Hi all,

After reading quite a bit here on the forum and getting some needed answers (I'm sure I'll have more), this is what I am planning to build. This took me the better part of 3 hours to modify Graham's GL1 ecosystem diagram:



Guys, I'd appreciate any feedback before I actually start building this thing. The tank is a 40 gal electric water heater. I am going to start with the easy stuff first (black pipe). I am in no great rush to build this, but the misses would like her half of the garage back!

Thanks for any and all comments, esp. you Graham if you're reading this!

This message has been edited. Last edited by: Bob_98SR5,

Nice design Bob, congratulations.
I haven't had time to really study it, but at first glance there's something that caught my attention: the HV2-HV8 combo.
Is there any reason for using 2 valves there? It seems to me one should be enough.

Just another little thing.
If you drain the BD through the standpipe, you will always be left with some perfectly good BD at the reactor's bottom.
I would connect the pump inlet to the same drain you use for the glycerol, with no standpipe.

Hi Bob,

I'm envious of your presentation skills. Good Job!

Using a dedicated standpipe for the input to the pump will cause a buildup of glycerin in the bottom of the tank. It will be settling during the mixing phase, taking methanol and catalyst out of the reaction when it does. Result: It will be difficult to pass 3/27 when drawing from a standpipe during mixing.

berny and rick,

ok thanks for the advice. i read over in the apple turnover that it was advised to add a standpipe to it, so thats what I did. if what you say (rick) is true about the glycerine buildup, i'll stick with GL's original plan. its already welded in, so i'll just plug it off

also, i realized i didnt draw the input from the pump to the venturi very accurately. ill need to use two maybe three 45 deg elbows. i will actually have to build it first to see if it works though.

i'm going to revise the drawing.

Bob,

Very nice diagram. Easy to follow. I agree with Rick. At least part of the input to the pump needs to come for the bottom if you want to maximize the use of your inputs.

Also, regarding mixing, consider having the top input to the tank point downward, possible restricted and into the layer of the fluid to enhance in tank mixing. Alternately, you can have upper and lower pump inputs, mixing the upper and lower layers and preventing any partial separation.

If you're going to use a standpipe in a turnover then make your draw available to the pump as well from the lowest point. The reasoning behind a standpipe in a turnover is the same as in a settling tank, you get to settle, drain off the overflow and then all you get is biodiesel minus glycerine to send either to a demething tank, or wash tank. You then use the floor drain to drain off the glyc.

In your diagram if you extend the piping from the glyc drain over to the pump with a T fitting and then move the redundant ball valve (HV8)over you can use it as a selector from teh floor drain and the standpipe making what I described above simple.
HTH

Guys,

Ok, I revised it, incorporating berny and rick's ideas. I believe I captured some of the same ideas that Legal and Andrew mentioned too. I also added explanations for the two new valves that are not on the original GL1.



Any additional feedback/critique is greatly appreciated. I want to make sure I am building this correctly for my own good and for the good of the community, should anyone follow this design. I have thick skin, don't worry

Bob

This message has been edited. Last edited by: Bob_98SR5,

I know GL has done lots on this, but does it really seem best to have the methoxdide input beyond the pump? Seems like you are missing a chance to get great mixing between the oil and methoxide in the pump. I suppose if your pump is pushing enough volume, this won't matter, but I wonder if that is really the best idea.

Anyone do any comparative testing on that?

Andrew,

Regarding your suggestion about pointing it downwards, I may have a natural (not created by me) solution. The concave shape of the now top of the processor limits how much pipe I can thread before it hits the inner bottom portion of the concave dome. This is illustrated by the yellow dome with the words "spray foam". I intend to use Great Stuff to fill that void once I do a leak test on the 1/2" pipe I will weld in for a vent line.

I took a rounded file and stuck it in until I hit the dome. I marked it and it measured about 3.5 inches. So I envision that the methoxide + oil mix will hit the dome and naturally flow downward when it contacts the dome. Do you think that is good enough? Here's what I'm talking about:



Or should I create a restriction and/or downward pointing device? I have a welded and I enjoy bashing things into shape.

I discussed this briefly with Rick on the phone about adding an eductor (sp?) but he felt it was unnecessary b/c I am not making large batch sizes. Hopefully I am quoting the same topic and Rick's logic.

Bob

This message has been edited. Last edited by: Bob_98SR5,

You may want to consider placing the venturi between the two 45 deg elbows. Otherwise if the venturi is mounted horizontally, it will need about a foot of pipe after the venturi and before it dumps. It needs a minimum of backpressure in order to work properly.

The oil return where it hits the upturned dome is pretty close to the top vapour outlet. Don't know if this will be a problem.

Because the vapour outlet is vertical up, the pipe will need to be really well insulated until it starts heading down again. Or else vapour will condense in the tube and fall back into the processor.

Will the suction created by the venturi be enough to lift your methanol or water up to join the oil flow? Might be more fun to add it before the pump.

You might decide not to worry about water 5% prewash. Last week I got 99% pure recovered methanol without prewash in my eco-system. The oil was dry for a start. Very pleased.

Andrew and Paulus,

Regarding your suggestion of routing the methoxide or water tube in front of the pump inlet, GL says that some pumps lose suction if you route it that way. He writes "Here's another way you can inject methoxide or water - it will be familiar if you've used the appleseed design. Some pumps cavitate and lose suction with this scheme. If in doubt, use the 1st design, with valve V4 going to the venturi inlet."

