paul,

i think you missed what i was saying. I wasn't talking about solidifying while in operation, i was talking about the need for the valve for isolation purposes to support maintenance/repair of the processor while you had a batch in it.

I see what your talking about bro. It's not 100% necessary the extra valve it's just there for the plain idea if you ever needed to isolate your tank from the rest of your system you could just close that valve and make what ever changes you needed to do. So far in my short biodiesel making career I have not had my pump or heating element fail yet but if they did fail during the process I would be able to isolate the tank and make my repairs.

That's one reason that i'm considering not using a submerged heating element. I'm thinking about using heat tracing or another method that doesn't require being submerged in fluid to operate.

quote:

I propose to get rid of it and then take a sample and do a titration, then slowly dose that amount of acid into the mix. It reduces valves, simplifies the plumbing etc.

The problem Graham reports with that method is getting a representative sample from a tank. There is always a gradient from top to bottom. Perhaps if you took two samples, one from the top and one from the bottome, and tried assuming the gradient was linear to see how that worked out...

I have suggested previously in the thread that it is not needful to neautralise all the catalyst to effectively prevent the reverse reaction in the time and temps involved. Removing most of it should make a big difference. So if you had the range of values that Graham's titrations typically give compared to the base amounts used you may be able to deduce a safe amount of acid to add that will always neutralise most of the base catalyst without ever going over and converting soap to FFA.

I have asked Graham for these numbers a couple of times but never had a reply.

good point. even if you add a standard dose based on batch size, then you should be able to simplify the process.

As far as a representive sample - if there are problems getting one, then that is an indication that the mixing action with the pump is not sufficient. i.e. flow rate is not high enough, because if you can get two different samples at different points in your batch then that means that the methanol + catlyst distribution is not homogeneous in the batch, which could lead to under conversion as well as excessive soap production.

Which bring us to the next question - do we really understand how much flow rate is needed to ensure homogenous distribution of the methanol and catalyst - my gut instinct is no. I believe Graham uses a pump that can turn the volume of hitch batch over twice a minute. If this is not sufficient as evidenced by the sample data (or lack thereof), then i would suspect that all these people running around with the blue harbor freight jobbies that might turn a 60 gallon batch over once every 5 minutes could have less soap and better conversion if there were to use a high volume pump.

So to you graham - could you do a couple of experiments for us - i.e. divide your batch size by two and then retry to see if turnover rate has a substantial impact on your ability to draw representative samples?

Wes

quote:

turn the volume of hitch batch over twice a minute.

Maybe i'm wrong here, but i think Graham has a pump that turns the batch over 1 time in 2 minutes or half a batch a minute.

I wonder if your oil had a very routine titration that you could have a standard amount of HCL to add. I would imagine that there's a curve that relates total catalyst neutralization to possible back reaction during distillation, but i think that "close enough" might be the best shot that some of us have.

I'm still very excited about this processor style.

quote:

but i think that "close enough" might be the best shot that some of us have.

I could be easily be wrong but my gut instinct is that it should be possible.

Graham finds there is variation from the expected values from the base amount of catalyst. He speculates that this is caused by variations in the reaction such as soap formation which can eat base catalyst up. Water content could be a crucial variable for example although Graham dries his oil pretty completely and still gets variations.

I would like to know the range of those variations so that I could speculate and then experiment on what might be a 'close enough' amount of acid to use.

Completely agree

I would then also like to see how much reversal you get if you don't bother to neutralize first and ideally see how this varies with temperature. Maybe if you can draw a deep enough value, then the reversal would be negligible enough such that it would not effect the 3/27 test.

quote:

would then also like to see how much reversal you get if you don't bother to neutralize first

Err... too tired ot look it up at the moment but I think that was the first thing Graham did. To my suprise it made a difference, hence Graham's experimentation with this process.

ok,

say you had a 100 L batch of wvo that you were going to process and this batch titrated at a 1.

This would mean you would add approximately 6.5 g/L or 650 grams of NaOH.

