am happy that the battery may stabilise but not actually influence any changes in power or output.

With the load applied, The voltage leveled out at 14.10V on B100 which was an interesting and novel brew my friend made from a load of Duck fat I came across. Still smells like regular bio from the exhaust.
Next I tried a mix of 20% RUG and WVO. Don't know what the WVO was specifically, it was from one of my 1000L upflow tanks so undoubtedly had a good proportion of cottonseed, canola and a few other things in it. Probably a good " representative" WVO mix.

As fully expected, the power was down on this blend and the voltage dropped to 14.03V, the lowest voltage I saw. The engine also ran noticeably rougher and unevenly. I was somewhat surprised the voltage was not lower. At first I thought I had got air in the fuel line but after making sure it was properly purged and re running all the fuels, the results I got surprised me in being identical.

Next I tried some SVO, the same base stock I mixed with the RUG.
Power was up to 14.08 on the unheated oil. I did see a noticeable exhaust haze on the SVO. I wouldn't really call it smoke, you had to look at it in the right light to see it, but it was certainly there.

The final fuel I tried on this run was some straight Avcat turbine fuel.
While the engine ran smoothly on this fuel, the output voltage was low at only 14.04V. there was no exhaust haze at all but the exhaust did smell somewhat Rich like there was an amount of unburnt fuel in the exhaust. Probably just am impression rather than reality but I have no way to test emissions other than observation.

I ran all the fuels at least 3 times and was completely surprised by the identical numbers I got each time even when testing the fuels in random order. Between some tests I removed the load to let the battery stabilize to full charge which was almost instant and re-applied the load. I came up with the same numbers each and every time.

I plan to run some more tests with different blends once I can put aside enough drink bottles to mix them all up in and going to a servo to Buy
:shock: :shock: :shock: some diesel :shock: :shock: :shock: to test straight and with some blends.
I'm interested to see if this test reflects my beliefs that a 5% Rug mix will have more power over straight SVO. This is of course a very different engine to the one in my car so if it doesn't show any improvement my opinion will not be changed as the the effect of it in my vehicle but if it does show more power, that will re-enforce my seat of the pants impressions. I think any increase will on this engine show up as a small one however given what I have seen from the other fuels I have run.
I would also expect straight dino should provide the highest numbers given it has more energy content than any of the other fuels and is fact what the engine is designed and optimized for.

I'm happy that this method is accurate enough to show changes in fuel variances and while certainly limited, if nothing else may be useful in adding weight to, or supporting theories and ideas gained from other sources.
_____________________

I don't give a damn about what might or could happen until a significant group of people can tell me it HAS happened to them.
Until then, it's just more endless gloom and doom Veg folk law.
David
 
Posts: 333
Joined: Thu Mar 06, 2008 4:12 am
Location: Sydney Australia

Postby bio_cowboy » Mon Dec 08, 2008 10:04 am

Hey David, I think using a generator for fuel-mix testing is a good idea. It allows one to create a steady load, from which one could monitor a number of things, like rate of fuel consumption, power output, etc.

The large incandescent light bulb makes and excellent constant load to monitor. So, I would monitor the voltage drop across that light bulb, which is just a high watt resister. Doing so will give you the current being delivered using Ohm's Law via the formula I=V/R. You could also calculate power or watts by P=IV. If you could add an RPM gauge to your generator it might give you another point of reference on the system. Also, measuring the resistance of your incandescent light bulb will help in the accuracy of your Ohm's Law calculations.

Rate of fuel consumption could be monitored, as mentioned in another thread, by making some kind of graduated cylinder that you can fill and valve into your fuel line to monitor the consumption of a measured amount of fuel over a fixed period of time.
bio_cowboy
 
Posts: 27
Joined: Tue Nov 25, 2008 6:01 am
Location: Amarillo, TX, USA

Postby coachgeo » Mon Dec 08, 2008 10:33 am

Some viscosity measure would be good for comparision too.

Same test at other loads would be interesting too. Say two or three light bulbs so the motor really has to work at high load like it is designed. Im sure the power band for the 1b20 can be found on the net somewhere.

I have heard of folk who found a way to change timing on these motors even though they are stated as "fixed"

Would be REAL interesting results you would get if you could adjust timing to burn the fuel your using to max efficency by watching the volts or amps go up in result of changing the timing.
Last edited by coachgeo on Mon Dec 08, 2008 10:50 am, edited 3 times in total.
Life; It's all in the Balance

Moderator
coachgeo
 
Posts: 569
Joined: Thu Mar 06, 2008 10:46 am
Location: North Texas

Postby coachgeo » Mon Dec 08, 2008 10:43 am

A twist to this might be hooking the motor to an air compressor and testing the PSI the thing can put out.

