Feathers
Veteran Member
Joined: August 2010
Posts: 169
|
Post by Feathers on Jun 26, 2011 1:39:48 GMT -5
Stoffe,
The engine has been started on propane, and for several tests, I idled the engine on propane, and cracked open the diesel valve to throttle it. Haven't got the guts to run on pure diesel yet (having heard the stories of frank whittle's first engine, and how pooling liquid fuel in the combustion chamber at low speeds lead to it's self- destruction) But I'm sure I will have to some time!
Feathers
|
|
wolfdragon
Senior Member
Joined: April 2011
Posts: 287
|
Post by wolfdragon on Jun 26, 2011 11:27:49 GMT -5
the pooling at the bottom can be countered with a drain down there, if all hell breaks loose just open the valve on that line, cut the fuel, and whatever liquid is there will blow down (flamethrower style, but it's still a safety for the turbine)
|
|
|
Post by stoffe64 on Jun 26, 2011 12:13:33 GMT -5
yes,wolfdragon is right there, a combustor drainage valve is a very good thing on a turbine engine with liquid fuel.
|
|
wolfdragon
Senior Member
Joined: April 2011
Posts: 287
|
Post by wolfdragon on Jun 30, 2011 6:52:16 GMT -5
So where are the pics and running video you promised?
|
|
Feathers
Veteran Member
Joined: August 2010
Posts: 169
|
Post by Feathers on Jul 1, 2011 23:05:20 GMT -5
About that... I'm having issues working out a PWM speed control modoule for my oil pump. At first, I was using a "victor" robot speed controller. This speed controller is rated 40 amps peak at 6-15 volts, and was supplied to me by the robotics club (surplus from years past) at my high school. Now, my little oil scavenging pump has a rating of 2.5 amps (with no load other than pushing oil up 1-2 feet to an oil pan). I don't have a servo tester yet, so I need my transmitter/reciever hooked up to the "victor" controller to supply the ppm that the controller reads. Running at 40 psi (less than 40% duty cycle on the pwm from the victor), and in warm weather (85-90*F) the thermal overload protection circuit on the victor trips after 5-10 minutes of runtime, and I loose oil pressure. This is, mind you, a 40 amp peak, 30 amp continuous controller. So either my 2.5 amp pump is drawing >30 amps at 40 psi, or something funny is going on. I just made today a simple speed controll circuit with a 555 chip and two irf510 power mosfets rated 5.6 A continuous (cooled by two relatively large heatsinks). Everything looked perfect on the scope so I hooked it up to the motor. I don't know a whole lot about driving mosfets, and I wouldn't think that the pump could be drawing more than 11 amps, but those suckers heated up quick, and after 5 minutes at ~75% duty cycle, it quit on me. No spectacular semiconductor fireworks either . So untill I find a way to achieve a steady, reliable, adjustable oil supply, testing is on hold . Ideas welcome! Otherwise I'm just gonna search ebay for a locomotive's diesel-electric powertrain to supply eanough juice to replace my apparently overdriven speed controlls oil pump . Thanks! Feathers
|
|
wolfdragon
Senior Member
Joined: April 2011
Posts: 287
|
Post by wolfdragon on Jul 2, 2011 8:03:58 GMT -5
Not surprising that you are blowing your driver. A brushed DC motor is nothing more than a glorified inductor with a degree of freedom for movement. Your oil pump has WAY more inductance than a servo, therefore the EMF kickback from the end of each pulse is hammering away at your switching components. Another problem is that a large inductive load makes for an insane impulse current requirement at the start of each pulse, which means you need a circuit that can sustain roughly 10-20x the normal load just to have enough capacity for those impulse hits. Note, these only exist for a couple of milliseconds. Example: Airsoft rifle, DC brushed motor and a 8-10 volt battery pack. When you pull the trigger, depending on the motor, the impuse current spike can be well over 150 amps, all the while the 15A fuse doesn't blow because as soon as that pulse is done, the motor is only pulling say 13A. Now make that pulse happen over and over again at high speed, now the fuse will blow because the duty cycle of the pulse lets it occur often enough to heat up that fuse to blow it. You could add a snubber circuit (a reverse biased diode across the motor terminals is the easiest one to check first) I would use a 1N4007, just make it point in the "wrong" direction to run the motor or they will just be a dead short when you want the motor to spin. But this will only get rid of the back EMF, it will do nothing for the impulse current requirements. So why are you pwm'ing the pump when you could just do this and run it at full voltage: www.mcmaster.com/#8088k14/=d017jhMcMaster Carr - Adjustable Brass Relief Valve 25-175PSI Adjustable ~$24 You'll probably spend more getting that PWM circuit right than just one of these in a bypass line.
