Monday, August 2, 2010

Yet More Leak Testing.

Well the past few days have been filled with leak testing. It has been frustrating. Good silver solder joints going bad. Which means they were not good in the first place!
This is my blanking cap for the top of the pump barrel. Its only 6mm thick. That is the thickest I had unfortunately. I cut a groove on the lathe. As you can see it chattered a bit. Its not a big deal because it not the mating surface.

I polished the outside ring where the o-ring would press against. That all worked out well.
You see, the piece of Al. was not big enough to be a flange too, so I used this piece of acrylic as a clamp. The real problem is that the acrylic bend when you start getting near the right compression on the o-ring (20%). This, I found was deforming the cap slightly. Enough to pop the J B Welded brass adapter off.

It held long enough to do the first few tests. I am doing them at 4 Bar.

Initially I had two small leaks. So I fixed them with great enthusiasm and retested. I then found I had four leaks...

This pattern went on for two days till I finally had it sealed. So now I believe my pump is air tight. Vacuum tight? We'll see.

New Pump Interior

After the failure of my two stage chimney design I was convinced to try a single jet instead. Above is a photo of the 2" copper pipe I'm using. Its upside down here. The small holes are for the condenced oil to flow back in to the middle and the larger ones, higher up are for vapor from around the outside to whizz in to the chimney and go up.
Again this is upside down. The skirt in this case is the same diameter as the interior of my pump barrel. This is only a cap over the condensed oil not a down jet deflector.

I cut 5 nicks in the cone with an angle grinder to allow oil to flow back in to the reservoir. The cone here is another stainless steel funnel.

I don't have a pic but the deflector is also one of these funnels. I have an 8mm clearance around the circumference of the cone for the vapor to bridge before it hits the pump wall.
Fingers crossed.

Saturday, July 17, 2010

Nickel Plating

I got a bit of Bright Nickel plating salts from Caswell Europe. I'm using a 'slow cooker' (which I pulled out of a skip) as my plating bath and a pure nickel electrode. This solution likes to be a bit warm in use.

I think the main things about plating successfully are: cleanliness and current control.

My power supply doesn't have current limiting so I have to use the voltage control and just watch the mA meter. Its probably gives a slower deposit or something, but I've had good results with it.
Here is one of my electrodes. I just did them for the fun, I really need to nickel plate the copper parts of the diff pump interior.

I did notice how slippy the surface became after plating. Even to run you fingernail over the surface is quite different to the feel of copper. It must become very flat and smooth at a microscopic level.

The Heater Section.

I found a 220v 600 Watt element from a travel kettle. I saw 4" pumps tended to be around this wattage so I'm hoping it s enough for my homebrew version. The good thing about this element is that it came encased in a lump of aluminium. This allowed a good surface contact between the heater and the bottom of my pump.

Unfortunately, when I soldered the base on to the barrel of the pump it domed very slightly. So the flat heater and the base didn't touch as much as they could. So I decided the best thing to do was grind them against each other.
I had some carbide grits so I used them. I did 80, then 220, then 400 and then Brasso. It took and hour or so. I was quite happy with the result. Not perfect but pretty close.

I needed to isolate the element from its metal box, electrically and thermally. I had kept a few blown plug fuses, thinking they might come in handy some day. Ceramic tubing!
I cut the ends off then emptied out the sand.The underside of the element had a three M3 tapped holes in a nice equilateral triangle. So I measured that and marked it on to the metal box. I drilled that out and put an M3 nut and bolt through the holes.(they were proud about 10mm, just enough to hold the fuses)

I put a bit of high temp silicone on to the end of the bolts and stuck the ceramic tube on.

I screwed short pieces of M3 threaded bar in to the underside of the element and these sat neatly into the tops of the fuses!
As this was going to be squeezed against the underside of the pump, I didn't fix it in permanently.

I had a nice ceramic connector and some high temp wire, from an old halogen security light. So I got the power out of the metal box where it could be connected without fear of melting.

Quick, quick clamp

This is a picture of a temporary clamp I made while I was waiting for the real clamps to arrive. It seemed to work OK. As you can see, its made of two pairs of split rings at 90 degress to each other.

8 m4 bolts are holding them together. I had a short ring of tubing and an o-ring inside.

Note the black marker on the acrylic to make it easier to reassemble. I made this by eye so nothing matched up unless reassembled in the same way it was drilled!

