Fridge Compressors

[rtbarton]

A shaft position/speed encoder is easy to implement.

I was envisaging extending the motor shaft through a hole in the top of the casing anyway to attach fan blades to, so something to detect the fanblades spinning by would be easy, except for the fact that the motor thrashes around a bit on a cold start, so the close proximity found in the usual LED/Photocell setup may be difficult.

I could mount a sensor on the motor casting, but it would get covered in hot oil.

alternatively, you could deduce start up procedure by measuring the change of pressure in the line to the NRV.

try to start a pump

close the solenoid

see if the pressure is increasing

open the solenoid

try to start the next pump

close the solenoid

see if the pressure is increasing, quicker than before (ideal factor of 2)

iterate for more pumps

I'd thought of that one, but it wouldn't work if air is being used at the same time, perhaps intermittently as with a tyre inflator. Also it would be long winded, say 30 pumps, each one starting in sequence taking 30 secs would give a 15 minute start-up time. Also if you work on a 5 psi pressure differential you'd end up with 100 psi after 20 pumps were running, but as a pump is not gauranteed to start at a pressure > 10 psi, especially when cold, some of the pumps wouldn't start anyway. Starting out with an empty receiver would delay the sequence even more.

For a warm start I'm thinking around starting each one at 1 sec intervals to avoid mains surges and then check for shaft rotation 20 secs after the last pump starts, kill any that haven't started and shut the dump valve.

This 20 secs would have to be extended for a cold start.

There are other probs not accounted for here, most significantly the way the thermal trips operates when a pump doesn't want to start.

[LandyManLuke]

I think Roger tried taking a standard pressure switch apart to look at it, at which point it distributed itself around his workshop, with the aid of the big spring (that was) inside it.

[rtbarton]

What about just reducing the size of the exhaust on the dump valve. Then it will require less flow to build up the pressure?

Si

Tried that - they're b*g*er to get at. Also considered trying to weaken the spring, but after the whole switch flew apart and I had to buy another I threw the towel in.

[rtbarton]

ATF - Got it up to 85 degC (motor casting was up to 100 degC) and quite a lot of vapour came off when I lifted the lid. Given that ATF is quite a sophisticated fluid and more expensive than cheapo engine oil I'll stick with the latter, although I'm going to try a thinner grade.

Cooling - a fan definitely helps, I tried a 4 inch one underneath which didn't help all that much, the smaller one that Luke used would be better as it concentrates its airflow onto a metal surface, rather than along the sides. The 4 inch fan worked best blowing on the side of the unit.

The motor with the extended inlet pipe ran significantly cooler than the one sucking in its own hot air.

I'm hoping to do some more scientific tests if I can find time later.

[simonr]

OK, what about using a check valve (non return valve) on the output of the pump. Between the pump & the valve have a pin-prick sized hole in the pipe to allow it to de-pressurise over a couple of mins.

OK it's a bit wasteful - but easy to implement.

Also, you could use a solenoid valve which closes when the pump is powered. When the pump switches off (when it's up to pressure), the valve opens and de-pressurises the pipe between the pump & check valve. This would have less hysteresis (delay) waiting for the pipe to depressurise before you can switch the pump on again.

Mains washing machine valves are pretty good with air - and available from any one of those dumped washing machines!

Si

[LandyManLuke]

Also, you could use a solenoid valve which closes when the pump is powered. When the pump switches off (when it's up to pressure), the valve opens and de-pressurises the pipe between the pump & check valve. This would have less hysteresis (delay) waiting for the pipe to depressurise before you can switch the pump on again.

Mains washing machine valves are pretty good with air - and available from any one of those dumped washing machines!

Si

this is the current set up i have, a washing machine valve tee-d off before the NRV. it works pretty well . just need to sort control now.

[TheRecklessEngineer]

Luke...

Slightly OT, but where did you get those 12v water solenoid valves, and what sort of price were they?

[rtbarton]

OK, what about using a check valve (non return valve) on the output of the pump. Between the pump & the valve have a pin-prick sized hole in the pipe to allow it to de-pressurise over a couple of mins.

OK it's a bit wasteful - but easy to implement.

