I ran the de-meth for 6 hours and got an even gallon out of my 120L batch. That just a little more than I expected (water?) The methanol looks like it has some bio in it as it is tainted just a little, so I was reluctant to put a hydrometer in it. The picture is a sample off the top that has been settling for a day and a half. It's not very clear. So are you saying I need to run it longer?

quote:

I do recover methanol from glycerol from time to time, but not always. I'm trying to find a way to make this more practical and economic.

I’m in the process of designing a GL based system right now with some additions, hopefully people can comment on whether they really make sense or not. I’m a mechanical engineer, so it’s in my nature to take something simple and complicate it for marginal efficiency gains…

Fine fileter pre-treat:
Basically do a really good pre-filter down to a micron or two. Just seems like the less crap that goes in, the better the product that should come out. One of the SVO guys has an inverted standpipe design that I’m going to steal for this (puts the new oil in the bottom of the barrel, pushes the settled oil out the top) that produces dry oil given enough standing time. Should reduce the drying cycle IMO.

External heat source:

The end goal is to combust a glycerin/sawdust log/brickette for heat via steam generation, near term a propane burner. Boil water, duct the steam to the vertical pipe between the pump and the top of the reactor (pipe in a pipe, basically). Collect the condensate for distilled H2O purposes.

I’ll probably vent the exhaust gases separately, and have a shield between the burner and the reactor so as not to make a flame thrower if it springs a leak.
My guess is it will be safer, just because the accidental uncovered heating element problem goes away, but never underestimate the power of determined stupidity.

Separate settling/pre-treat tank:
Thinking of a cone bottom 30-50gal plastic guy with a standpipe just so it’s easy to see and measure how much oil I’ve got in there.

Purpose is to use it for a by-product pre-treat as has been discussed in other threads.
Pump reacted oil to it, settle, draw off the biodiesel for meth recovery from standpipe into the main reactor.

Dump my next batch of WVO in and mix with the hot byproduct.
I’ll need 2 pumps for this setup and have the vertical pipe for the settling tank get some steam too to keep it around 50C.

If I understand it correctly this should ‘recover’ the methanol/NaOH from the byproduct and lower the titration level leading to less over-all usage of meth/NaOH?

Lastly a settling tank with a standpipe 1/3 of the way off the bottom for easy drain. I’ll probably throw a 1 micron filter at the outlet just for grins.

I’m going to model this, because I can (I have access to 3D CAD), once done there is a free viewer that people can use if someone can host the file somewhere, assuming anyone is both interested and patient enough to wait around for me to finish it…

Hi Rick

Are you using KOH?

Can you try this please ....

Divide the contents of that jar in 2.

In one jar, leave the lid on
In the other jar leave the lid off.

I'd bet there's methanol left in there, and you'll see settling in the jar without the lid.

Your airspace of 80 litres is enormous. That will alow around 1000ml of methanol to remain lurking, which will drop back into your biodiesel when things cool down.

Aim for minimum airspace, for maximum extraction.

I'm using NaOH. OK I'll try that, Yes I need to look at a smaller drum to do this in.

quote:

Originally posted by GrahamLaming:
Aim for minimum airspace, for maximum extraction.

Graham,

Is minimum airspace so important if the recovery system is based on your "GL" design?

I understand why minimum airspace is critical if you are just depending on heat, vapor expansion, and some convection currents to "coax" the methanol vapors out of the reactor and into the condenser.

However, doesn't the "GL" venturi driven, methanol vapor transport system minimize the problems of the above type methanol recovery system? This means keeping the airspace to a minimum is not so important in the "GL" system.

Yes/no? Comments?

quote:

However, doesn't the "GL" venturi driven, methanol vapor transport system minimize the problems of the above type methanol recovery system? This means keeping the airspace to a minimum is not so important in the "GL" system.

Maybe. My pump (similar to the HF) is rated at 33 litres per minute. The air flow rate through the venturi suction port is not that great. The GL method relies on evaporation rather than pure distillation, therefore the greater the circulated airflow, the faster and more complete the methanol removal will be. I'm hoping for greatly improved methanol extraction with my pitbull type pump (90 litres per minute). When we tested the pump with one of Jim's venturis, the vacuum was so strong, it collapsed the suction pipe! So, if you're shifting enough vapour, the larger headspace should matter less than if you're using a smaller pump or less efficient venturi. You're basically giving the vapour less chance to condense in the tank itself rather than sweeping it out to the condenser. I still find that a bubbler in the settling tank helps remove that all-important last trace of methanol.

Nick

Well, I found that I couldn't get much out at all until I heavily insulated everything. Maybe I could have gotten by will less insulation if I had less head space. I ended up with four layers of foil/bubble/foil on the side and about a dozen layers on the top before it would flow methanol.

Thanks for the response, Twenty4Seven. I find it very informative.

