The GL Push-Pull Eco-Processor

This is a more advanced version of the popular eco-processor, which adds these new features...

  1. Dewaters the oil at low temperatures, using vacuum
  2. Recovers methanol from biodiesel and glycerol at the same time
  3. Uses a non-siphoning heater tube to minimise the chance of accidental tank fires
  4. Uses a low cost vacuum pump designed for dewatering and methanol recovery

The system is compact and easy to make, but the processor tank must be able to operate at reduced pressure without collapsing. Normal 45 gallon or 205 litre drums are not likely to be suitable. (See here)

Propane gas tanks are ideal as they are strong, have domed ends which aid draining and are reasonably cheap.

See here for changes and amendments to layout, as I get suggestions from folks.

See here for a nice writeup by 'handcuff' on his build.

See here for Doug Weiner's photos and description. A video by Doug is here.

Basically it is a reactor with reversible routing for the pumped fluid. Normally suck from the bottom, return to the top during reaction and neutralization.

Then reverse the flow for distillation, so that the pump doesn't get burdened with ever thickening glycerol.The heated flow into the base of the tank adds turbulence to the glycerol.

The bunch of valves around the pump allow you to reverse flow. A non siphoning heater in the pump's outlet provides the necessary heat for the process. It has a failsafe narrow bore relief pipe, so that if heat is applied with all valves accidentally closed, it will prevent over-pressuring / rupture. JohnO of Moses Lake, made a great suggestion to use a pair of 3 way valves instead of 4 single valves, to control flow direction. That would mean the relief pipe could be omitted - Thanks JohnO! Make sure you use 'full-bore' valves, by the way.

Johno later suggested using 2 pumps, one pumping up, one pumping down, like this...



Then bernyjb has discovered that some Northern Tool pumps can be made run in reverse. See this post.

As you can tell, this is a design which is in its early stages of evolution, and there are sure to be loads of great suggestions by folks on ways to improve it - please keep em coming!

The vacuum pump is designed to handle either water vapor or methanol. Many folks seem to use refrigerator pumps and that concerns me, as there is a strong chance of condensate gathering in the pump and causing problems.

So I use a venturi, carrying heated, low viscosity rapeseed oil as the vacuum creator. It is heated to above the atmospheric boiling point of the material being distilled, to ensure distillate doesn't collect in the small rapessed storage tank, which is only a liter or two in volume. I have a sight-gauge on the bottom of that tank so I can spot any water which may condense, if I do something stupid like distil with the vacuum pump cold.

I use rapeseed oil as it is non-soluble with methanol or water and has a low vapor pressure, so wont vaporize as it passes thru the venturi throat, unlike water. It can become acidic and rancid after prolonged use but hey - I can make great biodiesel out of it if it becomes 'tired' of acting as a vacuum pump and needs renewing! The ever so slightly super-heated vapor then passes to a condenser at atmospheric pressure.

The thing at the top of the reactor is a foaming-trap, as the glycerol and biodiesel can foam as a result of the soaps. The trap contains steel wool, which the bubbles burst onto and the liquid is released to fall back into the processor under the liquid level at the right, to ensure the bubbles only come into it from the left. You will not need this trap if you can control the rate of temperature rise in the processor - a ramping PID controller would solve this for you.

The processor is designed to be used specifically with the following components...

  • Non-siphon heater
  • Venturi vacuum pump with heated rapeseed oil carrier
  • Plumber's delight condenser

Support information...

1. How to make a low cost venturi from standard plumbing parts.
2. How to make an efficient, low cost condenser from standard plumbing parts.
3. A safer electric heating system
4. How I made the methoxide / glycerol / distillate tank vents.

A few parts and sources if you are going to make your reactor from iron/steel pipe (recommended). Any prices are included just a guide, and can change at any time.

See also PipeFittingsDirect, whose prices can be better than BES for some items. I'm also working on McMaster Carr parts lists.

Some parts

Notes on neutralising the catalyst
After reaction, keep the pump running and do NOT drain the glycerol, and do NOT do 5% prewash.

Instead, add 1.5 grams of phenolphthalein powder per 100 litres batch and allow to mix thoroughly. You'll need the sight tube open to view the mixture, you want to be able to instantly see the colour change, so the tube must carry some of the pump outlet flow thru it. The pump must be drawing from the borttom of the tank during this phase.

Caution - highly corrosive chemicals!

WEAR CHEMICAL SAFETY MASK and GLOVES and very slowly (I mean VERY SLOWLY) dose concentrated HCl (you can use 'brick acid' or Muriatic Acid for this) into the inlet of the pump and note the colour change. Use a thin capillary tube to limit the feed rate, and to prevent possible blow-back from the inlet port, which could occur if the acid boils.

Click on images for examples of safety equipment suppliers...



When you get to about 0.1% dosed volume, you'll see the deep red will lighten up a bit.

STOP adding HCL at this point. The colour will darken again in a few minutes, DON'T add any more HCl or you will start to make FFA from the soap. The catalyst will preferentially neutralise first with the HCL, because of its higher reactivity, then the soap.

The NaOH or KOH should mostly be neutralised now, making it an ineffective catalyst for what follows...

Apply vacuum, reverse the pump flow, so you are pumping biodiesel into the bottom of the tank and add heat to the whole batch and start distilling off the methanol, you'll probably need to go up to 70C or thereabouts near the end, but don't go any hotter. We don't want to aid the backwards reaction.

When the methanol recovery slows to a trickle, stop.

The glycerol will drop out within a matter of seconds when you stop the pump, drain it straight away, before it cools and solidifies.

Then drain off the remaining biodiesel into a settling tank to allow soap to drop out, or you can wash it gently as is - it should wash fairly easily if you treat it gently for the first one.

Key things
1. Don't overdose on HCL.
2. Don't dose HCL too fast, or you'll make FFAs

For a 100 litre batch, you'll use around 100ml HCL, depending on titration, amount of lye, water in oil etc.

My pump is roughly 1/2 batch volume per minute.

Dose the HCL at a rate which will take around 30 minutes to complete, at the inlet of your pump so it mixes well.

This is info to the best of my knowledge to date - there is a lot still to learn and much can be done to improve the process, but it is showing good promise as a simple, fast efficient way to recover methanol from the whole batch and aid washing or eliminate the use of water.

I can't guarantee that the processor will work correctly or safely if you change the design of these components or if you install them differently to how I describe, or if you change any of the piping or valve layout, but I'd welcome the chance to discuss any design changes you may want to suggest.

Doug Weiner has published details about his push-pull processor.
Take a look here Doug Weiner's Push-Pull processor




Have I goofed?
Can I help with anything?
Want me to add something?

Feel free to contact me at grahamlaming@hotmail.com

All the best,

Graham Laming signature
Graham Laming