July 23rd, 2008
pic -> plates after welding bolts on
pic -> side view of perspex with plates inserted
pic -> top view of perspex with plates inserted
July 16th, 2008
Bit the bullet and re-hired some oxy and acetylene cylinders to enable the plate connection bolts to be fabricated. Have completed the first plate, only 15 more to go....
pic -> bolt - before and after beating
pic -> finished plate with connection bolt
pic -> finished plate with connection bolt (side view)
July 3rd, 2008
Mounted Smart Mixture Display unit into control panel today.
pic -> back of top panel (bit messy)
July 2nd, 2008
Finally had time to make some headway with the electrolyzer control panel (for inside vehicle cabin). Not sure until I connect it all up in the vehicle if the multi-core wire I used (about 3 metres) will be suitable for extending the thermocouple to the cell. At this point the temperature seems to measure correctly over that distance, same with the voltmeter and the ammeter. The 5v power supply on the temperature kit supplies enough to run the 3 LCD screens. Have to look at squeezing the AFR meter and Digital Fuel Adjuster units into the same panel as well.
Have to now look at knocking up a decent LPG torch so I can continue with re-jigging the electrolyzer terminals. Got rid of the oxy-acetylene cylinders ages ago - rental was a killer.
June 27th, 2008
Have machined up see through thicker tops for replacement cells. Used 3/4" perspex to give a decent top through which to bolt the connections (found the thin PVC was a bit too thin to support hot connections).
pic ->
standard PVC sleeve and end cap
pic -> standard PVC sleeve and end cap
pic ->
machined tops and perspex insert with 'O' ring seal
pic -> machined tops and perspex insert with 'O' ring seal
pic -> machined tops and perspex insert with 'O' ring seal
June 16th, 2008
Made a bit of headway with a few things lately.
Multi-timer/Lawton board: designed a circuit based on a 556 dual timer chip (rather than using two 555's. Got circuit drawn up, pcb done and populated. Board can be configured for use as timer + pwm, two independent timers OR a Lawton circuit. Tested fine on breadboard but as yet haven't been able to test built up pcb.
pic -> timer lawton layout pic -> timer lawton components pic -> timer lawton pcb pic -> timer lawton top view (bit blurry)
Current-limiting pwm: Gary (chemelec) has laid out a pcb for his modified version of ZeroFossilFuel's current limiting pwm circuit. Completed my first pcb from this today and have populated it - tomorrow, connection and testing. Used press-n-peel pcb film to do this latest pcb, wasted hours trying to find the right iron temperature but think I've now finally cracked it. Population of the board was done mainly from bits and pieces lying around (hence all pots don't face the right way, no big deal), but I had to make a special trip to find the big zener. No component overlay made it a bit hard for me, but I finally traced where all the parts go (I think).
pic -> current limiting pwm pcb
pic -> current limiting pwm pcb populated (50 amp shunt next to it)
pic -> current limiting pwm - press-n-peel
April 20th, 2008
School holidays taking a toll on time spent on booster related issues (looking after grandkids), but even so I started build/assembly of the twin booster unit for install in a Mitsubishi V6 3 litre 4X4 (yes I found enough space to do it), regardless of whether I can get the ODBII stuff working. The electronics units are all built except for the finishing touches on the high-temp digital temp kit (needs an insulated temp probe to connect to exhaust to ensure not running lean, one for unit, one for exhaust).
Since the 6-7 cell switch worked so well, I've elected to use a 2 X 8 cell config in each unit as I've managed to squeeze the extra plates in and still fit inside a 100mm PVC pipe (this gives 6/7/8 cells each side). Having both units running at the best level (just below the steam spewing point at 15 amps each), I should easily be able to consistently produce between 2.2 and 3 lpm between the two at 30 amps (this is what the bench testing has produced).
As far as switching between cells goes, I'll do that manually to start with as I'll be monitoring volts, amps, unit-temp, exhaust-temp and air/fuel ratio. Once I have more data on the most beneficial switching criteria, I'll look at some sort of auto management, which also could shut the unit down (or throttle back) in overload (amps), overheat or excess lean conditions.
April 7th, 2008
Received Mitsubishi ODBII connection cable but as yet, the software does not initialise properly. New s/w version expected shortly which will hopefully rectify the problem. All kits have been built and tested....waiting for ODBII connection for comparison and tests for baseline measurements.
In the last few days, I've tried pulsing many different types of inductors connected to the bench test cell. Although small HV spikes are noticed on the scope, nothing large enough to increase gas production at all when running 6-7 amps through the unit. Pulsing (and inductors) just limit the maximum amount of amps flowing through the cell - and hence the amount of gas produced. I guess an inductor may be used to limit amps as one way of stopping amp runaway with heat?. Even though I was hoping for a magic moment, there was none!!! I agree with Gary/Chemelec on this point now, pulsing in it own right does not increase gas production AT ALL. As far as inductors go, you'd need a really large one to have any effect at all - hence the humungous one that Boyce reportedly uses to gain an advantage (yet to be proven as far as I can make out).
Have ordered another PW3F board from Hydrogen Garage to provide drive for the 100 cell unit I have ready. While waiting for delivery, I'll get onto finishing the primaries on my Boyce toroid, been putting that one off for a while.
March 28th, 2008
Have purchased a few Jaycar (Silicon Chip) electronic kits to assist in getting an EFI vehicle running properly with a hydroxy booster . The Hand Controller, Digital Fuel Adjuster, Programmable Ignition System and the Air/Fuel ratio display.
Both Hand Controller and Digital Fuel Adjuster both worked first time (yay!) pic -> DFA and Hand Controller after setup/test. Now all to do is build the Air/Fuel ratio display then I can look at fitting out a test vehicle.
March 18th, 2008
Did some quick condensation tests on bench test unit output gas today. Used a large coil of about 15ft of 1/2" copper tubing in a container of cold water to see if I could condense any of the vapour (steam) coming from the unit.
The last bubbler test results may be due to slight gas leaks in the bubbler (to be verified) or actually may be that some of the vapour did dissolve in the bubbler water - will have to test this again and test the bubbler unit for leaks
(later) Tested the bubbler - no leaks. The outlet inside the bubbler had 6 x 1/4" holes in it for the gas to bubble from. Have now replaced this as the bubbles inside the bubbler were still huge. Now the outlet has 4 x aquarium airstones branching from it. All of these are not being pushed at their maximum from what I can see, but now there are a multitude of smaller bubbles instead. Did another output test straight from the unit and then through the modified bubbler. The output gas did not seem to change in volume at all, but now there is a marked reduction in mist emanating from the final outlet tube. The water inside the bubbler now gets caustic as the NaOH mist dissolves on the way through (can tell by the soapy feel of the water between my fingers - telltale sign of caustic).
