DIY Watermaker

[s/v Jedi]

Originally Posted by foggysail

So all I had to do was relearn C after 20+ years, in this case Microchips C18 software language, become familiar with one of their microcontrollers, a PIC18F4550 (about 260 pages) plus learning one of the newer serial communications buses I2C.

Huh... so I am not the only cruiser with micocontroller programmers and a stock of PICs aboard? Something's wrong haha Well I even have a BASIC Stamp aboard and normally use assembly to program them. They are more powerful than the CPU's I used to program (6802, 6510, Z80 etc)

I2C new? that's funny, you must be older than me

cheers,
Nick.

[goboatingnow]

Originally Posted by s/v Jedi

Huh... so I am not the only cruiser with micocontroller programmers and a stock of PICs aboard? Something's wrong haha Well I even have a BASIC Stamp aboard and normally use assembly to program them. They are more powerful than the CPU's I used to program (6802, 6510, Z80 etc)

I2C new? that's funny, you must be older than me

cheers,
Nick.

All the cool kids have gone over to arm cortex now ....!

Dave

[foggysail]

Originally Posted by goboatingnow

Agh well foggy that's what you getting listening to me. !!!

My experience ( I've since sold it) is that building your own can be cheaper but isn't necessarily so, especially compared to the cheaper end of the non-energy recovery professional units.

In my case I did a few things differently. Firstly I'm convinced that brass pressure pumps are adequate, subject to a few provisos. One they will not last anything as long as SS or titanium ( mind you I,m dubious about SS at the best of times ) but they are cheap. Secondly be aware that the oil seal can fail and contaminate the water that can be a problem on cheap heads. So learn to dismantle and inspect and replace. Thirdly I have a programmable unit ( mines uses a atmel controller) so I rigged it to always fresh flush the pump after every use, never leave seawater stagnant in the pump.

I used a needle valve and motorised it with a simple motor gearbox unit. That's was still cheaper then a cat back pressure regulator and has the advantage that the system can be adjusted remotely. It's very handy to not have to bring high pressure piping any distance to allow access to the pressure regulator.( as foggy pointed out this also applies to remotely sensing pressure and flow. Also my unit offers the possibility to balance pressure against product flow to compensate for salinity and water temperature. ( I didn't do that yet that's for version 2) I have a purely electrical panel that uses a 3.5" windows CE panel PC to display and control the watermaker and this communicates to the micro controller. As I said this keeps the high pressure runs very short. From inspection I'd say I'll would have gotten 4years from the pump. This assumes good sediment filtering etc.

I built mine because I like building things. If you do, great if not, its not a good way to save money. The small things like reliable high pressure fittings and hoses etc can get you.

Danfoss are now beginning to ship small volume rotary energy recovery units and hopefully as prices fall this will let us DIY with them as well as spurring more competition to the Clark pump method, which while useful and cheap is a poor example of good energy recovery.

Dave

Hey Dave--

Thanks for coming back!!! And yes, keeping one's control panel free from plumbing is the way to go!!! I have the CAT 7070 and I did think about a stepping motor instead. What stopped me was the mechanical interface between the motor and the needle valve shaft. I also have a SS needle Parker needle valve (4A-V4LN-SS).

Brass pump! That is what I am using. I also have an automatic flush cycle programmed.

Any chance of sharing info on your stepping motor and mechanical interface? I WOULD LOVE TO RID that CAT from my control panel! My plumbing would be soooo much easier without the risk of HP water spraying all over the place.

OH--- I picked up a couple of Newhaven serial display modules. The displays I have now have 4 alphanumeric lines, 20 digits/line and of course, I2C. They are inexpensive, less than $25 each.

Foggy

[s/v Jedi]

Oh.. you should pop a wifi module in and create an iPhone app for control

ciao!
Nick.

[foggysail]

Originally Posted by s/v Jedi

Huh... so I am not the only cruiser with micocontroller programmers and a stock of PICs aboard? Something's wrong haha Well I even have a BASIC Stamp aboard and normally use assembly to program them. They are more powerful than the CPU's I used to program (6802, 6510, Z80 etc)

I2C new? that's funny, you must be older than me

cheers,
Nick.

Yeah Nick, I'm older than dirt! Yeah, Philips originated the I2C way back in the late 80's but it was still new to me. And assembly????? You must really like pain!!!

Way back in the days of yesterday (like the Lone Ranger's days) I was an engineering manager. I almost broke some programmer's knuckles once when he wanted to use that stuff in one of the designs. It was lucky if I could get 10 lines of usable code/day out of C from them.

Really nice to see there are folks here who have interest in this stuff!

Foggy

[foggysail]

Originally Posted by s/v Jedi

Oh.. you should pop a wifi module in and create an iPhone app for control

ciao!
Nick.

Now Nick! I am in my 2nd year with this thing!!! Got to pass water first!

[Tellie]

I swear if you guys don't post some pics....
I did enjoy Thomas blog on his build. I remember when I built DIY watermakers. It's not for everybody to be sure, especially anyone who dares value his time. But it is rewarding when you finally take those first few sips of fresh water from a device you created.

[foggysail]

Originally Posted by Tellie

I swear if you guys don't post some pics....
I did enjoy Thomas blog on his build. I remember when I built DIY watermakers. It's not for everybody to be sure, especially anyone who dares value his time. But it is rewarding when you finally take those first few sips of fresh water from a device you created.