Thursday, July 15, 2010

Leaks, leaks and more leaks

Its a moral tale... and the moral is: pressure test everything as you go along.

I'm guilty of rushing a few bit here in retrospect. I got a bit carried away with the excitement of nearly having the pump finished and didn't pressure it at the silver solder stage. Because once you start soft soldering you cant go back and fix your bad silver solder joints.

So for the pump barrel I pumped the system down as much as I could with the backing pump and painted some shellac on to the area that was leaking. That fixed that for now. How well it stand up to the 220 degress plus, I dont know.
This was in interesting discovery. I found a bad lead solder seam which looked ok and worked ok for the last few months. But now I found it was leaking. Naturally enough it was the last place I looked. I desoldered and resoldered everything else, assuming that one was good. Then in exasperation, I pressure tested it. And there it was...

Today I hooked it all up and pumped a but of water through the cooling sleeve. There was a leak there too. Actually two leaks:

One I thought I'd fixed, close to the bubbles in the top photo.

The other at the top of the sleeve. This one only became apperant when I was pumping the water through. I had tested it initally just with gravity and it was fine. But the small pressure of the pump pushed a bit of water above the inlet level and squirted out the top.

High temperature silicone!

Fingers crossed.

Friday, July 9, 2010

The diff' pump video. Looky here.

I've just uplaoded a video to youtube fo the diff' pump as it stands. There are a few posts missing between the last one andthe pump as it is in this video but I'll put the missing info up over the next few days.
In the meantime watch the video! <-- click that!

How to shake a 100mm hole in a piece of steel.

This is me drilling the 100mm hole for the ISO flange for the top of the diff pump. Its a bit of a late post but I didn't have the pictures at the time.
The external measurement of my stainless pump barrel is 102mm. The hole saw was cutting a little too large. I think it wasn't exactly a circle. I've noticed this on other hole saws too.

I decided I'd bend the teeth in a little to give me a tighter fit. Well, you can't do that. I is a very brittle alloy. And I happily snapped off two teeth without too much pressure.

So plan B was to sand the outside of the saw to remove the outer edge of the teeth. Now that did work. Always remembering not to let the teeth get too hot or they will loose there edge very easily.
Here it is whizzing around. I set the pillar drill too its lowest speed. I predrilled the pilot hole too. I wanted to get as little wandering as possible. The whole drill press shook like hell for the first few minutes, until the groove developed.
Again, I stopped quite of the to oil the groove and make sure it was cutting freely. I went all the way through from one side this time. The clamping of the disk was a pain and I didn't think I could line it up again from the other side.
It worked out just fine. In fact, it was bang on. A friction fit onto the tube without any sanding.
That doesn't happen very often...

Sunday, July 4, 2010

Electrical feed-through

Today I decided to leave my base plate intact and feed everything through the 'belljar' instead. I had a few reasons for doing this, ease of access being one.
So I got out my trusty hole saw and drill and cut two 40mm holes in the top of the gas cylinder. It went through without a problem. I stopped every now and then to add a bit of 3 in1 into the groove.
Then I sanded back the paint and exposed some clean metal for the epoxy to stick to.
I cut two 55mm squares of 4mm glass to act as insulation for the electrodes. I drilled these wet with a tile drill bit. 6mm. I went three quarters of the way through from one side then flipped them over to finish the hole. This stops a large flake being pushed off the underside of the glass when you get through.
I used plenty of water to lubricate. The drill was running slow and I was only pressing till I could hear the hiss of the glass being cut.
Then JB Weld again! I hope this stuff works...

These are my electrodes for the filament power. They are 6mm brass rod. I silver soldered a flange on to compress the o-ring, which I hope will seal it up. the square bit is a baffle to stop the aluminium from depositing on the glass and causing a short circuit.

Thursday, July 1, 2010


I was cutting a new opening in my vacuum table and I dropped the bit of ply on to my Bourdon Gauge. It popped off at the lead solder. Eejit.
I put wet cloths around the other lead soldered components and fluxed and remelted the solder back into place. As good as new... or secondhand anyway!