March 14th, 2008
Tested cold start with series perspex bench test unit today. Wired as 7/2 series/parallel to start with, 13.8V only drew 5A of current and produced 0.425 LPM, temp was 28.5C (cell and ambient to start with, in the middle of a heat wave at the moment, this is as cold as it gets in the morning). Switched wiring back to 6/2 series/parallel setup, then 13.8V drew 20+amps and immediately upped production to 1.18 LPM.
Once the unit temp warmed up a few degrees to 32C and the cell got going properly, current draw gradually rose up to and above 25A. At that point wiring was swapped back to 7/2 config where current dropped back to 20A at 13.8V, output was 1.02 LPM
I think auto switching of plate config would be better off based on current draw rather than temperature unless some form of current limiting is introduced (maybe pulsed DC with feedback circuit). At that point a combination of current limiting with temp control would be the best mix.
March 13th, 2008
Over the last few days have tested various configurations of the bench test cell. The SAME plates and container were used for all tests.
pic -> sealed unit with tool dip and gladwrap covering (original series layout before perspex covering added - tool dip failed at higher temps)
pic -> step gap test (a la Smack booster) (with new container cap, 6mm spacer nuts and 2mm nylon washers)
pic -> sealed unit with equalisation holes (drilled at staggered 1" gaps so min distance between any two holes is 1")
pic -> sealed unit with equalisation holes filled back in
Results for different tests:
| Configuration | W/LPH | Perc of expected Faraday output | Voltage | Amperage | Amps/SqIn | Litres/Min |
| Step-gap config (2P4S) Smack-type | 6.06 | 45.23% | 13.8 | 20 | 0.667 | 0.76 |
| Series perspex cell with 1mm holes | 4.99 | 54.45% | 13.8 | 20 | 0.667 | 0.92 |
| Series Gladwrapped cell | 2.92 | 99.68% | 13.8 | 4 | 0.133 | 0.32 |
| Series perspex cell holes blocked | 2.88 | 96.49% | 13.8 | 11 | 0.367 | 0.88 |
| Series perspex cell holes blocked | 2.88 | 89.03% | 13.8 | 15 | 0.500 | 1.11 |
The step-gap design gained heat very very rapidly. In fact it only took 10 minutes for the cell to heat up to 45C and start spewing heaps of steam/vapour. In contrast, the latest series perspex test at 15 amps, only gained 2 degrees during the 1/2 hour run time (on outer connections with two lots of 7 series cells in parallel internally).
Last series test at 15A (7.5A/plate), started spewing steam/vapour similar to the step-gap design. With the small plate size of 6" X 2.5" (15 sq ins), amps should be restricted to a max between 10-12 amps overall (0.375 A / sq in). With two duplicate units run in parallel, the overall amperage draw could vary between 20 and 30 amps, with output ranging between 1.7 LPM and 2.2 LPM. Since I have heaps of these offcut plates lying around, that might be worthwhile looking into for a short term solution to get something into a vehicle and running.
March 8th, 2008
Machined some 1/4" perspex panels for the sides and end of test unit. Fitted, glued, cured and now testing again.
pic -> upgraded unit with new perspex panels
March 5th, 2008
Pulled finger out and finally got around to machining an adaptor to mount a burnt-out angle grinder head to the vertical spindle of my 3-in-1 bridgemill, to allow better control when cutting slots (angle grinder used to be clamped to vertical spIndle). Pretty chuffed with the final result.
pic -> burnt out angle grinder
pic -> grinder gearbox ready to mount
pic -> adaptor mounted to vertical spindle
pic -> cutting slots longitudinally
March 4th, 2008 (later)
Unit has been performing well all day up until the time it got up to 45C (ambient temp was 39C at the time - hot day today). After that I noticed a progressive lowering of output? Tonight after powering down, I disassembled the unit and found the side clingwrap/tool-dip panels had shrunk somewhat, leaving gaps at the edges of the plates - hence the voltage and output loss. The tool-dip (only) layer on the connecting tabs and the back of the two outermost panels was still intact, only the layers including shrinkwrap had changed.
Tomorrow I will go ahead with what I was originally going to do and slot two perspex side panels and an end panel for the plate pack, and just hold them against the plate pack with a couple of s/s hose clamps (or something like that?). Could glue all three perspex panels together and just slip them up and over the plate pack, then fix at the top somehow I suppose, will see what happens.
March 4th, 2008
Have now used 15 plates of set in a new lyzer using a 4" PVC casing and screw-on lid (much less leakage). TIG welded some mounting/connecting tabs onto centre and outer pair of plates on each set (to give 2 x 6 or 2 x 7 cells). Quickly tried it as an open bath series/parallel before continuing any further (similar to layout of Smack booster without step-gaps) - too much current leakage, drew 4 amps with only 2.5V using outer plate set (2x7 cells), gas released from centre and outer plates only - no visible activity from inner floaters at all (just wanted to see what would happen).
Next, I laid the assembled plate unit one side at a time, onto a layer of clingwrap with tool dip plastic painted onto it. When tool dip dried, did same to two other sides. When finished, covered the whole outer surface (over clingwrap) with a final sealing layer of tool dip. When dried, tested in a clear container with KOH mix and found NO leaks into the plate gaps at all (didn't fill up), even after poking a hole in each cell side with a needle (holes sealed themselves). To retain a level balance hole in each cell, I think I'll have to heat the needle to melt the plastic sandwich so hole does not close back over.
After tipping unit sideways so the cell gaps filled, connected it back up and got 0.315 lpm for 13.8V @ 4 amps initially. Putting these figures through the calculator gives approx 100% Faraday - good place to start. Will run unit for a couple of days to re-lay protective oxide layer and then start the series of test planned before putting into a vehicle.
pic -> upside-down new config pic -> PVC container pic -> outer plate pair wired
The initial parallel/series sealed series cell used slots to mount horizontal plates, this new one uses vertical plates (separated by machined insulating HDPE spacers). Output is actually slightly better using vertical plates - I thought it would be the reverse???
February 23rd, 2008
Tried neutralising some NaOH (Sodium Hydroxide) today with H2SO4 (Sulphuric Acid) to see if the neutralising equation worked out. According to the equation the net result is H2O (water) and Na2SO4 (Sodium Sulphate). The sodium sulphate seems to precipitate out of solution. I put the resulting filtered mix into the lyzer to see what would happen - it is a really weak electrolyte.
Cleaned it all out and then tried a 10% H2SO4 electrolyte mix to compare output/reaction to caustic mixes. Seems to output approximately the same although less acid is needed compared to caustic (could be an erroneous statement, just my impression). With sulphuric acid, there are very small bubbles that pop quickly so NO mountains of foam like with caustic solutions of NaOH and KOH (used in soap production - hence the bubbles).
