reverse CM

[manu de hanoi]

basically a VM with a siphon.
The amount of cooling would control the amount of output (and not the amount of reflux as in regular CM).

untitled.GIF (6.93 KiB) Viewed 3103 times

++ No uncondensed vapor escapes from the product condenser
++ the coolant use in the condenser is optimised all the time
++ There is a good chance that the VM valve be useless because of the suction strength created by the condenser. without valve this design would be cheaper than a VM

-- if too much coolant some air may be sucked in from the top reflux condenser, so a safety distance is required between the arm and the reflux condenser. A purge valve may be used during the run to remove air from the condenser (coolant is stopped so that vapor expels air from the condenser, same shall be done starting the device )




That would perhaps be a good contestant for the VM in the most convenient design. who wants to test ?

[LWTCS]

I like the trap concept at the discharge. How will this set up cope with a huffing condenser do you think?

[kiwistiller]

interesting. for one thing the valve would still be needed for getting the column in equilibrium, right?

[manu de hanoi]

I like the trap concept at the discharge. How will this set up cope with a huffing condenser do you think?

No huffing possible

[manu de hanoi]

interesting. for one thing the valve would still be needed for getting the column in equilibrium, right?

good point but if the coolant is turned off and the siphon filled, it will get the column to equilibrium without a valve.

[LWTCS]

No huffing possible

How so manu?

[manu de hanoi]

No huffing possible

How so manu?

because the only thing that draws vapor to the product condenser is the void created by the condensation of vapor that was previously there. There is no vapor "push" from the column because the output of the condenser is blocked by the siphon.

In short, the product condenser will pull as much vapor as it can condense.
The only huffing possible is from the reflux coil.

[HookLine]

Very interesting idea, Manu.

To control the take off rate via the product condenser I think you will need to have very fine control of the condensing rate (coolant flow) in the product condenser. This is the critical design issue, as far as I can see.

A standard size Liebig may not work so well. It has a fair bit of thermal mass and at low take-off rates may take too long to respond to small changes in the coolant flow rate. A mini-Liebig about 150 mm long, or a small coil condenser or Graham condenser, might work better as they have less thermal mass.

Reflux and product condensers will need separate coolant flow control valves.


If you can control this still properly with the product condenser then you would not need the VM valve, and that means you can safely seal the top of the column above the reflux condenser. If so then you can run this still in two, and maybe three modes.

1. Pot still mode, where coolant flow to the reflux condenser is completely turned off, so that all vapour is condensed in the product condenser. Output rate is controlled by power in. This mode does not need the liquid trap at the bottom of the product condenser.

2. Reflux mode, where the reflux condenser is condensing 100% of the vapour that hits it, and output rate/reflux ratio is controlled by the product condenser. This mode needs the liquid trap at the bottom of the product condenser.

3. Partial reflux mode, where the reflux condenser is condensing only part of the vapour that hits it (achieved by carefully controlling coolant flow to the condenser). The excess uncondensed vapour must go somewhere, so it travels down the product arm and condenser. I suspect that this mode is very sensitive to coolant flow rate in the reflux condenser, and maybe also to the interaction between the two condensers (their condensing rates). It may be difficult to get consistent results, and may require some trial and error. I think this mode also needs the liquid trap at the bottom of the product condenser, but not 100% sure. I recall Riku made some comments on partial reflux in a VM still before. Ask him.

Really looking forward to somebody trying this one.

[manu de hanoi]

If you can control this still properly with the product condenser then you would not need the VM valve, and that means you can safely seal the top of the column above the reflux condenser.

very interesting comment. But if the top is closed, then that would be running a vacuum still. And if the coolant flow is too strong then the liquid trap will be sucked in the still, if the coolant is too little the trap will be blown out. That would be hell to control.

[olddog]

Have you tried this build Manu? it looks like it would be a simple build to try.


OD

[manu de hanoi]

Have you tried this build Manu? it looks like it would be a simple build to try.


OD

the condenser is built, but I am working on a new type of packing so the test has to wait:

[HookLine]

If you can control this still properly with the product condenser then you would not need the VM valve, and that means you can safely seal the top of the column above the reflux condenser.

very interesting comment. But if the top is closed, then that would be running a vacuum still. And if the coolant flow is too strong then the liquid trap will be sucked in the still, if the coolant is too little the trap will be blown out. That would be hell to control.

Fair point. Closing the top of the column may only work without the liquid trap.

[LWTCS]

Fair point. Closing the top of the column may only work without the liquid trap.

It will work without the trap. How well will it work?

Come outta the top of the column and that condenser will behave like a dephlag.
takeoff acts as the vent and with adquate cooling the thing can be put into 100% reflux.

ABV ??? I dunno. Column height should take care of that.

[manu de hanoi]

If you can control this still properly with the product condenser then you would not need the VM valve, and that means you can safely seal the top of the column above the reflux condenser.

very interesting comment. But if the top is closed, then that would be running a vacuum still. And if the coolant flow is too strong then the liquid trap will be sucked in the still, if the coolant is too little the trap will be blown out. That would be hell to control.

Fair point. Closing the top of the column may only work without the liquid trap.

Doing so would result in something very similar to a normal CM

[Master-Peter]

From my very theoretical point of view it might work, if a real constant water pressure is given and the cooling flow is absolutely fine adjustable (expensive needle valve). But if it works it makes column building far mor easy. So really worth a try.

For me probably no option, for I have a changing water pressure between 2 and 3 bar...

[manu de hanoi]

I just finished a test run, It works !!!!!

There is just a trick: the vapor wont fill the condenser at the beginning because the path to the reflux is way more attractive than my tiny syphon. So you have to "prime" the pump by either :
- turning off the reflux collant and block the top of the still so that vapor is forced to expell the air in the liebig or
- plug a hose siphon with a manual pump (the aquarium thing) to the output of the liebig and squeeze 2 or three time remove air and to "prime" the condenser

No video for now, i forgot to charge the batteries. Here is a pic of the setup :

[kiwistiller]

Well done mate, good work. looking forward to performance feedback

[HookLine]

Nice one, Manu.

Also look forward to more detailed performance results.

[manu de hanoi]

Nice one, Manu.

Also look forward to more detailed performance results.

like what ? the only thing i can show is a video of the thing working.

[HookLine]

Nice one, Manu.

Also look forward to more detailed performance results.

like what ? the only thing i can show is a video of the thing working.

Like how well it concentrates heads, how stable it is, output rate compared to LM or VM, etc....

Only if you want to do it, dude. Wasn't an order. Just curious.

[manu de hanoi]

Like how well it concentrates heads,
- that is not related
how stable it is,
- on the video
output rate compared to LM or VM,
- max output rate doesnt change btween cm/lm/vm it's just how you control it.

[kiwistiller]

how about ease of control?

[manu de hanoi]

it was controllable. prolly not the best on planet earth because I was running it on my water cooling loop fed by a too poweful pump. The water need to have a low pressure so that the coolant flow is a more or less linear function of the valve opening.