Gas Cylinder Belljar

This is going to be my vacuum chamber. I found it dumped in a ditch so I took it back to the workshop and took the brass valve out of the top. They are in there pretty tight! Before I cut into it I filled it with water to get any remaining gas out. This is a very important step. I then pored most of the water out and boiled the remaining water to get the steam to draw any that might remain out with it.
I cut the base off with my angle grinder. I did all this a couple of years ago in my initial efforts to get a vacuum system going. The base flange was a huge job. I think I cut it with a cutting tourch and spent the rest of the day grinding it into the right size.
The view port is a bit of 50 by100 mm box section and a face plate made of 3 by 25mm ribbon steel. All the steel is brazed together. I only have a stick welder so the brazing is a better option for gas tight seals.
I'm using JB Weld here to stick on the glass view port. I don't know if this is going to be a problem in vacuum but I've read some good reports about it. So what the hell! Its the slow kind by the way.I picked up a radiator bleed valve for €1.40 today. Its new and has a metal on metal seal. I hope it works. I threaded the steel washer which I brazed to the top of the cylinder. I will JB Weld the threads to get a good seal.

Today I have decided to do my electrical feed throughs in the cylinder instead of through the baseplate. I seems like a better option for practical reasons.

100mm Hole in 20mm Al. Baseplate.

Last week I bit the bullet and decided to go ahead and cut the hole for the diff' pump in my base plate. I got the piece of Aluminium plate from a local scrap dealer.(€10). He had mixed cement on it and it took a bit of cleaning with patio cleaner. Its dilute hydrochloric acid I believe. Any way, it attacked the cement well and loosened it after a few goes.
So I decided to use my router and an 8mm straight wood cutting bit. As you can see from the picture I did a test cut in MDF first to see if it would work.
First I drilled a 6mm hole in the MDF and pushed in a pivot made of 6MM brass tube. You can see that sticking out of the wood.
I then carefully measured two points on the base plate of the router one at 102mm dia and the other at 98mm dia. I centre punched them and drilled them out at 6mm. (Don't forget to measure from the correct side of the cutter!). So then I had my pivot point on the router. Easy!
Here is the first pass in the aluminium plate. I oiled the hell out of it. On the cutter and in the channel. I did 8 or 9 pases on one side then flipped it over and needed 4 untill I felt the cutter clear into the groove in the oppisite side. I forgot to mention that the 6mm brass pivot went all the away through the 20mm plate, so it could be used on both sides.
This was a critical point...
As you can see I left 3 tabs holding the plug in place. Aluminium is very sticky to work and I was a bit concerened that at the last minute it would catch the router bit and fuck it all up. The tabs were thin enough to be broken off by hand and sanded back. The process was quite painless, the cutter bit was still sharp at the end. I suppose it took a couple of hours in total.
Note: I firstly cut the 102mm circle to about 3mm depth and then 98mm the rest of the way through. This was to imitate the ISO100 flange design. Just in case I get a gate valve one day...

Tuesday, June 15, 2010

What Flange?

Today I was trying to figure out the flange size that I should make up for my pump. Its a little confusing at first. You have to get to know your ASAs from your ANSIs( which are the same!) from your ISO-Ks and your ISO-Fs!
Now a 4" ASA/ANSI flange is 9" dia with a 5 and a bit inch hole. Maybe a 3" would be better...
Well to make a mornings research short. I'm going with an ISO fitting.
ISO-F 100 specifically. So here is a rusty piece of 6mm thick mild steel I had knocking around. I got the dimensions from the Net and marked it out.

After a bit of cutting and loads of grinding I got the disk below. I've just drilled the pilot holes for the bolts. I'm going to see if I can find a hole saw a bit less than 100mm to cut the disk from the center.
Looking good so far!

Monday, June 14, 2010

The Diffusion Pump begins

To reach a good vacuum you need a diffusion pump or other fancy pump. I decided that a diffusion pump would be my best bet to make myself so this post is about my progress so far.
It seems its possible to make a pump from copper. But I found that many wonderful shapes and sizes of parts are available for very low prices in stainless steel!In the above image you can see the parts. I think the image is self explanatory! I silver soldered all the parts together with Oxy-acetylene. My soldering getting better as I went along.
I'm planning to put a KF/QF fitting on the top of the outlet to the roughing pump. I've just got to find a brass blank to make it from.This is the thermos lid with the center being cut out. it was thin enough a snips would do it. Below is the M6 SS nut to hold the jet nozzels in place. Its also silver soldered in place. Before I soldered the base on! This will hold the threaded bar which keeps the jets in place.