I have a "GL" type set up with a slight variation. There is no venturi in the system. Vapor recirculation is achieved with a small air operated diaphragm pump. The pump draws from the vent line, after the condenser, just like GL's. The vapors are reinjected through a separate port in the top of the processor.

I can vary the recirculation rate by adjusting the air feed rate to the pump. My pump moves so much vapor early in the cycle that my condenser is flooded and methanol vapors can be lost to the atmosphere. Near the end of the cycle I can't move enough vapor to tax the condenser.

My question, then, is can you quantify how much air you add through the bubbler to drive out the last small amount of methanol? How do you know when to start air bubbling? At about what temperature is your BD when you start your bubbler?

quote:

Originally posted by RickDaTech:
Well, I found that I couldn't get much out at all until I heavily insulated everything. Maybe I could have gotten by will less insulation if I had less head space. I ended up with four layers of foil/bubble/foil on the side and about a dozen layers on the top before it would flow methanol.

Yup, that ties in with what Graham said. With no insulation and a large head space, you've got yourself a huge reflux condenser. You also need to lag well any pipework from the top of the tank to the top of the condenser. The length of the condenser and temperature of the cooling water are important factors too.

Cheers

Nick

quote:

My question, then, is can you quantify how much air you add through the bubbler to drive out the last small amount of methanol? How do you know when to start air bubbling? At about what temperature is your BD when you start your bubbler?

I use the very scientific "trail and error" method Three hours methanol removal and three hours bubbling work well for me. Everyone's rig is different and there are really so many variables that it's dificult to provide a definitive answer. After heating to 75-80*C and circulating for three hours, the temperature of the bio drops to 55-60*C at which point, into the settling tank it goes for the three hours bubbling.

Nick

Just a followup,

My sample above only needed another day of settling. I can see all the way to the bottom of my big batch as well.

When removing methanol, using the GL Eco process; how low of a methanol content must be achieved before you start seeing the soap and glycerin fall out of solution? The current US standard is now similar to the EU std. at 0.2% vs. in the past 0.08%.

GCG

Folks,
I see recurring reference to heating the fluid up to temperature and then running a process of settling for some number of hours. But I don't see any references made to how long to hold the target temperature. Does all of this heating consists of only heating to a target temperature and then cutting off the heat? I'm new at this, still absorbing the basics.
John

JBoy,

Unwashed bio contains about 3% methanol. so for a 100 Liter batch, you would keep the heat going until you recovered about 3 liters, or until the methanol stopped flowing out the condenser.

Thanks, Rick. That clears the fog quite a bit.

To All,
I'm working through the elements of a successful system design based upon Grahams approach with an eye toward using single tank instead of two. I think a single tank can act as reactor/processor tank and settling tank if I set up the plumbing correctly. Any obvious issues I may be overlooking in my ignorance?
John

I'm nearly done building mine thanks to Graham for his plans, just tested venturi and pump and piping just need to test safety tube heater and build condensor before applying insulation will post results when done.

I have a venturi placement question. I am buidling an Apple Turnover style + GL1 Eco-system. See the picture below:



The problem I'm having is that I am trying to use the drain as the location where I'd like to run the venturi pipe. The pipe would hit the bottom of the concave bottom (now the top) of the water heater.

Graham says that it should be approximately 30cm in length. That's a shade under 12".

So I am wondering if its the length of the pipe that matters most or that the pipe needs to be inside of the water heater. If its the former, then I imagine I can have it as shown in the picture above. If its the later, I'll need to move it down to the bottom element (now the top element) or weld in a new fitting somewhere between the drain and the bottom element (now the top element).

Venturi experts please chime in!

Thanks,
Bob

You could get a similar effect by screwing a 3/4"x2" nipple into the drain, then a 45 deg elbow onto the nipple and the venturi into the elbow. It worked for me anyway.

Rick,

Thanks for your reply. So I gather that:

1) it does not necessarily need to be a one straight run of 30cm pipe after the venturi

2) it does not necessarily need to be 30cm of pipe after the venturi in order to work

If you say it worked for you, I will try it but I hesitate just a bit given Graham's writeup says it needs around 30cm of pipe. I'm definitely not discounting your suggeestion, Rick! I just want to do it by the book within limits. So keeping with what Graham says, how about something like this? See attachment.

alternate_venturi_setup.jpg (32 KB, 21 downloads) using a 90 deg elbow with 6" nipples

You need the equivalent resistance to 30cm of pipe. A 45 deg elbow will get you there, a 90 would be too much.

What we are trying to prevent is air pockets in the venturi mixing zone. if the pipe is too short then the oil will fall down from the top of the pipe and only flow in the bottom half with an air gap taking up the top half. Such an air gap would kill suction.

Of course by moving the venturi off the horizontal, you achieve the same end. In this case gravity keeps the oil packed fully in the venturi mixing zone.