For interest, I kept the H2SO4 solution electrolysing for about an hour. The amount of crap that came out was unbelievable. Most was the layer that the previous caustic mix had deposited, but additional to that, the water took on a dark blue/green tint - similar to a copper sulphate solution. Cleaned out the lyzer with distilled water and had a look at a few of the plates - WOW!!!, it looks like they had been acid etched, one side more than the other, but both sides were really really clean. Now mixed a fresh batch of NaOH solution and started a test run with that. Immediate difference noted - the huge amount of foam does not appear now, it generates very small bubbles somewhere between the caustic and sulphuric bubbles. Next time I'll run the H2SO4 both ways to clean both sides of the plates. Unit efficiency has increased from 76% to 87% or thereabouts. I know Bob Boyce said not to use acid to prep plates (I think he might have meant nitric/phosphoric like some were using). Any sulphuric acid not neutralised by water wash would be immediately neutralised by the incoming caustic mix - can't see any problem with that? Will see what happens once the caustic based protective layer is re-deposited after a couple of days use.
February 22nd, 2008
Completed reshuffle of configuration so now by using different connection combinations, can use the unit to test series 1, 2, 6, 7, 8, 12, 13, 14, 15, 16 cells, and series/parallel for 6x2, 7x2, 8x2. The series/parallel combo will be used for testing 12V (6=2V, 7=1.71V, 8=1.5V) or 13.8V (7=1.97V, 8=1.72V). This will allow introducing an adjustable temp controlled relay to switch higher and lower voltages (and vice-versa) by switching in/out more plates. The aim is to allow automatic heat regulation to some degree, and also add a separate over-temp shutdown in case the unit gets totally overheated for some reason (like has happened on bench before).
Have also wound a small toroid with 43:63 turns to allow stepup of 12V to 18V for overlaid pulsing test (bit rough but hopefully will do to prove/disprove my theory)..
pic -> new config bench test unit
February 17th, 2008
As I go through the cleanse/prep stages using 15VDC and 2.5-3 A, I've noticed different effects of electrolyte levels and/or foam. IF the electrolyte level is raised even slightly above the top of the plates, there is immediately a drop in gas production and voltage (drops to around 13V). Once the level is down to the point where the meniscus drops just slightly below the top of the plates, the voltage comes back to 15V. Having the top of the cell filled with foam doesn't seem to impact the voltage/production at all ??? (still to test with cells half full for definitive results).
If the side electrolyte storage areas are topped up and then the cell is quickly sloshed from end to end (like it would travelling in a vehicle), the electrolyte level almost immediately equalises at the end across ALL the cells before the electrolyte wave drops back - production seems to be only affected slightly during this adjustment. So it seems that just adjusting the storage area level(s) would be enough to maintain levels right across a small cell such as this (17 plates).
Cell connected via outside connections (8 cells each side), doesn't heat up like the big unit did even though it's pushing 40C here at the moment. This may change once the amperage is ramped up later.
Also have noticed most cells stop producing foam after a few hours and just produce clean bubbles within each cell ??? This is with the electrolyte level down about 1/8"-1/4" from the top of the plates.
February 15th, 2008
Got the bench test unit together today and ran first test. Had a few minor fluid/gas leaks (yes, I didn't wait for the sealant to go off again - impatient). After initially taking the wires out through the top, I've decided to change this slightly and take them out the end instead and put some sort of compression seal on each one - it's too much trouble trying to line them all up and keep the top seal in place while putting the top on.
Initial test with 14.9V and 3.2A gave 2.039 litres in 557 secs - not bad first up, 0.22 lpm whereas Faraday expectation was 0.296 lpm = 74%. Over the 8 cells (each side), that was 1.6A with 1.86V per cell. Can't wait for prep/clean cycle to be complete to crank it up. If the ratio stays the same, I should get about 1.85 lpm with 20A (10A each side). Then on to the waiting experiments....
pic -> bench test unit top view
pic -> bench test unit side view
January 28th, 2008
Also started work on a small 9 plate series unit for bench testing. Plates 8 and 9 will have separate connections so I can test different options. Firstly, automatic plate switching idea with relays and temp controls. Idea being to use 8 plates ( 7 cells) to start with and when unit reaches a preset temperature (like 50-55C), then the 8th plate will be deactivated and the 9th plate will be powered instead with the 8th becoming a 'floater' for that period. This should reduce the average cell volts from 1.97V down to 1.72V (for 13.8V supply). Having the switching controlled by an electronic fan relay (or similar for appropriate current throughput), would automate the heating and cooling cycle. This is expanding on the switching idea put forward in Patrick Kelly's documents. If things don't cool down when running with 1.72V, then may have to look at a 10th plate to bring voltage down to 1.53V (switch between 8 and 10 instead of 8 and 9).
pic -> temp control for small series unit
The other thing I want to try before continuing too far with the big cell (or while it's cleansing etc.), is test a small unit with a base low current supply to keep the cells at 1.5V, then pulse up to 2V per cell with maximum amps over the top. I've seen a circuit somewhere recently that allows amplitude to be controlled while pulsing. Will have to see if I can find that one again and try to modify that to do the job. Being able to modify the input voltage (via amplitude) hooked to a temp feedback circuit would be great for later.
December 17th, 2007
Completed single spiral cell tests today using both straight DC, pulsed DC (full range of duty cycles and frequencies) and pulsed DC with electromagnet on bubbler unit. NO difference to original results was observed (wasn't worth measuring, could tell just by bubble stream the rate stayed the same). Tried a quick test at 15 amps and although the output increased slightly, was nothing like the 2.125lpm needed to be anywhere near Zach Wests' claims. At this point I think it may have been a bogus claim and setup. Have now disassembled and stowed on the shelf.
December 14th, 2007
Finished winding secondary, waxed and taped.
pic -> toroid sec during sec waxing process.
pic -> toroid sec waxed and taped.
December 12th, 2007
Ordered 2 extra 250W cartridge heaters and upping the size of the injection moulder to cope with a 6oz shot. Instead of wasting hours machining more slots, I'm going to put the time into machining an injection mould. New section drawing pic -> segment drawing. If this build method works, it should make it easier for everyone to get into series cells. Basically all you'd need to start would be a pair of end HDPE panels, set of clamping bolts (don't need to be stainless - cheap), and a minimum of say 8 plates and 8 cell panels. That way people could start out with a small cell to get the hang of things, then UPSIZE the cell by substituting longer clamping bolts, more plates and cell panels and water fill tube with 'O' rings - to get up to a 60, 100 or 120 cell unit. MAY be able to shave off a bit off the cell thickness to allow usage of thinner stainless shim later. Not sure at this point whether stainless shim will allow resonance or not.
900 watt 12V/110V inverter arrived today along with PWM3F parts.....much to do.....
December 6th, 2007
Ordered the PWM3F parts that I need to build the output section from DigiKey yesterday.