Tellie---

The only pictures I could share at this point is that of a microcontroller development board along with a digital display attached. When I get to the point where I test the software with actual solenoids and simulation, I will start with some pictures. At that point there will at least be components that are recognizable.

Remember a year or so ago when I posted something similar to "Anybody who builds a watermaker to save money is probably building it for the wrong reason?"

Foggy

[goboatingnow]

Foggy, I'm appalling at documenting things I build , ie photos etc. but in my case I used a dc motor gearbox from RS components, I then used a shaft coupling ( it was in my odds and ends box, but I think it was from lenze uk) to connect the needle valve shaft to the gearbox output shaft ( after drilling out for correct shaft dia). There then was a bit of a metal lash up to hold the motor and to prevent it turning ( basically a tie bar to the valve body. It might be easier in retrospect just to mount the whole lot in a cheap aluminium die cast box. I relied on the valve stem to support the little motor/gearbox. Since it all turns slowly, it wasn't a problem.

Oh there's no limit switches. The control circuit detects stall current at the end of the shaft travel and in operation the pressure sensor effectively acts as a feedback sensor. ( I effectively monitor motor current on a ADC line on the controller. )

I added a high pressure pressure sensor, an input paddle wheel flow meter and product flow meter and tds sensor ( actually a simple conductance probe). This gave me an all electronic panel ( of course the required divert and flush solenoid valves. The only fitting in the high pressure lines are the pressure sensor and the needle valve. There is a standard relief valve built into the pump that's acts as a over pressure protection.

The input flow meter also acts as a warning of filter sediment buildup in the input filters. The micro controller also controls motor startup and feed pump startup as well. It communicates by RS422 to the little windows CE display device. My intention was to have graphs etc. but ..... Version 2 etc

The motorised needle valve also let's me do slow startup and low pressure flush operations very nicely. ( as I run the pressure pump during flushing to ensure seawater is ejected.

Ps, if you keep listening to me you'll never finish it.

Biggest problem I had was bloody metric to npt conversions. It took me time but i sourced the hose fittings swaged with npt, the pump was bsp so i needed a convertor. So much RO stuff is US sourced. There must be a good European supplier but I never found one. ( ie for pressure vessels). I was tempted to have one welded out of stainless and pressure tested, I met my local agricultural stainless pipe man recently a d he says it didn't look difficult and the pressures were not a problem. ( welded one end , threaded o ring the other. ) , but it was a chat over a pint type of thing.

Dave

[s/v Jedi]

Originally Posted by foggysail

Yeah Nick, I'm older than dirt! Yeah, Philips originated the I2C way back in the late 80's but it was still new to me. And assembly????? You must really like pain!!!

Pain? After opcodes and operands with hex pad programming, assembly is luxury!

This is what opened this binary world for me while in school:


Amazing that these pictures are on the web

ciao!
Nick.

[foggysail]

Thanks Dave---

Like the microcontroller idea---- a stepmotor driving a needle valve IS THE WAY TO GO!!! I really cringe at the thought of routing a high pressure line of any type from the pump assembly to my control panel. Horror of horrors!!! I can just visualize it now.... raw salt water alll over the place at over 3 gallons/minute!! No way!

The important thing in your above post for me is that related to the stepper. I have a draw filled with flow solenoids both low and high pressure plus flow transducers and one high pressure transducer. For fittings--- I already have numerous Swegeloc stuff in my desk drawer. Probably will require more. They are readily available on Ebay.

And yes--- you have excellent advice so I do pay attention. I would rather sacrifice a slight additional implementation problem if it accomplishes getting rid of high pressure water from my control panel.

Thanks again--

Foggy

[foggysail]

Nick--

I remember an into to push, pop, stack pain with Microsofts 8 bit stuff. I will avoid it if at all possible. See---Cs for me. Maybe if I was doing this for a living I could become motivated. But I'm doing this for fun......unless you listen to Wifey! And hex programming???? Is that not a little like stone age? You DO like pain! : )

Foggy

[goboatingnow]

I didn't use a stepper just a simple permanent magnet dc motor with integrated gearbox, cheaper and easier to control.

Yes micro control is the way to go, for example I have a routine in software that detects sudden loss of high pressure, virtually instant shutdown. This prevents high pressure leaks from spewing water every where.

Yes you seem to have the plumbing sorted better then me. Of course next time round I'll know exactly what to get. !! hint sell your first one then build the real one. !!

Dave

[foggysail]

A plain DC motor! Now that is interesting. Sell the first one??? I might just do that. Second time around would be a piece of cake.

Foggy

[Sam Plan B]

I have been wondering lately about an alternative to a conventional WM. It seems to me that forcing sea water through a semipermiable membrane takes an awful lot of energy. Besides, there is the problem of filtration to remove sediment when at anchor. However, compressing gas does not require so much force. So, since we are always in an atmosphere of 100% relitave humidity, can a WM be made that produces condensate on a coil. The condensate would be harvested and piped into the boat's water tank. It would always be pure water without the problems of sediment filtration.

Also, years ago there were air conditioners on the market that worked on amonia. I believe Arcal was one brand. The amonia was in a closed loop. When it expanded it absorbed heat. When the amonia was fully expanded it would condense to a liquid. The liquid amonia would be pumped throug a tubing section that was exposed to a gas flame. The amonia would expand again and repeat the circuit. So, can a WM be made that operates with a propane flame and a 12v circulating pump? Just wondering.