The molecular weight of NaOH is roughly 40 grams/mol.

The molecular weight of HCL is roughly 36 grams/mol. Also, there is roughly 354 grams of HCL per liter (@ a 30% concentration)

So, let's figure this out. Lets say that the titration amount is used up making soaps. So that leaves us with 5.5 grams / liter (550 grams all together) of NaOH to neutralize.

therefore, 36/40 = 0.9 therefore, to figure out how many grams of 30% HCL we will need, 550 * 0.9 = 495 grams of HCL, which is 495/354 = 1.4 liters of 30% concentration HCL.

How does that jive with what graham has been adding? Is it even close?
Perhaps, there is more NaOH that is used up making soap than just the excess added to neutralize the FFA's?

anyone have any thoughts? (or see any mistakes that i've made?)

quote:

Originally posted by Dogma721:
. . . . Maybe . . . then the reversal would be negligible enough such that it would not effect the 3/27 test.

The 3/27 test is not likely to show a reversal.

The only product that "drops out" in the 3/27 test is triglycerides as they are insoluble in methanol.

The transesterification reaction proceeds, step-wise, from TG to di-glycerides to mono-glycerides to methyl esters.

A reverse reaction proceeds, step-wise, from ME to MG to DG to TG. The excess amounts of MG and DG formed during a reverse reaction will be more than enough to cause the sample to fail a GC test for total glycerol.

But, being mostly soluble in in methanol, the reverse reaction products are unlikely to cause a 3/27 test failure.

quote:

Perhaps, there is more NaOH that is used up making soap than just the excess added to neutralize the FFA's?

That is the reason Graham titrates. In practice it is not so simple as assuming you need to neutralise the base amount exactly. It seems to use less more often than not. If I knew the results Graham was getting eg no more than 5%, 10%, 20% or whatever, less than expected, I would titrate a sample and make it up to full base amount by adding caustic before applying the 10% (or whatever) less than ideal amount of acid to see if the uneutralised caustic was enough to cause a reverse reaction using the methods Graham describes to check if it was happening. This would simulate a worse case scenario assuming you titrated and measured correctly when you were making the bio in the first place.

Better yet Graham could do it as he is all set up already lol.

If no significant reverse reaction took place then we would have a far simpler method of achieving the same goal.

I'm not convinced that a "reverse" titration on a sample wouldn't get us "close enough". Much depends on how much room for error we have. If indeed we only need to get below a particular threshold, then there should be plenty of allowance.

My only test was with a beaker-sized sample, and I was amazed how little acid was required to drop below the phenolpthalein indicator limit (about 8.2). All I would have needed to do was to continue adding acid until I reached some neutral value (say, 7.0) to get a sense of how much wiggle room we're talking about. Since I'm not planning to make another batch soon, perhaps someone else would volunteer to try it?

Cheers,
JohnO

Wow, what a technical thread!

Sorry to be the ignorant one, but is the push pull system taking the place of the GL 1 day? It seems to be so. Is the GL 1 Day easily adapted to be a push pull system? It seems not to be so....

Thanks

Tof

Hi folks

For a batch made with 100 litres of oil, I use about 100ml of 34% HCl to neutralise.

I use NaOH, 5g + titration, typically 7g/litre.

I tried titrating for the HCl but maybe my pump is too weak, I never got consistent results and I suspect it is due to glycerol dropout continuall causing the lower region to be rich in glycero/catalyst and the upper region weak in it.

with a beefy pump, I'm sure you could reliably titrate.

Don't forget though, we only want to neutralize the catalyst, we don't want to convert the soap to FFA, so a little less acid is probably better than an overdose.

Somewhere on this forum I went thru all the quantities I use, but I just can't find it now. It was to do with minimising methanol usage in the reaction ,and Andrew Morris was one of the main contributors to that thread.

I think we need to figure out the relationship between temperature, catalyst load and the rate of reverse reaction.