Winding tape and PTFE wire has now arrived - ready to have a go at the toroid wind, just have to wait for a few days clear, don't want any interruptions for this process.
December 3rd, 2007
Toroid core arrived already from Magcore P/L in VIC. $131.25 + freight & GST = $158.13. Expensive but quick.
November 30th, 2007
Ordered solid teflon coated, silver plated toroid winding wire today.
Beeswax (1Kg) picked up yesterday. Now just have to wait for everything to arrive....
November 29th, 2007
Have decided to bite the bullet and ordered a MicroMetals toroid core.
Waiting for silver plated wire contact to reply.
Winding tape is on back order, will wait for a week or so and then try elsewhere if not forthcoming.
Have also ordered a PWM3F board from HydrogenGarage. Intention is to half build (only output section), and hook up to ARM7 pwm emulator.
November 11th, 2007
Prepped cell only very marginally improved, still so close to Faraday. Will
try winding Zach's electromagnets (through a bubbler water column) into the
pulsing circuit to see if that impacts on the cell output one way or the other -
at least I'll get to use a bit of the heap of 2mm (12 AWG) wire I got for the
D18 experiments.
November 7th, 2007
Single spiral test cell is now up and running going through the cleansing and prep
stages. Hooked to benchtop power supply - input current double
checked with analogue meter (PS shows 4 amps, analogue meter shows 5).
Zach West's claimed output of 17 lpm for 8 cells gives a target per cell of
2.125 lpm at 15 amps. So for 5 amps would need around 0.708 lpm per
cell to be in the ballpark.
Initial actual output (measured twice) at 5 amps, 23C ambient temp, is 300ml per
5 mins (0.06 lpm) - almost spot on Faraday calc, but a
loooooong way off target.
Things still to do:
1) Persevere until thoroughly cleaned and prepped and re-measure output.
2) Hook up to PWM and try various frequencies and/or duty cycles to check
outputs.
3) Try spiral in opposite direction (anti-clockwise) for southern hemisphere
(water vortex down the plughole spin in reverse down here).
October 29th, 2007
Finally received injection plans, cartridge heater, adjustable thermostat for same, plus 0.004" thick shim stock for Hasebe/West single test cell.
Have guillotined shim into 6" x 10" sections, then cross-sanded as per Zach West description. Am also using Zach West soldered/insulated wire connection to plates for replica cell -I want to replicate it as closely as possible to remove any possibility of changing something that will prevent it from giving the same results as Zach reported.
October 13th, 2007
While waiting for the injection plans from the US, I thought I'd knock up a Hasabe test cell since there seems to be quite a bit of interest in this type of unit since someone appears to have run a 250cc motorcycle off 8 small units producing 17 lpm. From what I can tell, the motorcycle setup didn't use magnets or pump as per the Hasabe patent but still produces quite a bit of hydroxy with pulsing. The shim stock is available as Shim-in-a-can off the shelf in vary thicknesses, widths and lengths (0.005" seems to be a good thickness to go with). Already have the pump setup required from previous experiments. For a simple test I'll just get enough for one or two cells, build them up and test the output in different configurations before going any further with it. I've often wondered about the effects of a magnetic field on the electrolysis process - guess now I can find out.
October 10th, 2007
Updated section drawing pic -> segment drawing
October 8th, 2007
Changed my mind again! - I want to build an 8 cell test unit to test different configuration ideas BEFORE I built a full unit again.
Rather than build a whole solid unit this time, I've decided to look into injection molded panels to mount the plates and provide a casing etc. for the electrolyser, all-in-one. This follows on from the idea I found on one of the watercar groups by Renmak, although it looks like he's doing a multi-part configuration. After doing some research, I found and ordered David J Gingery books on how to build your own plastic injection/vacuum molding equipment for small items. To do this, I'll machine a set of molds (L&R end-plates and insert pieces) once I have Gingery's instructions to hand.
The beauty of going down this path is I can get dirt cheap plastic offcuts and re-use some free plastic containers as source material. Another benefit is, if one particular model of the electrolyser doesn't pan out, rather than waste the case material, I can recycle it into a new case. This will save heaps of money as the thick plastic sheet was really expensive.
Also using this technology (if it works OK), will provide an easy means of making different size units just by adding extra insert pieces, plates and longer clamping bolts.
pic -> initial segment drawing
The squiggly path of the water on the left of the diagram is an idea of Smoky's that I'm testing - extended water equalisation path = minimal current loss???
There are two segmented holes through the panel at the top (apart from all the clamp bolt-through holes), one is the gas outlet at the top of the gas expansion chamber, the other is for a water fill tube like Boyce's with a small 'O' ring inside to control pressure flow - all molded in...
October 1st, 2007
Disassembled original perspex case to find source of leaks. Found that smoothly machined surfaces squeezed glue out so only a very thin layer was left (and hence a few leaks). These joined pieces broke apart easily. The end that had a small gap to fill with the glue, actually shattered the original perspex panels without the glued parts giving way. Obviously there needs to be enough glue to react with the perspex properly. From now on, with perspex will flycut with slight rough surface so as not to squeeze out the glue.
Next steps - re-sand s/s panels by HAND, machine up HDPE panels for new config.
September 20th, 2007
Got enough electrolytic caps to filter the 4 amp DC supply today (7 x 680uF). Fitted same and started a test from cold - started at 134VDC @ 4A at unit (to get max action while cold), plenty of foam and gas created, variac output was at 112.8VAC.
Over the next 2.5hrs the variac was adjusted down when necessary to keep the amps at 4. Temperature gradually rose from 19.6C to 30.5C (with a max of 31.5C). Final measurements were 111.4VDC, 4 amps, 30.5C. At that time a quick gas measurement was made - 1.25 litres in 35 secs which gave 2.14 litres per min. With the size of the plates, this unit should be able to handle 9 amps which may produce just under 5 lpm.
Caps were taken out and the unit started again right away (same temp, amps etc) but unfiltered voltage was 112.5VDC. After about a minute, the gas output was again tested from a now unfiltered supply. Result was 1.25 litres in 40 secs. This gives 1.875 lpm which is slightly less. Will do some longer tests tomorrow with a greater volume to see if I can verify the difference a filtered supply makes.
pic -> chain of electrolytic caps
September 8th, 2007
While the plate cleansing is still happening. I've been tilting the 61 unit
all over the place to see what would happen if it was in a vehicle and going
up/down a hill, (same as fast acceleration/braking), travelling on a max 30
degree camber and off-camber slope (as well as going up/down a hill). Most of
the severe scenarios would only be likely to happen in a 4X4 vehicle. Best
all-rounder is still the current version 3a of the waterfill pic ->
updated waterfill proposal The 61 unit would be mounted at least at a
30 degree sideways angle (if unit is in-line with the vehicle's motor - best
way). The absolute best scenario would be mounted at a 45 degree angle which
would give the most protection all round (and put the expansion chamber close to
the centre). Even in the severe case of a 30 degree off-camber slope, the
vehicle would be in other trouble before the cells flooded the expansion side
chamber.
