Witte CD-12.... how slow can you go?

[bschwartz]

I have a Witte CD-12 that is rated 12 HP @ 750 RPM.  The generator head is an ST-12 (1800 RPM 4 pole head) with a 13" pulley run off of the 32" engine flywheel.  There will be times when I don't need all 12HP (6KW) of power, and would like to slow the engine down for greater fuel efficiency. 
I have a 8.25" pulley that will fit on the generator.
Running it off of the 32" flywheel, would put the engine at approximately 464 RPM.
Assuming the oil pressure is where it needs to be (~30 lbs.) can I run the engine this slow, or will it cause problems with the engine?  Potential issues as I see them are poor lubrication (I'd think would be OK as long as there is oil pressure), fuel timing too advanced for the lower speed, increased bearing pressures.

Do you think any of these would be a problem at low speeds, or are there any other issues you can think of?

Thanks for any input.
-Brett

[mobile_bob]

provided you have enough oil pressure at the lower speed
i don't think you will have any issues

timing really won't need to be changed as there is really little change in rpm
even though as a percentage it might seem that it should.

what you may want to do is this

running the engine with no load at the higher rpm, carefully measure how much fuel it consumes over a fixed period of time, with the engine thermally stable.

the repeat the test at the lower rpm and see if there is any real/appreciable difference in fuel consumption at the lower speed.

you may find that the fuel savings are not significant enough to warrant operation at that speed.

also you will want to try to determine if there is a critical speed somewhere near the target low rpm speed, if for instance the engine produces some odd vibration(s)  somewhere around this lower rpm, you would be well advised to not run at this lower speed. even if the rpm is just above this critical speed, what might happen under a load is the engine might droop just enough to get into the critical speed operation and that is where things will break. sometimes minor things, sometime major parts like a crankshaft.  just stating this fwiw, just i case you hadn't thought of this concern.

it might well be that under testing you will find the engine is too close to critical speed operation, but still runs with a reduction in fuel consumption significant enough to go after. if this is the case, you might consider running another test at another higher rpm, maybe 500rpm, maybe again at 550rpm, etc. the goal being finding a lower rpm that captures most of your fuel savings goal while operating well above the critical speed operation.

from my experience obtaining a 5% reduction in fuel consumption is very hard to do running at a lower rpm. that is probably about the most reduction you will see.  so my advice would be to not risk damage chasing minor gains, unless you do some careful testing?

anyway that is my thinking on the subject

fwiw
bob g

[BruceM]

Isn't the bigger issue the need to keep the engine loaded adequately to maintain temperature and avoid a carbon fest?

I thought that was the main point of slowing down rpm for smaller loads.

[bschwartz]

Yes Bruce.  In my long winded post, I forgot to include that reason as well.
I agree it is probably the most important reason to slow down the engine when under a smaller load.
So, any guesses as to problems, or do you think I'd be OK?

[Jens]

what you may want to do is this

running the engine with no load at the higher rpm, carefully measure how much fuel it consumes over a fixed period of time, with the engine thermally stable.

the repeat the test at the lower rpm and see if there is any real/appreciable difference in fuel consumption at the lower speed.

I would think that this would be a perfect red herring. I doubt that any change in rpm of an UNLOADED engine will tell you anything of value.
I would think that any test would require the target load, run at low speed and then again at high speed (same load) to get an idea of the possible gains. Run long enough to see if there are carbon issues.

BTW, I would agree that gains would be minimal as you only save a bit on windage and running friction which could easily be reduced by inefficiencies of a slower than designed engine speed.

My wild guess anyhow ......

[bschwartz]

I think Bruce really nailed it with loading the engine at a hgher percentage at low speed.  If I only need 3KW, running it at 750 RPM could cause carbon build up.  That would really be the important reason to slow down the engine, fuel consumption really isn't that important.

With the listeroid I didn't mind experimenting too much as head gaskets are cheap and available.  Pulling the head to check for carbon, or fouled rings, or injector tips are no big deal.
I'm a little (OK, a lot) more afraid to 'try' things with engine.  It may be more robust, but if something goes wrong, it could cost me DEARLY!!  Even pulling the head to check for carbon could be a problem if I can't reuse my head gasket.

