OK ... Think it's time to come out of the shadows and divulge one of the much speculated "secret" methods I've tried.
This is Paul's (Carrington's) idea so I take no credit for it's merit or otherwise.
I have to say at the outset that the process involves operations I consider potentially dangerous, and I'd recommend that it is not tried unless you are careful, knowledgeable of chemicals (especially acids) and use the required personal protection equipment.
I'm not putting myself forward as particularly careful and certainly not knowledgeable about chemistry, but I did seek third party advise from a friend who has a PhD in chemistry and survived to tell the tail.
The method is very similar to Paul's titrated acid wash method, except it uses no water. Paul believes that there is a chance this method may prevent the formation of HMPEs.
The sequence is as follows ...
Process normally to a 3/27 pass, by whatever method you choose.
Allow Glycerol at least 8 hours to settle then drain.
(By draining the lower pipe work I managed to settle over 24hrs, draining several times)
Titrate to quantify acid requirement ... probably a bit of a mix with the pump before hand will collect glycerine off the processor walls and give a more representative result. I believe a concentrated acid is required to eliminate as much water as possible.
Add acid to Methanol, the acid being circa 20% of Methanol quantity.
(This is the dangerous bit! I used roughly double the Methanol quantity, believing it to be slightly less hazardous. The reaction is highly exothermic and the acid should be added very slowly. I did so using a pipette, but even then an audible fizzing and gluping was noticed if more than a few drops were added at one time.
Add "Methacid" to processor and mix, say 20 mins
(My plan was to introduce the "Methacid" via the venturi, but again I had concerns about this. There were pockets where the "Methacid" could come into contact with copper and brass and not have a chance to get washed away by the bio. So I decanted bio from the sample point and added small quantities of the "Methacid" to the bio and entrained that into the processor via the venturi and all went well)
Demeth the bio.
( I usually WBD, but as the bio had cooled to around 20°C, the glycerin had been removed and there's the school of thought that WBD can promote HMPEs, I decided to bubble the Methanol out. So I added Coldflow, mixed and pumped out.)
Pump to settling tank.
(I bubbled for 24 hrs to remove Methanol)
Settle for X days as you see fit.
(After a week the bio was still cloudy, but gave the best 50/50 test I've ever seen. A test with litmus paper showed a pH of around 7. To remove the cloudiness I returned the bio to the processor and distilled out about 2 litres of Methanol ... unless for prolonged periods, bubbling is obviously only good at removing residual Methanol)
I've still to get a controller for the redundant freezer and run chilling tests, but I'll stick a sample in the kitchen freezer and report results.
I have to say that, having tried it, unless it returns startling results in the way of HMPE reduction, I'll not be trying the process again ... I found it rather scary! That said, many thanks to Paul for hitting on another variation of the process, these things have to be tried, let's hope it has some effect on HMPEs ... watch this space!