So that is why I routed it the way I did (playing it safe). It does make sense that the oil and methoxide hitting the pump vanes is probably the best way to assure complete mixing of both. Rethinking it again, I think the pros of vane mixing outweighs the risk of incomplete mixing if designed as I have it in version 1.5. So I'm going to take your suggestion and going to design it with the methoxide meeting the oil prior to hitting the pump. I'll give that a try first. I'm revise the drawing once more.

Paulus: I'll insulate the heck out of that vapor outlet too. Thanks for the advice. I recall GL has a pic of which pipes need the most insulation too. And good to know your results on not doing the 5% prewash. Another step removed is always a good thing.

And about the vent being affected by the shooting oil, I thought about that and I'll just see if it does or not. However, that got me to thinking about the 30psi T&P valve. The only place I could put it is in the lower (now top) heating element---which is exactly right under the venturi and the oil coming down. That's not a good place, I imagine. So I drilled another 1/2" vent hole at the top and I'm putting the T&P there.

Rick: Ok, got it re putting the venturi in between the elbows. On a side note, if I were to mount the venturi horizontally w/ about a foot of straight pipe after the venturi, the only place I could do it would be at the bottom (now top) heater element hole. And by doing so, that would leave a pretty big void at the top of the reactor. I recall that GL advises that your reactor should be 75% full.

And that raises another question: if my tank is 40 gallons, does that mean if I fill it to the top, it will be 40 gallons? I vaguely recall that this is not the case.

Your alternative is to switch ports. Put the return on your current vent and the vent on the current return. Then you place the venturi on top of the tank horizontally with one elbow to point it back to the tank.

If you are using the little blue clearwater pump, they can handle more entrained air than most, up to about 20% before they cavitate. The grounfus ?sp? pump Graham uses can not take nearly as much entrained air.

Hot water heaters are calibrated "to deliver" they hold ("contain") about 10% less than they can "deliver". Make sure you leave room for your methoxide as well.

Rick,

Like this?

Re venturi placement and running the return lines, as you can see, there's a long length of pipe running after the venturi, down 2 45 deg. elbows and straight down into the tank. I beleive thats what you were saying would be optimal. Note too that on the other end of the pipes and elbows coming from the pump, I used 2 additional 45 deg elbows and added a sight tube just above the pump.



Oh, I forgot to ask: I've seen unions used in various processor pics and plans. Most have been at the input and output on the pump. Are there any other places where you recommend unions be used? These little devils sure are expensive though!


I think this is it, fellas.

This message has been edited. Last edited by: Bob_98SR5,

You're wasting HV8, it's not really doing anything.

Now if you extend HV8 straight down to a T fitting you then have the option of drawing from the short standpipe for transfer should you do a double batch and settle in the main reactor. You can still draw a sample using HV9 during reaction.

After settling, purge the level via the floor drain and then a little from the SP to drain the settled glyc in it and then you are good to go draining only biodiesel leaving the residual glyc below the SP for draining later.

The original idea of using a Standpipe in a turnover was to settle in it as well as a space saver, and then to maximise leaving the glyc behind while transfering to demething or washing.

Some are space challenged and don't/can't have a separate settling tank, so the SP turnover to the rescue.

Just posting to watch easier .

quote:

I've seen unions used in various processor pics and plans. Most have been at the input and output on the pump. Are there any other places where you recommend unions be used?

The unions at the pump are a good idea so the pump can be replaced or repaired without having to take everything apart. It is a good idea to have a couple of spare unions on hand when you do the plumbing as sometimes things wont screw together as planed.
I welded the fitting for my condenser on the uppermost part of the top of the reactor as this is where the fumes will be most concentrated, If you weld it low in the dome I wonder if you will be losing out on some of the methanol fumes.
For the vertical pipe leading to the top of my condenser I heat it with heat tape up to about 170F, this made a huge difference for the better on how much methanol I can recover.
Jon

I'm wondering if valve HV3B is redundant. Without it you can still isolate the pump, introduce oil, drain glycerol, and pump. Thoughts anyone?

Also, if you could get the pump inlet level with or slightly higher than the four-way tee coming out of the processor then your HV9 purge valve is unnecessary. There's no point in shutting the pump off while draining glycerol, as it will be in the pump lines too. In practise glycerol always sinks to the bottom and so drains from all areas equally, including backwards through the pump lines. As it is you're just creating a glycerol trap necessitating HV9.

I think also your sight tube should be plumbed into the pump inlet manifold, after valve HV2. Then you can suck it out with the pump, if necessary closing slightly HV2 to redirect suction. One of the biggest benefits of this is that you will not have unreacted oil trapped in the sight tube line, plus also being able to drain it saves the tubing from hardening.

Also, what is the red funnel for? You could plumb the top of the sight tube into the vapour line coming out of the processor, or to the vent. That way when you fill or empty the sight tube you won't be introducing new atmosphere into the reactor, just swapping like for like = fewer emissions.

quote:

if my tank is 40 gallons, does that mean if I fill it to the top, it will be 40 gallons? I vaguely recall that this is not the case.

Hi,
You'll probly be getting a 20-25 gallon batch.

Also, if you have a welder and dont mind re-insulating the tank, you turn the concaved 'top' into a dome. This will get your recovery vent at top center.

Here's how I fixed the dome problem;
http://biodieselpictures.com/viewtopic.php?t=669
If you have the means to customize the tank, I would, and do.

I have a question. What software are these diagrams made with? My old release of AutoCAD wont work on my new pc. (64bit)

Brian