Phoenix Turbine Builders Club

Previous Back

Fourth Quarter 2006

Electronically-controlled air-pressurized oil jet system for 4.5" Experimenter's Tesla Turbine

October - December 2006

In our last article we mentioned a problem one of our Club members was experiencing with the oiling system for his 4.5" Experimenter's Tesla Turbine. It seems that oil was feeding too quickly into the bearing block and subsequently blowing past the air seals. In this article we are going to discuss oiling systems and DIY electronic metering to deliver the correct amount of oil to cool the bearings without flooding the bearing compartment.

The first thing we have to do is determine how much oil we will need for our bearings. As mentioned in a previous PTBC article, there is a very good technical data section in the front of NTN Bearing Company's ball and roller bearing catalog. Most bearing companies have similar sections in their catalogs. Technical information may also be downloaded from company websites. 

The oil we are using is DTE24 -- a machine oil with a viscosity of ISO VG32 (light duty). This oil rating is good for an operating temperature as high as 60 degrees Celsius and a maximum of 80,000 rpm for all types of bearings. According to the formulas and charts supplied by NTN, a good starting point is around 200 cubic centimeters or 200 ml of oil per minute.

Oil delivery systems

There are several methods of delivering the correct amount of oil to the bearings.

  1. Vane/gear pumps with flow limiting orifices/valves

  2. Electronically controlled diaphragm/gear pumps

  3. Air-oil mist system with electronically controlled oil quantity

  4. Air-pressurized oil jet with an electronically controlled pulse gate valve

For reasons we will mention in the following section, we chose option (d) as our final approach.

Experiments

Initial tests of pump using waterWe began our experiments with the vane pump we purchased from Surplus Center. As we showed in last year's experiments, the pump worked just fine driving a fluid such as water with low viscosity. Once we switched the working fluid to DTE24, well, that was a different story. The higher viscosity of the oil created so much friction in the feed lines, the flow through our oil jet slowed to just a trickle. Checking the pump outlet pressure with an automotive manifold gauge indicated a mere 1 psi -- much too low for a jet spray into the bearings. The solution here is to find a pump with a higher psi rating.

Our next experiment utilized a gear pump from an oil oil furnace spray head, and a VW fuel injector. These gear pumps are set to around 60-80 psi, which is too high for our (KIP) solenoid valve, but just about right for fuel injectors. The problem we ran into is that the orifice in a fuel injector is so small, the amount of deliverable oil was too low for our oil jet, resulting again in just a trickle.

There are two ways to work around the problem:
bullet

Decrease the operating pressure of the pump to 10-20 psi (our KIP valve rating is 20 psi with a 1/4-inch orifice), or

bullet

Order in a new solenoid valve with a higher psi rating (smaller orifice).

Since we wanted to explore all options, we moved on to the next proposed systems.

We next tried an electric automotive fuel pump. When powered from a power supply with sufficient amperage, the Bendix pump delivered 2.5 psi, which was borderline usable pressure for our oil jet. The problem is that the pump used too much amperage for our pulse circuit.

Zeroing In on the Solution

In our last two experiments we moved into pressurized air instead of an oil pump per se.

First we connected our air supply to an off-the-shelf oil misting system (for cooling tool bits), and set it to 30-40 psi of deliverable air. The result was amazing -- a smoky cloud of finely vaporized oil -- but not very usable in our oiling system.

Moving on to our final experiment, we partially filled a small metal pressure tank with DTE24 oil, pressurized the tank with 20 psi of air, ran the tank exit line through our solenoid valve, and finally to our oil jet. After connecting our pulse control circuit to the solenoid valve, the system worked flawlessly. We were able to set the air pressure back to around 10 psi, again with flawless results.

The Final Design

PTBC 71 - Diagram 1
Diagram 1

Diagram 1 shows the mechanical layout of our air pressurized oiling system. Air enters a pressure tank (a) filled with machine oil, forcing the oil out of the tank and to the control valve (b) The control valve is pulsed with our proprietary (12 vdc) PWM circuit (Schematic 1). Oil is gated to the oil jet (c) at a rate specified by the bearing manufacturer, and determined experimentally.

Ken Rieli's PWM Oil Pulsing Circuit - part 1
Schematic 1 (part 1)
Oil Pulsing Circuit - turbine bearings

Ken Rieli's PWM Oil Pulsing Circuit - part 2
Schematic 1 (part 2)
Oil Pulsing Circuit

PTBC 71 - Oil pulsing circuit, power supply
Power Supply Circuit

Schematic 1 shows our proprietary electronic design for controlling a solenoid valve with a 12-14 vdc variable frequency, variable width waveform. 

When developing the prototype circuit, we used an old satellite dish actuator transformer with a 12.25 volt secondary, fed it through a bridge rectifier, and then to the pulse circuit power feeds. (You can use any 12 volt source that is able to supply at least 1 amp at 12 volts DC.)

We assembled the low power components on a breadboard, and the voltage regulator and power transistor on a non-copper clad phenolic circuit board. In the future, all of the circuitry -- except the transformer and bridge rectifier -- will be mounted in a project case.

The potentiometers shown in the circuit control the charge and discharge rate of the 50uf capacitor (connected between pin 6 of the 555 timer and ground). R2 (500 kohm) resistor controls the frequency, and R3 (5 kohm) resistor controls the pulse width or "on time" of the oil valve.

Conclusions

This article will help move experimenters in the right direction on oiling systems.

The few experiments demonstrated here should give you a pretty good idea of the developmental process. And while we did come up with a good working system, there are other approaches to the problem, as we mentioned. There are also enhancements to the proposed system to make it a completely closed-loop system that were not covered in this article -- we'll leave that part to other club members, or to a possible update in the future.

For those of you who wish to duplicate our design, you can find the electronics parts at Jameco, Mouser or Digi-Key. The solenoid valve we used is a KIP Inc. part #U651118-02. For higher pressure pumps, order a valve with a smaller orifice. Another source of valves in 24 volt DC is Burkert (6011 or 6013 models).

In determining flow rate, simply measure the amount of oil exiting the bearing case in one minute of operations. Increase or decrease the frequency of pulse width of the oil valve with pots 2 and 3.

One last reminder/comment. Make sure you use heat sinks with the 7805 regulator and 3055 To-220 transistor -- also use a bit of heat sink grease when mounting the parts.

Ken Rieli

Last updated: July 02, 2008 11:21 PM

Phoenix Turbine Builders Club

FREE Open Source Forum

http://phoenixnavigation.com/ptbc/home.htm