September 7th, 2007
Each day brings new challenges and the lessons keep coming...
Finally got all the plumbing on the "GMC-12" completed today.
Although it leaks a bit because of all the irrigation fittings, it has already
served it's purpose in allowing me to see what happens inside the cell when it's
cranked up enough to get it foaming well. Had to apply
140VDC @ 4 amps to get it foaming properly but it's still within Boyce's 2-2.5V
per cell for the cleansing process. I need to keep that
happening for a few hours at a time over a few days to cleanse the plates before
conditioning them (Boyce method). I've already cleaned
it out once and filtered the water and there is indeed very fine particulate
left behind in the filter paper.
Since I overtightened two bolts and cracked the casing badly (told you I was
heavy handed), while the cleansing process continues, I'll order
some ABS and get a replacement case under way. From what I've learned already,
the slots on the side foam dispersal/waterfill idea I had
from before should work really well. I'm sure now I can amalgamate both
functions into the one side unit, and this time I will slot the
underside of the lid as well to seal the plates all round (bar the exit slot).
pic -> assembled with tubes side view
pic -> assembled with tubes end view
pic -> updated waterfill proposal
September 5th, 2007
Talk about false economy, I elected NOT to go to the trouble of buying a finishing tap for the threading kit - BIG MISTAKE!!!. As a result I only managed to get the thread inserts down 1/8" below the top of the 1" threaded holes in the sides, bottom and ends. Tightening the top down to seal against the rubber gasket, popped two inserts and broke two bits of perspex off the sides. Will take the two bits out, remove the inserts, get a finishing tap, tape the holes right to the bottom, replace the two inserts at the bottom of the holes where they belong, then glue the broken bits back on top. That should do the job until the conditioning process is complete, then I'll look at changing to ABS.
I mixed up 8 litres of demineralised water using 10 grams of KOH per litre start with. The 60 cells took just under 5 litres of water to fill. I think I'll have to get some more water and dilute the strength of the KOH as it seemed to be drawing a lot of amps at 100V (didn't measure with meter, will do that tomorrow).
Switched on the variac based DC supply today and cranked it up to 100VDC. Had little geysers of foam and gas spewing out the 120 outlet and waterfill holes as I haven't fixed the outlet balance pipe on yet (waiting for glue to dry). All plates seemed to bubbling evenly which is a good sign. Water level throughout takes a few minutes to even out but eventually enough water leaks between cells to do so (some cells are real slow).
Since I elected to make the outlet and water fill mechanism fit into the lid only (so I can change it without affecting the base unit), I think I might order enough ABS to replace the bottom, sides and end pieces. Can also take the opportunity to change the end and side plates so the end plates go right across the cell and the sides plates are only as long as the length across all plates. Won't take long to whip some more slots through them. Then after the plates are conditioned (and if the unit needs to be cleaned - not the plates), I'll swap them over into the new casing. Using ABS means the case will be more resilient too, it only took a bit of a knock to chip a big chunk out of the bottom perspex plate - it's so brittle. This is likely to happen again as the unit is so bloody heavy - 20.8kgs plus 5 litres of water etc.
All in all though, a good result for first fire up!!!
September 3rd, 2007
Finally finished preparing the 61 plates - long day...assembled base panels with perspex glue. Realised it would have been better to have the end panels overlapping the sides rather than the other way (as they are now). Unfortunately the glue seemed to go off really quick so I don't have a good feeling about how well the unit is sealed - will find out after 24 hours curing of cement. Installed all plates to make sure they would all fit before glue goes off.
pic -> dry assembled unit (no top gasket yet)
pic -> close-up through plates in unit
pic -> dry wieght is 20.8Kg (~46lbs) heavy little sucker
August 31st, 2007
Wow, everything happens at once...thread inserts and CNC plans rolled up today.
Have most of Boyce/Banki unit threaded and screwed together (except for top) with rubber gasket on base. Also started cross-sanding the plates with #80 grit belt on sander - this bit is hard work and difficult to do without touching the plates. Tried rubber gloves but they got ripped to shreds very quickly, leather gloves left slight greasy residue. Have resorted to just moving plates by holding the edges.
August 29th, 2007
Test cut of water fill slots in perspex side - NO GOOD, perspex is too brittle, should have used ABS type plastic instead, maybe with next unit.
In the meantime will use some miniature irrigation fittings and silicone tubing to provide outlets and water fill inlets for each individual cell. Will stagger holes slightly so they are not all in line and have a bit of room around for the fittings themselves. The outlets will be taken up to a central PVC pipe so the foam and gas can settle before being removed to manifold. The water inlet tubes will be higher than the outlets and be vented with a one-way valve to the central cylinder so water in the inlet tubes will settle down to the same height as the water in each cell. If not vented, the water fill tube may provide a current path. The outlet pipes into the central expansion cylinder should allow the foam/gas to settle down and away from the outlet fitting and thereby stop the foam providing a current path as well.
60 outlet pipes of 1/8" ID provide a total outlet area of 0.737sq ins compared to 0.1964 sq ins which is one common 1/2" ID outlet hole through 61 plates.
August 26th, 2007
Ordered CNC plans from http://solsylva.com/ Will probably take around 2 weeks to receive.
August 25th, 2007
Another version of the water fill idea is based on Bob Boyce's angled plate scenario. By angling the electrolyser unit by as little as 20 degrees, the amount of wasted plate area at the top is reduced markedly - gives a very small wedge shaped free area.
August 24th, 2007
First cheapo grinder burnt out today, even before finishing one side. Replaced unit with a better one and re-aligned everything. Slot cutting went fine for rest of the day and the three sides were completed. Dry assembly showed all slots aligned perfectly (yay!!!).
pic -> grinder_replaced - cross-slide driven by cordless drill
August 15th, 2007
Outrigger angle grinder fitted, 1/64" blade used in end as has slight run out, now does perfect slot for 1.2mm plates.
Been thinking about series cell refill scenarios, preferably one that doesn't have some 180 X 1/16" holes and moving parts inside cell.
Basic requirements:
Current idea looks like this Water Fill Idea#2. The theory goes,...plates go all the way to the top of the cell, there's a 2 to 3mm slot on each side of the cell (angle may not be necessary). The highest part of the lower slot defines the water level when level. In operation all output including gas, water and foam goes through outlet slot (left) to side chamber. Any foam or water will run away down the side and therefore not be available to bridge the next cell. At timed periods, the water will be evacuated from the side chamber by pump#1 back up into the fill tube and then into the cells. If the low sensor activates (not enough to fill cells and half side chamber), pump#2 is activated to fill the rest of the cells until they overflow enough into the side chamber to activate the high sensor. If at anytime the overfill sensor activates the previous activity is re-run.