Not sure if there would be a way to change the load this way though.
Life; It's all in the Balance

Moderator
coachgeo
 
Posts: 569
Joined: Thu Mar 06, 2008 10:46 am
Location: North Texas

Re: Testing fuels and blends by connecting to gen set and lo

Postby coachgeo » Mon Dec 08, 2008 10:49 am

David wrote:...As fully expected, the power was down on this blend and the voltage dropped to 14.03V, the lowest voltage I saw. The engine also ran noticeably rougher and unevenly. I was somewhat surprised the voltage was not lower. At first I thought I had got air in the fuel line but after making sure it was properly purged and re running all the fuels, the results I got surprised me in being identical. ...
Could the RUG blend been getting hot in the fuel line enough to begin to be turning the gas to vapor. This coulc be an additional plausable explaination on the rougher/uneven running.
Life; It's all in the Balance

Moderator
coachgeo
 
Posts: 569
Joined: Thu Mar 06, 2008 10:46 am
Location: North Texas

Postby zoochy » Mon Dec 08, 2008 11:49 am

To summarize your tests:
14.10 = B100 (du canard)
14.08 = WVO (cottonseed and canola)
14.04 = Avcat turbine fuel
14.03 = 20%RUG 80% WVO (cottonseed and canola)

I’m very interested in these results. Although you claim they are not scientific, they’re probably the most scientific fuel comparisons I’ve seen on this site to date.

20% RUG is a bit high. I suggest trying the tests again with half as much RUG. Also, non-waste canola as a baseline for your WVO.

Thanks for your efforts.
91 Toyota Hiace 3L engine w/ a 300W ½” heated fuel line, lift pump, additional 10um fuel filter, and a 2nd diesel tank for starting.
>160,000 km on:
88% canola WVO
11% old gasoline
.6% turpentine
.3% acetone
.1% eye of newt
zoochy
 
Posts: 66
Joined: Fri Mar 07, 2008 8:21 pm
Location: Victoria, BC, Canada

Postby bio_cowboy » Mon Dec 08, 2008 4:05 pm

I agree with coachgeo, if David had a way of testing the viscosity of his blends it would improve the reliability of his results. However, I doubt a difference of .01 volts is a reliable measurement, because David's DVM most probably only reads to 2 decimal places, so the variation in measurements could easily be instrument drift, or subtle differences in surface of the contacts.

I also forgot to look at the load verses the output capacity of the generator. One would want a constant load of at least 50% of the capacity of the generator for these results to be meaningful.
bio_cowboy
 
Posts: 27
Joined: Tue Nov 25, 2008 6:01 am
Location: Amarillo, TX, USA

Postby David » Mon Dec 08, 2008 4:59 pm

Thanks for the interest and replies.

It would be I think pointless to measure the voltage drop across the Bulb itself given that I am running it from an inverter and this device will aim to supply a constant voltage through it's electronics. There would definitely be some compensation going on with that which would skew the results or at very least hide small variances. Given the fluctuations are less than 1/10th of a volt on the supply side, I doubt the inverters output would change given the tiny percentage this represents at a nominal 240V output.

I don't have a tacho I can use for the tests and the only one I know of, one of those Tiny tachs is more than I want to spend on the exercise and I'm not sure I could find a suitable place to attach the sensor on this engine anyway.

I will try to measure the fuel consumption but varying the RPM by what I imagine the tiny amounts needed could be a problem. I did some other tests on another Industrial diesel I have with some HOHOHO gas and leaving the throttle in a fixed position still showed variations in consumption although the load I was testing which was the gas generator itself was quite large and this was enough to activate the governor. I'm not sure this will happen in this test but one way to find out.

Viscosity I can measure relatively by simply letting the fuels run through the serological pipettes I have. Won't give an actual viscosity rating but would just show the relative differences.

For additional load I could run some 12V lights I have but I might need quite a few. The alternator will max out before the engine gets too loaded. Perhaps the inefficiency of the car alternator will be my friend here in putting much more load on the engine than just the outright electrical current draw.