Sunday, May 16, 2010

This isn't metalizing I know, but I did it today...

Mini Neon Driver
From Disposable Camera Flash

This is a high voltage DC circuit made from an old disposable camera flash. Its similar to the popular Joule thief I believe. Here is the diagram:
This is exactly as it is in the flash except I've removed the rest of the circuit, like the high voltage capacitor. Ouch! Also the switch in the doodle is closed permanently.
In its place I put a mini neon bulb. I like freeform circuits.

Below its running at 1.5v. I don't know whats the voltage across the neon. (over 100v but less than 400v)

Here's the power supply. Its reading .5 amps at 1.5v it goes up quickly when the voltage rises any more. As this is a DC circuit only one side of the neon glows.Would it be possible to do AC by using two of the above circuits at 1.5-o-1.5v?

The neon in the power switch is more impressive...
The Guts of the Thermistor Gauge.

I refer you to the link on my earlier post of the circuit by Roy Schmaus. Its based on an LM324 Op Amp. I'm just trying to get my head around these devices now. They seem to be used for everything under the sun. Here is my circuit nearly finished.

I've included a 7815 voltage regulator on the board too. You can see that on the right. The three components behind the black croc clip. There is a 7 watt resistor on there which is waaaay too big for the thermistor I used but it was 470 ohm, just what I needed, so I went for it!

At this stage I've yet to put two trimmers on. One to be part of the LM324 divider circuit. And the other to drop the output voltage/current to be used in a moving coil metre.

I got this cheapo multimeter from a Chinese Shop in Barcelona. I don't always go so far to buy components! I took the board out, cut the battery holder and screw posts. This gave me a flush back which fit neatly on to a project box.

So I put in a DC socket and an old mobile phone charger cable out for the thermistor connection.
I managed to squash the circuit in.
Here it is, with thermistor soldered on for testing.

The circuit is amazingly sensitive to air movements around the thermistor. A couple of tips if your building this are:
-There will be a delay after you power up before things get to the levels/temperatures/voltages they are supposed to be at. So give your circuit a few seconds before you take a hammer to it.
-Try putting the thermistor in a little cardboard tube to protect from drafts and your breath which may cause weird readings.

Friday, May 14, 2010

Zero and Beyond!
Thermistor Vacuum Gauge Construction.

I had been thinking for a long time about getting a gauge for very low pressures. I came across this little project a year or two ago but it was beyond my understanding and abilities at the time.

It came from

Even with this nice circuit to work from I still had the problem of electrical feed-throughs. Then one day not too long ago the solution pop into my head.

Strip fluorescent lights!

I had a T4 size fluorescent tube which is just under 16mm dia. Now that's just larger the the external measurement of a 1/2” copper pipe. So the brass fitting could be adapted to fit. Using an O–ring to keep it air tight.
Strip lights have two electrodes nicely sealed in to each end! Perfect! So I cut the tube on the lath and with a hot wire. I then washed out the fluorescent powder. I believe these lamps have may have mercury vapor in them so I didn't breath while I was doing it...

Then I covered the electrodes with a bit of Al. foil to protect them as I rounded the cut end with a flame to de-stress it a bit.

I popped it all in some safety pickle and got the Dumet wires back to nice clean copper. I could then , with a bit of flux, solder the thermistor to the inside.

Once I fitted this to my home-made ‘QF16 from plumbing pipe fitting’ my feed through was complete!

This is turned from a 1/2" tank fitting. I'll go into this in another post.

I'm just waiting for my Viton o-rings to arrive and I'll hook it up.

Wednesday, May 5, 2010

The filament power supply

The power for my filament comes from a rewound microwave oven transformer
(MOT). There is loads of info' online about doing this so I won't go into it here. I'm getting about 6v tops out of this one.

There is also a 'dimmer' suitable for use with inductive loads. It came from a company called 'Brocott'. I got it through You can see the little VR going through the lid of the box to vary the power. (Its a really crappy VR wobbling all over the place. I may have to replace it with something a bit more substantial. Saying that, it doesn't matter at all for the brute force output.)

I'm going to put an illuminated power switch and a 50 amp ammeter to keep an eye on the power without burning my eyes out watching the filament!

Anyone know if you can convert a DC ammeter with external shunt to AC? I've gone and bought the wrong one!