Theory is that foam will drain away and any cell sloshing that causes overflow, will just cause a bit of fluid to go into the side chamber (or maybe a little bit into the fill tube). Overall it should keep the cell water levels approximately at the right level. If foam tends to fill side chamber, there is the possibility to recirculate the water within the side chamber and spray it within the chamber to dissolve the foam. Whether this is necessary will only be determined during the testing phase, but it's an option.
I'm aware that sealing the top of the cell will have to be done differently because we will have 1/8" slots across the sides all the way along, but with a couple of reinforcing bars along the top, end bolts could hold the whole lot down. I don't see this bit as a problem...
Because vehicle acceleration/deceleration/cornering affects the water in the side compartment, there may be a requirement for baffles to stop surge and flooding the end cells. Each baffle would need a small (say 1/8") drain hole or notch in the bottom of each so the level can balance the whole compartment over a few minutes.
After now having a bit of exposure to the AVR microcontroller environment, it seems it would be much easier to use one of the smaller AVR chips to control all the sensor/pump requirements with a small basic or 'c' program. Since these controllers have timers and I/O interfaces built in, FET's can be driven directly to control pumps/solenoids and sensor input lines (inc temp and TPS) can be read directly. If brute force DC is being used, a small interface board could be easily produced to handle the sensor/pump side of things. If microcontroller controlled resonance is being used, level and pump functions could be included as part of the main resonance controller. Could be done either way. It's much easier to have software based flexible logic to control things rather than hardware based, so much so that using a small menu, led display and some mini buttons/keys is all you need to change parameters to suit any unit.
August 12th, 2007
OK, now that I've been set straight about the importance of having a sealed cell with no/minimal current leakage, I'll stop farting around trying to find an easier way to assemble. Obviously the only way to go is with slots as they locate the plates in fixed places. This is necessary for the methods that Bob Boyce and Les Banki use to maintain water level and then seal the cell afterwards. I see what they mean - this is definitely one of the hardest bits of the whole sealed series cell design. Might see if I can alter Les' design slightly to have the same effect but without using a separate casing. Will think about this whilst machining the slots this week. Fortunately last night I worked out if I make an outrigger angle grinder mount (with slitting saw blade) on my 3-in-1 vertical column, I will just have enough table movement to cover the whole area needed. So this is the first task for this week, followed by the slots and then water fill design.
August 11th, 2007
Tried a few different methods of sealing/spacing plates today.
Bob Boyce replied in OUpower group today and re-posted his explanation of why holes or leaks (current) are bad bad bad in series cells, especially high voltage ones - great explanation so I thought I'd keep a copy where I know where to find it:
************************************************************
Re: Question about current leakage
I explained this before, but let's see if I can explain this once
more in a manner that might be understandable.
Say you have 7 cells with 13.8 VDC applied, ie 1.97 VDC per cell.
Now, let's say the barrier potential of the electrodes used allows
electrolysis to take place at 1.47 VDC. So, 1.97 VDC - 1.47 VDC =
0.50 VDC per cell above barrier potential. Each cell has that 0.50
VDC of extra attractive potential to speed along the anions and
cations between their respective electrodes.
Now, drill some holes, and lets see what happens...
The above happens, but... The barrier potential remains the same per
cell, as it is ALSO the same from end to end (through the holes).
Let's figure that attractive potential. The applied potential 13.8
VDC - 1.47 VDC = 12.33 VDC. So now, we have the relatively weak 0.50
VDC attractive potential in each cell, and a much higher attractive
potential of 12.33 VDC from end to end. Where do you think the anions
and cations are going to want to travel?
Maybe now you might understand the magnitude of difference, and why
anions and cations will squeeze through even the smallest of holes to
bypass all of those barrier potentials in series in between. The path
of least resistance is to bypass the plates in between and their
barrier potentials, even though they are closer and have potential.
The actual amount of I*R losses depends on many factors such as hole
size, electrode spacing, electrode configuration, electrolyte
concentration, applied potential, ect.
As bad as this is for a 13.8 VDC booster, just imagine the potential
bypass forces involved when a cell stack is made up of many dozens of
plates and running on ~ 160 VDC (rectified ~ 120 VAC) or more. This
was one of the reasons the welder units from George Wiseman had such
poor comparitive efficiency as compared to a similar unit with no
holes drilled in the plates.
Bob
************************************************************
August 9th, 2007
Holy shit !!!! - don't know what's going on, maybe the planets have all aligned for me at the moment? Went to pick up my 20mm perspex pieces I ordered cut-to-size today (quoted ~$140AU all up), well they ran out and can't get any more 20mm for a while, BUT they had enough to cut mine but didn't quite have enough for a full size spare piece that I ordered as well, and the pieces had a few scratches here and there (minor) - so they gave me the whole lot including a large tube of solvent glue for $70AU - stoked!!!
Picked up the 0-240VDC variac today as well, even that was good news...found a new electronics store has opened up just around the corner and they have all that stuff as well. Saves a long trip to get resistors, cap etc. from now on....does it get any better than this?
Maybe NOW is the right time for me to be working on this electrolyser?
----------------------------------------------------------------------------------------------------------------------------------------
After doing some accurate measurements today now I have the perspex, have decided to change to a 61 plate booster with inter-plate gap of 3.1mm. This will give me an overall unit measurement of 300mm and will then fit into the current 2nd battery cradle. I thought originally I might be able to squeeze it in but it will be too tight a fit and may cause problems with perspex cracking later. Found I already have a 3/64" (1.2mm) slotting blade (from a cheap ebay purchased chinese set), will adapt to a spare 41/2" grinder (with speed controller) so I can cut the slots along the 3-in-1 bed rather than across - then I get to use the power feed (at least for the sides). Also I can plunge cut the bottom slots so I end up with blind slots and not have to worry about the sealing of slots later.
Although the perspex was cut to size, unfortunately the edges are not exactly
90 degrees so it's just as well I'm cutting it all down slightly so then I'll
end up with nice fly-cut edges to fit together.
August 8th, 2007
Read a few posts from Bob Boyce today that finally convinced me to shell out for an autotransformer. Always have talked myself out of it each time the subject has come up in the past. Been trying to work out what sort of voltage I want to run/test this 81 plate unit on the bench - don't feel like rewinding a toroid over and over just to try different settings. Bob said best thing is to use an autotransformer so you can adjust the voltage up/down to what's needed. The ones available here in Oz are $199AU 240v units so with a bridge rectifier on the output, I should be able to dial in anywhere from 0V to 240VDC (not going to worry about filtering - can't find any big enough caps anyway). Once unit is conditioned and I've done some output testing, will have a better idea what vehicle converted voltage should be. Interesting that Bob Boyce's hex controller board uses an Atmel AVR microcontroller chip (with inbuilt multi PWM units). My recently purchased AVR development kit handles these so maybe I can bypass building a PWM board and go straight to an AVR controlled multi-PWM setup (or Les Banki's, whatever is quickest and most efficient).