I fully expect that the RUG concentration in the blend was too high but I don't think the problem with the poor running was necessarily vaporization but it may well have been a factor. This engine has a high return rate of fuel and when swapping from one fuel to the next the air being purged in the return line is very evident. I saw no unusual bubbles in the return when running the RUG blend but any vaporization could have been occurring past the fuel return point or the vapors could have re condensed by the time they exited the return. The return fuel on this engine does seem rather hot so vaporization is a strong possibility.
I think the blend is likely to be causing an upset with the combustion timing and is just a lower energy fuel.

I plan on doing some more tests with different Rug Blends and my own bet for best power is with a 5-10% mix and I may even get inspired and try a 7.5% mix just to see which way the numbers go.

The last run was really just an experiment in seeing if the idea really worked. The observations I made initially were in the process of doing something completely different and I wanted to make sure I wasn't basically seeing things. Having convinced myself that this setup is a valid indicator and is capeable of showing the differences in various fuels, I'll make up some more mixes and run them in as carefully controlled a manner as I can create in order to get some numbers that are as factual as possible.

My multimeter has a logging feature but I have never used it or seen the cable that hooks it up to a computer in a long time. If I can find the interface and software I'll drag out an old computer and hook that up so I can record the runs and note any variations.

One other test I thought of ( and aren't there always an endless number that a person could try!) is to give the engine a spray of water for a while and see if that changes the numbers afterward with the idea of it possibly cleaning any deposits out and allowing for better efficiency. It could also go the other way and if things in the engine are worn ( and I know it has done some decent hours!) perhaps any carbon may also be helping seal the engine and I may see a power drop. Either way a difference would indicate the water has had some effect.

I'll update when I can get all the samples I want to run organized and give it another go.
_____________________

I don't give a damn about what might or could happen until a significant group of people can tell me it HAS happened to them.
Until then, it's just more endless gloom and doom Veg folk law.
David
 
Posts: 333
Joined: Thu Mar 06, 2008 4:12 am
Location: Sydney Australia

Postby David » Mon Dec 08, 2008 5:28 pm

bio_cowboy wrote:I agree with coachgeo, if David had a way of testing the viscosity of his blends it would improve the reliability of his results. However, I doubt a difference of .01 volts is a reliable measurement, because David's DVM most probably only reads to 2 decimal places, so the variation in measurements could easily be instrument drift, or subtle differences in surface of the contacts.

I also forgot to look at the load verses the output capacity of the generator. One would want a constant load of at least 50% of the capacity of the generator for these results to be meaningful.


UGGHHH!

And this is what I eluded to right from the start!!! :twisted:

What I am doing is backyard testing with what I have available. As I tried to acknowledge right from the beginning ( see paragraph of first post in BOLD text!) it is not scientific, reliable or dependable in any way. It's simply observations based on the setup I have at hand which I also acknowledged no doubt has many flaws. By virtue of the fundamental and basic fact I am running these tests on a single cylinder, air cooled, industrial engine when the engines people are running in their vehicles are completely different makes any numbers a guide at beast and anyone that wasn't obsessed with a pre-determined agenda would see and recognize that.

If people want meaningful, creditable tests that are up to scientific standards and scrutiny, go spend YOUR Time, money and effort in setting something up and running numbers that satisfy your standards, or go commission a testing facility that can.
My bet is no matter what anyone or any Lab does for that matter, someone will always come along and say the tests don't mean anything because of this, that or some other reason they dream up.

The thing I notice when ever anyone does something like this is that if the results agree with peoples pet cause and ideals, then there is never a problem. The minute the results show something other than what people want it to, then they start picking the whole thing to pieces. I have yet to see someone challenge results they wanted to see no matter how dodgy the procedure in the experiments may have been.

As far as instrument Drift or contact surface changes go, I think that is a load of complete and utter crap. I have actually had the meter I am using properly calibrated ( admittedly last year) and if it were drifting, it would be higher odds than winning the lottery that when I tested each fuel 3 times or more, the numbers always came back the same as before for each fuel! That to me says what I am seeing is dam accurate because it is repeatable multiple times. I expect when I do another long series of tests I will see some variance through the difference in climatic conditions over the time frame running all the fuels multiple times will take. The results should all be affected equally but still, to allow for these changes beyond my control, I'll average the numbers results out.

I'm doing this out of my own curiosity and going further than I would have other wise through the interest of the people here.
If you don't think my numbers mean anything, FINE! I never proposed them to be reliable fact in the first place!