OK, oil pressure check.  No strange sounds or vibrations check.
Anything else?

[mobile_bob]

if the goal is to reduce fouling, that is fine, but
do you know that fouling is a problem at part loading running at 750rpm?

is the engine cooling system tstat controlled? or hopper cooled?

my mention of the no load tests was based on engines having a baseline consumption number, as mentioned by jens engines have frictional and windage losses. these losses are not always linear with rpm changes.  how much of a change would need to be determined by testing before i went any further and it is very easy to do.

measuring under load would require a different pulley drive ratio, and that ought to be about the further down the  test list in my opinion, one should exhaust all the other testing that requires no additional cost in hardware first. or so it would seem to me.

so i would test in the following order

1. first determine if the engine will run at the target rpm, if ok, go to step 2

2. determine if it has enough oil pressure at that rpm, if ok, go to step 3

3. drop an additional 50rpm and see if the thing runs without undue vibration
and cycle from that rpm slowly up past the target rpm to maybe 50rpm above, to see if you run into vibration issues. if ok, go to step 4

4. run a baseline fuel consumption test unloaded, both at the rated 750rpm and at the target rpm. if there is significant reduction in fuel consumption, go to step 5

5. run a specific fuel consumption test at full load 750rpm, part load at 750rpm, and at whatever load you would like to get at the target rpm but running at 750rpm, you might consider a test at 6kwatt, 5kwatt, 4kwatt, 3kwatt, 2kwatt, and at 1 kwatt, so you can plot a curve. then go to the next step

6. run a specific fuel consumption test  with the new gear ratio at the new target rpm at no load, at 1kw, at 2 kwatt and at 3kwatt... so you can develop a curve, then

you can compare the two output curves, and the info contained to see if there is significant reduction in fuel consumption to be had.  whatever the case might be i would certainly look into doing these next two steps with hardware if you haven't already.

1. get the coolant setup with a tstat, and running as near to 195 as i could get it.

2. if it already has a radiator and a belt driven fan, i would remove the belt and install a thermostatically controlled electric fan. this will aid in keeping a more stable coolant temp and also cut some power needed to run the old belt drive fan. power to drive the fan goes up with the cube of its driven speed iirc, which if that is the case running at a lower rpm might reduce the power needed to drive the fan very significantly?

fwiw
bob g

[mobile_bob]

another thought that you might want to consider

that is the governors ability to regulate rpm at the new lower target rpm
it might be that the droop and recovery is such that you will have to address
this in some manner, maybe with spring changes, maybe with weight changes, maybe with linkage changes, some/part/all?

also as is widely accepted now, the first kwatt produced is the most expensive in terms of fuel consumed, because of the baseline frictional and windage losses of the engine, the goal of reduction in fuel consumption generally isn't as large as we hope for.

now as far as the goal of reducing fouling, that might be a bit harder to attain, every engine type will be a bit different in this regard.  although getting and maintaining as near as 195 F is likely to get you where you want to be in this regard.

what i would like to see is some well documented testing, so that we might learn more about how all this inter relates.  with good testing and some good results we ought to learn enough so that we might predict certain outcomes vs certain operating parameters for the witte.

we already have some good well documented numbers at various loads, rpms, temperatures etc on the changfa's and listeroids,  it would be nice to add witte to the list and see if there are some similarities that illustrate some common ground.

at which point we might be able to answer questions for the next guy down the road that is considering such operating parameters, when he asks what gains can be had by changing the parameters and what pitfalls or deal breakers is he likely to encounter.

otherwise we are doomed to have to revisit and repeat history over and over again, having to reinvent the wheel as a group.

not saying that is a bad thing, but maybe not the most efficient use of our time or money?

some of the most interesting testing i have seen done was done by folks like Ronmar on his 6/1 some years ago, and Veggie with the slowspeed 195 project.  in Ronmars case, we were able (he and i ) to predict with amazing accuracy the BSFC at various loads by first knowing the BSFC baseline unloaded.  in Veggies case i for one learned alot in regard to what is possible with the changfa running at dramatically lower rpms.  i think we not only need more of this sort of quality testing and documentation, but we probably out to archive this under a topic in the white papers section?

just some late night rambings

;-)

bob g

[bschwartz]

Bob,

Thanks for the insight!!
I think I probably should NOT have started the thread with "fuel consumption" issues, as loading was really more important to me.  That's what happens when I start a thread when I'm tired.