August 7th, 2007
Went to back to s/s suppliers and ended up getting 190 of 1.2mm plates 152mm X 178mm for $1.36AU ea - ripper!!! Also went to see about some 20mm perspex offcuts but found it was cheaper to get pieces cut to size rather than do it myself - pick case pieces up tomorrow. Using 20mm perspex I should end up with a 81 plate unit measuring 310mm X 222mm X 196mm. Hmmm...not nearly as big as I thought it was going to be, should fit in an extra 4X4 battery tray for easy mounting. Also going to use the ribbed rubber mat instead of cutting slots - plate gap is 2.1mm using 1.2mm plates. Would have been nice to use 0.5mm plates but then I would have to pay full price for a new sheet and guillotining.
Found someone's results from sealed cell testing today using different plate numbers and 13.8V. Turns out the most efficient was 8 cells as the voltage per cell was 1.725V. Below this (9=1.53) the cells needed external heat applied to get things going and over this (7=1.97+) the cell produced it's own heat. Decision made - will use 1.75V per cell initially as I won't be creating a Boyce/Banki multiple freq toroid to start with. With 80 cells that gives a target voltage of 140VDC. To work out the most efficient but satisfactory gas production, I'll run some tests after the unit is assembled and plates are conditioned, to work out the best ratio of KOH to use to provide the best current density.
Theoretical current density should be between 0.025 and 0.040 so I'll aim
somewhere in the middle at 0.0325. From this, using a full plate size of 15.2cms
X 17.8cms, plate area is 270.56 sq cms and this gives a target CD of around 8.79
amps. Shit that's 1230 watts, maybe I'll see what the actual tests reveal first
but I'll have to at least design the electronics for 2.5 amps output (maybe 4
amps if possible, absolute max 560watts). Since I'll be controlling things with
a PWM chip, the actual used power may be substantially less - I hope!!!
August 6th, 2007
Went to s/s suppliers today to see about a bit more s/s sheet and getting my spare bits cut down to 100mmX100mm squares. BUT, while there I noticed a couple of hundred 150mmX180mmX1.6mm 316 s/s offcuts lying by the guillotine. Got a price for 160 of them cut down to 100X100mm - $1.50AU each. Will go back tomorrow and see what I can get the lot as-is for, no extra cutting - hopefully for less than $1AU each.
These will do fine for a few 81 plate Boyce style electrolyzers. Should be able to fit 81 plates into a space about the size of a car battery and so will be able to fit in the engine bay of my current 4X4. Since I'll have to create the electronics to provide 120-136V for the unit, will have to decide if I want to target 1.5V (Bob Boyce recommendation for pulsing) or 1.7V per cell. Was reading some research and experimental data today that mentioned that other than the theoretical point of electrolysis of 1.23V, there is an extra 0.07v required to liberate the H2 and 0.4v to liberate the O2 from the electrodes. Even though this only gives 72% efficiency overall, the extra voltage provides enough to get things moving properly.
I have various examples of push-pull circuits based on TL494, SG3525, UC3526's. From these I should be able to create one suitable to drive this unit as brute force DC unit to begin with.. Will also make a jig on the bench to speed things up when it comes to cross-hatching the plates with a belt sander. So much to do, so little time...
August 5th, 2007
Have enough spare s/s lying around for one 485mm X 100mm dia tube, about 73 X 95mm circular plates and also a 300 watt 12/240v inverter to have one last go at a booster. I'll get enough extra s/s to make up 161 plates (160 cells) - should have about 1.5v per cell with 240v. Tube length needed: 161 X 3 = 483mm.
Evidently the current density (amps/cm²) range should be between 0.015 and 0.060 with the most efficient being somewhere between 0.025 and 0.045. Using 95mm dia circular plates, I should have somewhere around 70.91 sq cms area per plate. This will give a CD around 0.014 with 240V @ 1A (=300/240*0.8 : 80% efficient inverter). To get the best efficiency I should really be using between 1.77A and 3.19A with 240V, that's between 425 Watts and 766 Watts - bit beyond capability of a 300 Watt inverter. Depending how things go, will look at creating a PWM controlled flip-flop inverter to suit, this will also provide a means of controlling the output by varying the duty cycle.
August 4th, 2007
While wandering through the rubber supply stores (again), I noticed some very
small weatherstrip, grabbed a couple of metres to try it out. This stripping
measures overall about 4mm high by 3 mm wide and seems to suit both 1.5mm down
to 0.9mm plates (not sure about 0.5mm, haven't got any to try). You end up with
a 2mm gap between 1.5mm plates. There was thicker stripping available but I
wanted to see whether this small profile one would possibly do the job.
End profile:
000_0287.jpg
Cut 90 dgree notch at corners (not all the wat through):
000_0289.jpg
Fit up two 1.5mm plates:
000_0290.jpg
Two plates together:
000_0291.jpg
Two plates gap:
000_0292.jpg
Not sure this would be of benefit but if you're looking for a flexible (ha
ha) self-adjusting, non-slot cutting alternative, this may be the one.
Continuing this train of thought further, I now realise why a couple of
electrolyser designs are made of circular pieces of s/s inside large PVC tube
(never thought of it before now). Diagram of area differences to exposed/usable
electrode area: ComparisonRoundSquare.jpg
- I think I calculated the area right, but ccould be wrong...
Using the small weatherstrip profile (and maybe a small smear of soap to aid
assembly), the circular pieces could be slid into a tubular container very
easily - a tubular container you say???....hmmmm, where could I find one of
them...
Yes, this leads me to the possible use of end plates, cylinder and threaded
rod from any spare non-used D18's. If the weatherstrip is used for
cell-sealing/mounting, then approx 100X95mm dia round plates should fit inside
the standard 12" D18 (or adjusted dia to suit). The centre threaded rod
would be isolated using plastic tubing, and maybe a smaller 1/4" rod added
lower/higher to lock the plates into position. As far as outlet goes, the top
could either be cut off the plates, or 1/2" hole drilled through them. To
balance the water level, a small 1.5mm hole could be drilled near the bottom of
the plates and this should be small enough not to cause too much current leakage
(been used elsewhere). Other than that, the other D18 water level schemes may be
adapted as well.
The overall plate area would be around 1100 sq ins which is roughly equal to
half of a 61 plate 6"x6" setup - but used more effectively (less
unused plate area). All that would be required for final connection is to bolt
the two end s/s plates through the lexan/perspex end plates and use those two
connections for whatever voltage is put through. If two of these units are run
in parallel, the voltage and current should be able to handle the same as one 61
plate 6"x6" unit (like Boyce/Banki).