|
|
Feathers
Veteran Member
Joined: August 2010
Posts: 169
|
Post by Feathers on Jul 2, 2011 21:43:56 GMT -5
Wolfdragon I think you misunderstood my description of the victor The victor unit is not meant to drive servos, it is used to drive CIM motors. It uses the ppm signial you'd get from a servo, and translates it to a 12 volt pwm signial which drives a large motor, designed to move a 115 pound robot (a motor bigger than my oil pump motor). This is the only reason that the issue perplexes me. I've got the reverse-bias diode in the speed controll circuit, bridging the positive rail and the mosfets' drain(s) I would go with the regulator, but I plan to integrate a microcontroller as a PADEC, which will have a feedback program keeping the oil pressure wherever I tell it to go . I've ordered some irfz44n 45 amp mosfets which should be arriving shortly, which I'll install into the existing 555 pwm circuit, and cool with a "BAMF"ish heatsink. If that doesn't work, I don't think anything will. Thanks! Feathers
|
|
wolfdragon
Senior Member
Joined: April 2011
Posts: 287
|
Post by wolfdragon on Jul 2, 2011 22:29:44 GMT -5
Beefing up the circuit like that will help, but you need to hook up your scope and watch the actual voltage on the motor + line, and then you need to put a hall effect sensor inline so you can see actual instantaneous current draw, an amp meter is RMS and you won't see your switching components getting pounded As for the victor, I know exactly what it is, it's an old FIRST robotic motor driver, but those motors are lightweights compared to your oil pump Another thing to consider, what is the PWM freq coming out of the victor, on small DC motors it is not really an issue because of the light loading but a big motor (read large inductance, which I know your oil pump has in relation to a CIM) at more than a few hundred Hz will have really bad tuning issues which will manifest itself in more back EMF (which I am fairly certain is what is heating up your mosfets). Luckily you can try a very simple snubber circuit across the motor terminals might just fix you right up: 0.1uF Cap and a 10K resistor. That should give you enough of a punching bag in the circuit that everything will calm down. As for the feedback controlled oil pressure... just seems like adding a complication where one need not be, why do you want to dial in the oil pressure dynamically? Usually mechanical systems, oil film bearing especially, just want a steady flow at a given pressure... Just seems like an overly complex thing... and I made this: projectwolfdragon.com/Projects/Project%20BlackBox/ProjectBlackBox.htmAnd where is a pic of that popped elbow? Carnage is always the most popular part of a lessons learned discussion. I hope you don't think I am being overly critical, I am very impressed with your design and am very interested to see what you make this engine do.
|
|
|
Post by racket on Jul 2, 2011 23:09:24 GMT -5
Hi Wolf
I'll go along with you when you recommend a simple lube system , it needs to cope with a 30minute run at full oil pressure using cold oil to load the system up as much as possible , substandard lube systems are a recipe for disaster , why put a lot of hard work and money at risk, our lube systems are as important as the rest of the engine if not more so , they aren't just an ancillary, they are the heart of the engine , ..................if it stops suddenly, the engine is dead meat :-( ........................an interlock with the fuel system is essential to prevent such a sad end .
Cheers John
|
|
Feathers
Veteran Member
Joined: August 2010
Posts: 169
|
Post by Feathers on Jul 3, 2011 0:38:50 GMT -5
I've hooked up my scope, and voltage readings are the typical square wave with the inductive signiature of the motor. I do need to get SOME current measurment! My meter is broken! I have no actual numbers as to what this motor draws on straight 12 volts!! The CIM motor is actually a bit larger than the oil pump motor, with a 2.8 amp no load current draw at 12 volts and a stall current of 133 amps. I am not aware of the frequency that the CIM motor runs at, but I'll assume that this unit has been engineered to work with a large DC motor . The working frequency of my homemade speed controll varies +-about 10% with temperature changes in the mosfets but hovers at 250HZ, so I don't think a snubber circuit will be necessary. I will get a hall probe on the scope and take a look to make sure everything is okay! Sounds like a really good idea. And the reason I want this system for the oil is because: #1, this engine will be running mainly hands-free when I work out the arduino control #2, as the engine warms up, the oil viscosity drops, requiring constant manual adjustment of the oil pump speed to maintain the proper pressure for the first 5 minutes of runtime. With the feedback system, I can set the oil pressure wherever I want it to be, and the arduino will do whatever it needs to do to the oil pump to retain that pressure . And racket, I have a 25 PSI pressure switch hardwired into the fuel pump and solenoid, such that if oil pressure suddenly drops below that already low number, fuel is immediately cut and my engine will hopefully be okay. This system has really saved my @$$ the couple times my fussy victor unit shut off the oil pump! Thanks for all the input guys, I really appreciate it!! And I would post pictures, but I have no idea how to with this posting setup!!! Feathers
|
|
|
Post by Richard OConnell on Jul 3, 2011 1:17:41 GMT -5
I made image posting a bit easier now Added a "post image" area to the Reply page (not quick reply area). Then all yu have to do is copy over the code it gives you for "Forum Code." I might simplify this process further as time allows. As for controlling oil with the arduino.. Should be ok, but you need to set redundant kill switches to react the moment things go wrong. Will you have any way to store "flight data" for lack of a better term? Might make troubleshooting easier should anything happen further down the road.
|
|
Feathers
Veteran Member
Joined: August 2010
Posts: 169
|
Post by Feathers on Jul 3, 2011 8:36:22 GMT -5
Hi there! Thanks for the tip Richard, And I do have some safety switches hardwired into the unit. As for the pictures:
|
|
wolfdragon
Senior Member
Joined: April 2011
Posts: 287
|
Post by wolfdragon on Jul 3, 2011 9:08:18 GMT -5
I can understand why your friend started laughing, there is coffee on my screen now...
that is priceless...
|
|
Feathers
Veteran Member
Joined: August 2010
Posts: 169
|
Post by Feathers on Jul 3, 2011 9:21:55 GMT -5
I know right??? It scared the piss out of me at first, but its just so cartoonish...
|
|
|
Post by turbochris on Jul 4, 2011 8:37:39 GMT -5
You all may not know this but Ernie owns Compucar N2O systems. I have never seen or heard of one of his solenoids fail. I would use one as a safety on any turbine running on propane. His valves will not be damaged by liquid flashing to vapor or anything like that. If you're feeling really ambitious, Ernie says his solenoids can be pulse modulated with the proper electronics. The ultra popular NOS (favorite of fans of FAST AND FURIOUS stinkin rice burners..) solenoids are crap.
Also, when you run the oil pump on the PWM controller, what is the voltage drop across the speed controller? If you're putting 12 volts into the controller and the pump only needs 6 volts you could try lowering the input voltage. It's not supposed to matter on pwm controllers but excessive voltage drop (especially at high current) will make them hot.
|
|