You put a lot of thought into methodology for testing and documenting consumption rates, first KW, windage etc.

I am just not the guy for that kind of record keeping......
One of the reasons I do IT for a living is there is almost no paperwork.  Someone reports a problem (network outage, virus, hardware failure, printer out of paper....) and I go fix it.  Then I go on to the next...

Let me go over a few things you did ask as best as I can.

     "if the goal is to reduce fouling, that is fine, but
do you know that fouling is a problem at part loading running at 750rpm?"

I don't even have the engine mounted yet, but I don't know how I could test for fouling without pulling the head which I am afraid to do (parts availability) unless I have to.

     "is the engine cooling system tstat controlled? or hopper cooled?"

The engine is hopper/condenser cooled, so no thermostat is possible. See this picture from this thread.
http://www.microcogen.info/index.php?topic=2883.0
Coolant fills the water jacket around the cylinder, and is just high enough to be slightly into the hopper.  As the engine heats, the coolant boils in the hopper, and rises into the condenser (looks like a radiator, but it isn't liquid filled) where it cools and condenses, then dripping back down into the hopper.  Elegantly simple, but I don't see any way to regulate temperature.

     "1. first determine if the engine will run at the target rpm, if ok, go to step 2"

A quick test run while sitting on el cheapo wheels (just designed for moving the engine, NOT portability) show that it seems pretty smooth at fairly low speeds.  Actually I was afraid to run it up to full speed due to starting to jump around 

     "2. determine if it has enough oil pressure at that rpm, if ok, go to step 3"

YES, it develops full oil pressure at a very low speed, so we should be good there.

Your next questions pertain to 50 rpm changes, and until mounted, I can't dial the speeds in well enough to determine.
You then go into fuel consumption, which really isn't my real concern.

     "1. get the coolant setup with a tstat, and running as near to 195 as i could get it."

Wish I could.... see above.


     "2. if it already has a radiator and a belt driven fan, i would remove the belt and install a thermostatically controlled electric fan. this will aid in keeping a more stable coolant temp and also cut some power needed to run the old belt drive fan. power to drive the fan goes up with the cube of its driven speed iirc, which if that is the case running at a lower rpm might reduce the power needed to drive the fan very significantly?

Oooh, I forgot about the belt driven fan....
I have absoulutey NO IDEA what effect the slower fan could have on this situation.  Being hopper/condenser cooled, I may need a certain amount of airflow even at lower loads.  I just don't know.  As fuel isn't really the issue, and loading is, higher loads from the fan actually would help keep some load on the engine (normally an efficiency drawback, but a positive in my case)
     
     "another thought that you might want to consider
that is the governors ability to regulate rpm at the new lower target rpm
it might be that the droop and recovery is such that you will have to address
this in some manner, maybe with spring changes, maybe with weight changes, maybe with linkage changes, some/part/all?"

This would be a deal breaker!!!!
I am not going to modify internals on an engine like this (see earler posts about willing to experiment with a listeroid, but not an antique)
If the governor wont work at the lower speed, then this thought process is done.
With modifications, I'm sure a lot is possible.  This whole line of questioning is about what potential problems I could run into.  I'm hoping to determine IF it will work, not how can I make it work at the lower speed.

I guess what I'm hearing is "just try it and see"

Thanks for helping me think through this a bit Bob.  I appreciate your time and insight.

[deeiche]

Brett

If you are not really looking at fuel consumption why not just add in a load dump during period of light load.

Something like this below.   

[Thob]

Even on an engine with pressure lubrication, some things may be splash lubricated.  I think the cylinder walls, wrist pin, and possibly the big end of the rod may be splash lubricated.  I'm not sure about the crank mains, this Witte has tapered roller bearings, doesn't it?  If you run the engine too slow, there may not be enough splash to properly lubricate everything.

Does anybody have a manual for this beast?  Maybe there is a manufacturer recommended speed range?  That would really be helpful.