I agree with what Graham just said - obviously we are not helping ourselves out if we start converting soap to FFA's - which will probably happen to some extent.

Perhaps if we drop temperature low enough and apply a deep enough vacuum we can eliminate this step all together?

Graham, do you know how deep of a vacuum that you draw with your setup? If we used a commercial eductor/air ejector with a beefier pump I wonder how much effect that would have on the temperature required to distill the methanol? I'm sure that since we know the rule of thumb of halving reaction time for every 10C increase in temperature, i wonder if the reverse reaction would follow a similar progression?


At 30C the vapor pressure of methanol is 163.97 mmHg. I'm not sure what that corresponds to in terms of inches of vacuum, but, its about .2 ish atmospheres, which is 2.5 psia, so, its a pretty good vacuum but not entirely unachievable and i believe the commercial eductors with a sufficiently sized pump should be able to pull that. At 30C, the reverse reaction rate out to be pretty slow. Also bear in mind that due to partial pressures, the temp will have to be a bit higher to be able to distill the methanol.

Graham, i think you mentioned this before but what were your temps during distillation under vacuum? and do you have any data on how deep of a vacuum you were able to draw?

Hi Dogma,

I draw a very poor vacuum, about -7psi
but this seems good enough.

Your idea of a lower pressure, with accompanying lower temps is really good.

The trick when running low pressure is to limit the rate of heat input, otherwise the vacuum pump can't keep pace with the vapour output, pressure rises, temperature rises and it all foams up.

A simple way to limit the rate of temperature rise is to use a PID controller with ramp feature.

Draw as hard a vacuum as you can, and then get the PID controller to raise your temperature very gradually, say 1/2 a degree per minute.

This will help to ensure you don't foam, and will keep the pressure and temperature under control.

hopefully this should help...

i'm also interested in the bp of water under vacuum and in solution so i can refine the methanol recovery into two steps rather than deal with distillation.

quote:

csp97
Member

Posted 16 October 2006 01:38 PM Hide Post
Revised for 26" Hg vacuum (2 psia) using Antoine.

Biodiesel / Methanol Solution
% MeOH--- Boiling Point, F
5.00 ---------- 114
4.00 ---------- 121
3.00 ---------- 130
2.00 ---------- 146
1.00 ---------- 176
0.50 ---------- 211
0.25 ---------- 252
0.00 ---------- 467

Glycerol / Methanol Solution
% MeOH --- Boiling Point, F
45.00 --------- 88
40.00 --------- 91
35.00 --------- 95
30.00 --------- 99
25.00 --------- 105
20.00 --------- 113
15.00 --------- 123
10.00 --------- 139
5.00 ---------- 169
1.00 ---------- 259
0.00 ---------- 436

Location: Texas | Registered: 29 July 2006

quote:

For a batch made with 100 litres of oil, I use about 100ml of 34% HCl to neutralise.

I use NaOH, 5g + titration, typically 7g/litre.

Is this the least amount of acid you ever need with those amounts of catalyst?

quote:

Don't forget though, we only want to neutralize the catalyst, we don't want to convert the soap to FFA, so a little less acid is probably better than an overdose.

Exactly so if we try using 80mls of 34% acid per 100litres of oil then we should be safely neutralising the bulk of the catalyst without forming FFAs. It would still be important to add the acid slowly whilst circulating rapidly to avoid local acid concentrations from forming FFAs.

With 80% of the catalyst removed would we expect much of a bio reaction? So why would we get much of a reverse reaction? Especially using vacuum and low temps to slow possible reactions down and reduce the risk.

it could be a simple way to make good bio with complete one shot meth recovery using a standard GL reactor.

If 80 % doesent work well enough we could push to 90% and still be safely under the making FFA limit.

I find it hard to believe there would be much or any reaction with little catalyst, low temps and the fastest time we can manage.

quote:

obviously we are not helping ourselves out if we start converting soap to FFA's - which will probably happen to some extent.

I would hope making FFAs can be avoided.