You are free to either ignore the numbers I come up with or get off your A*%e and go run your own that you will no doubt be happier with when they show exactly what you want them to! :roll:
_____________________

I don't give a damn about what might or could happen until a significant group of people can tell me it HAS happened to them.
Until then, it's just more endless gloom and doom Veg folk law.
David
 
Posts: 333
Joined: Thu Mar 06, 2008 4:12 am
Location: Sydney Australia

Postby SunWizard » Mon Dec 08, 2008 6:08 pm

Interesting results so far, thanks for posting about this. I agree that a bigger load would probably make the variance between blends more apparent. The fuel consumption would also be a good indicator of differences but its more of a pain to measure than voltage.
YVORMV - Your veg. oil results may vary.
95 Dodge Cummins 4x4 SVO WVO conversion.
81 Mercedes 300D- stock and happy on V80/D20 blend.
Low fossil net zero house- 100% solar power and heat.
SunWizard
Site Admin
 
Posts: 1720
Joined: Mon Dec 11, 2006 2:53 pm
Location: N. Colorado

Postby Burbarian » Mon Dec 08, 2008 11:38 pm

David,

Interesting test. Go for it!
To adjust the throttle setting in small controlled increments, use a turnbuckle.
Measure the current going into the inverter. The inverter + bulb is the load.
Crude viscosity testing could be done with a coffee or bean tin with a nail hole through the bottom, a thermometer and a stopwatch.
A cheap tachometer can be rigged using an inexpensive frequency counter, available at Radio Shack or eBay.
1987 GMC Suburban 6.2L V8 IDI
1985 Merc 300TD
1968 CAT D4D 3304 dozer
1971 Waldon 4100 loader
1981 IHI 30F excavator
1995 Changfa 195 w/ ST 10kw genset
Burbarian
 
Posts: 195
Joined: Wed Mar 05, 2008 12:14 pm
Location: Vermont

Postby John Galt » Tue Dec 09, 2008 12:52 am

One might get more consistent results if a light bulb or other pure resistive load is connected directly to the generator output, and then measure the voltage drop across the resistive load. That's similar to how a dynamometer works. The batteries and inverter ' muddy the water' somewhat.
John Galt
 
Posts: 526
Joined: Tue Mar 04, 2008 9:17 pm
Location: Coldest North America

Postby Burbarian » Tue Dec 09, 2008 10:02 am

Unfortunately, a pure resistive load like a lightbulb changes with voltage fluctuation. As the supply voltage goes up, the load goes up. With a stabilized load that draws a constant voltage regardless of supply voltage, the amp draw is a constant. Hence this acts as a static ballast, allowing you to measure fluctuations on the supply side.
1987 GMC Suburban 6.2L V8 IDI
1985 Merc 300TD
1968 CAT D4D 3304 dozer
1971 Waldon 4100 loader
1981 IHI 30F excavator
1995 Changfa 195 w/ ST 10kw genset
Burbarian
 
Posts: 195
Joined: Wed Mar 05, 2008 12:14 pm
Location: Vermont

Postby John Galt » Tue Dec 09, 2008 1:54 pm

I apologize if my comments implied that the voltage regulator should be bypassed. That of course shouldn't happen.
John Galt
 
Posts: 526
Joined: Tue Mar 04, 2008 9:17 pm
Location: Coldest North America

Postby bio_cowboy » Tue Dec 09, 2008 4:47 pm

John Galt wrote:One might get more consistent results if a light bulb or other pure resistive load is connected directly to the generator output, and then measure the voltage drop across the resistive load. That's similar to how a dynamometer works. The batteries and inverter ' muddy the water' somewhat.
I agree with John, and since an incandescent light bulb is such a pure resistive load, then that 300 watt incandescent light bulb should work fine as such a load, and then measuring the voltage across the lightbulb should not be a problem.

Also, David, I meant no offense to you or your equipment when I mentioned meter drift, but I assure you it is a reality, so no reading +/- a few digits on the far right of the decimal point will ever be relevant, on any meter. At best one could consider +/- half a decimal place for a good steady meter with good probes that are properly placed with good clips. However, there is likely to be greater variation in voltage output directly from the alternator with a steady load, when the only variable is fuel blend.
bio_cowboy
 
Posts: 27
Joined: Tue Nov 25, 2008 6:01 am
Location: Amarillo, TX, USA

Next

Return to Single Tank Systems and Blending

Who is online

Users browsing this forum: No registered users and 0 guests

cron