July 23rd, 2007
What a huge disappointment...new 11 LPM pump arrived today and after fitting and testing, decided to do gas measurement tests of booster unit with inverted water bottle etc.
BoosterCellOutputMeasurement
Thought my output was quite good (2-3 1/2'' bubbles per sec) but now I have a better idea what the REAL gas output is - bloody miserable. Until I can find a proven method that produces something in the order of 130-135 LPM I won't even try and install anything into my vehicle.
Upside is ..... now I can get back to testing with the mini test cell without being distracted by wondering what a booster unit would do.
July 22nd, 2007
While waiting for new pump to arrive, I have assembled my plates again with 2 layers of electrical tape on each edge - gives about 0.025''gap still, then trimmed to be about 1/8 wide from edge. Fitted plate unit into a 75mm PVC tube and used a couple of s/s bolts for connections. Blocked gaps around plates with expanding foam - messy but good enough for a first test. If I could have found some 50mm square PVC downpipe, would have used that instead and made plates to fit. As usual sealing the electrical connections was the hardest thing to accomplish. Did notice that two of the end plates have shorted (probably when I bent the plate tabs over) - not a big deal just means theres one less neutral plate on that set. All seems to be bubbling pretty evenly as it is.
Anyway, finally bench assembled both plate unit and bubbler. Now have it setup on a timer so it switched on for 1 minute every hour, 24 hours a day. Will leave that for a couple of days to pre-condition the plates. It currently draws about 4.5-5 amps with the small pump off, but goes to 5.5-6amps with it on. You can tell by the flow of things through the connecting tubes that the pump can't keep up. Natural flow created by the bubbles themselves is enough to circulate electrolyte slowly. Bubbles coming from the bubbler unit stay at about 2-3 x 1/2" bubbles per sec.
After running the unit solid for 1 hour to start with, the unit was still cold - means I'm not wasting too much power heating water.
Found a better way of creating new bubbler, less complicated and smaller - same as plate unit at moment. Just create a 75mm PVC cylinder, glue one plain end cap on, and use a couple of turns of pipe teflon tape around top before pushing the top cap on. The teflon is enough to create a low pressure seal, but still allows the unit to be disassembled if need be. The other benefit is that it's an inbuilt pressure relief in case of flashback - top will just pop off...
Want to find a small pressure relief valve to fit on bubbler so when the electronic solenoid valve closes when ignition off, any pressure built up in the bubbler unit is released at say 3-4psi and vented underneath the vehicle. Since the whole unit will be running under vacuum, only a very small type of rubber flap valve would probably be enough.
Have found that the gas does not bubble up through the filter wool as small bubbles so I'll reduce the layer of filter wool, bring the outlet pipe up a bit higher above the filter wool and fill the rest with diffusion material (like a scourer) between the filter wool and top of the inside of the bubbler.
Updated booster layout
July 17th, 2007
Found a mail-order marine 11LPM inline 12V pump today for $49.95AU plus postage. Will order one tomorrow.
Did a few test runs of different spaced slots today and found a gap of around 1mm (0.03937") will be a good compromise from 0.025". Gives a lot more strength and still keeps plates close.
Updated booster layout
July 16th, 2007
Bit of a backward step, disassembled the plate unit to address the sealing problems and found a few of the 0.025" spacer pieces had broken off during assembly. Current plate unit design is scrapped. A different design of top, side and bottom panels will be machined. This time I'll increase the size of the plate spacer pieces to 0.045". Hopefully I can make the sides larger and bolt over and under the centre slotted pieces (with a full size rubber/silicone/nitride gasket to seal them), thereby getting rid of the screws.
Found a couple of better 12V inline pumps - throughput of 7LPM and 10LPM (rather than 3.3LPM). The latter being $84AU - bit expensive so I'll wait until I see how things go with a slightly larger plate gap. Maybe the foam won't be as much of a problem with a slightly larger gap, time will tell.
Won't be using silicone sealer again, makes too much of a mess and takes too long to cure properly. Will stick with real gaskets from now on.
July 15th, 2007
Latest progress report...found pump does NOT have enough throughput to dislodge the small foamy bubbles before they form into large ones therby bridging the .025" gap between plates. This stops electrolysis at that spot until the large bubble dislodges (if it ever does). Will remove small pump from bubbler unit and insert an in-line sink pump in return pipe to move things along and see if that cures the problem.
Also, found once the pump was on and unit was under power on bench, ANY weak spot developed quickly into a small fountain - cannot rely on silicone type sealers to do the job. Will machine a new end-plate and re-jig the 4 power tabs to output via two s/s bolts which I can seal properly with 'O' rings. The unit has to tolerate a small amount of pressure even though it will be under vacuum while running in vehicle.
White foam spews out of the plate unit when under way with 5 amps, am hoping once the rockwool filter material is fitted into the bubbler unit, the amount of foam making it through the rockwool will be negligible. Then I might be able to crank it up to 20amps - the target.
Have fitted a 38mm rubber bung plug to top of bubbler unit to provide a weak spot in case of an explosion. Hopefully it will just pop the bug out and not damage anything else (if it happens).
Top of bubbler and end of plate unit (connection end plate removed)
Top view of plate unit (connection end plate removed)
July 5th, 2007
Photos of booster construction...slow progress
Proposed booster layout
Twenty two plates cut - not blasted yet
Cutting the casing slots
Bubbler and pump unit
Bubbler internal 200 LPH pump
July 2nd, 2007
Going to add more plates to booster and make it 21 cell. Have finished cutting the extra plates to make 22 in all (4 x tabbed, 18 x plain). Next, machining casing slots, then shotblast and prep plates before fitting and assembling. Quick end view of what I'm planning for the plates is below. Not sure if this series/parallel setup will work but time will tell.
21_cell_layout
June 29th, 2007
Have decided to create a quick booster unit to reduce my ongoing fuel costs while we all work towards solving Stan's puzzle. Got some 0.9mm 316 offcut stainless sheet today and set about cutting some into 14 of 2" x 8" plates. Buggered up my guillotine blade in the process so next time I'll just order it cut to size as the place that supplies my stainless, also does job cutting etc. - much much easier (and cheaper if you count damage and time involved).
Dry assembled my 90mm PVC bubbler/storage container (top unit) today as well. Found a small 12v fountain pump that will fit inside the bottom of the bubbler so I can forcibly circulate the electrolyte from the bubbler through the cell and back (with gas). Having glass wool filter material in this unit will also filter out any impurities (and brown scum) along the way - that's my theory anyway.
Will be machining my plates to size this weekend and hopefully make some progress on the booster unit.