I'd thought of that, no need really because although there is a NRV if the compressors themselves when the pump is stationary there is a slight leak anyway which will drain the pipes between the pump(s) and the receiver. This still won't allow enough air to bleed out whilst the compressors are starting up, which is why I'll be opting for----------->>>

Also, you could use a solenoid valve which closes when the pump is powered. When the pump switches off (when it's up to pressure), the valve opens and de-pressurises the pipe between the pump & check valve. This would have less hysteresis (delay) waiting for the pipe to depressurise before you can switch the pump on again.

As Luke says, just need to arrange the control!

[LandyManLuke]

Luke...

Slightly OT, but where did you get those 12v water solenoid valves, and what sort of price were they?

They came from RS, because i'm lazy. No doubt they could be sourced locally for a lot less than the £9 (each) i paid for them, though at that price they're still significantly cheaper than a 12v or even 240v air solenoid valve.

240V version would be in plentiful supply at your local scrappy!

The valve is the same unit, the coils are removable and swappable.

they're ok for simple use, but they are only truely one-way, and they don't like operating with a small pressure difference across them.

[rtbarton]

Just done some quick tests. I won't bore you with figures but..

1) Extending the inlet didn't do much in terms of temperature rise or rate of pressure increase, but does significantly reduce oil consumption and contamination, worth doing for this alone.

2) Changing from 20/50 to 10/30 oil made a big difference to starting up, and reduced final temperature by about 10 degC.

When starting from cold the motor spins a bit and then the overload device cuts in and takes about a minute to reset and the cycle repeats until the oil has warmed up enough to allow the motor to run. The oil is heated by the motor windings.

Ambient temp 8 degC 19 tries (14 mins) to start with 20/50

Ambient temp 7 degC 10 tries (8 mins) to start with 10/30

I allowed 24 hours between tests and changed from 20/50 to 10/30 with the motor hot and ran it for an hour to completely circulate the new oil.

My measurement devices were a bit crude, Temperature only read to nearest degC, Pressure graduated every 5psi and ambient temps were 7,8 & 9 degC. I'd be interested to repeat the experiment with automatic data logging and more controlled ambient temperatures.

[rtbarton]

One of my pumps seizes solid at TDC and you need pliers to free it off, so I thought I'd investigate.

Removing the head there was a large amount of carbon on the inlet valve, which is right above the piston and the latter was hitting it.

Cleaned it all off and made a new gasket, which leaked - they don't make cornflakes packets like they used to - anyone know of a good gasket material and where I could get some? It has to work at 200 psi and 100 degC. Also air was leaking back straight through the pump.

Head off again and close examination showed the inlet valve was slightly bent, so that was staightened up OK (The valves are just flaps of springy steel).

Decided to dig further and have a look at the outlet valve. This was absolutely bunged up with carbon.

I can't believe they get like this in normal service, and I reckon it's the oil getting too hot - another good reason for extending the inlet pipe and fitting a fan.

I'll try to post some pictures tomorrow.

[pugwash]

just out of interest but why aren't you using compressor oil?

i have a couple of hydrovanes at work (small ones) and seized one very quickly when i used the wrong oil once- have no problems with the compressor oil i buy from the local tool place- it always comes out nice and clean when i change it!

[rtbarton]

just out of interest but why aren't you using compressor oil?

i have a couple of hydrovanes at work (small ones) and seized one very quickly when i used the wrong oil once- have no problems with the compressor oil i buy from the local tool place- it always comes out nice and clean when i change it!

When I first started messing with this project the inlet to the pump was inside the housing so it sucked in a lot of oil and expelled it as emulsion in the air, thus it was a total loss system and with Machine Mart's synthetic oil for the Ssshhhhhhhh compressors coming in at £10 per litre a cheaper alternative was sought.

I did experiment with vegetable oil, I can get small amounts free and save the Planet Blah Blah but that didn't work.

I'll keep an open mind and carry on experimenting, but a scrap fridge motor is a whole lot cheaper than a Hydrovane!

[pugwash]

are these compressors greaseable?

[rtbarton]

are these compressors greaseable?

No, they need to sit in an oil bath for cooling. AFAIK The bearings are pressure lubricated, with the lid off and when warm they'll fling oil a good couple of feet radius all over the place. The oil runs down the inside of the housing for cooling and some of the mist gets sucked into the pump.

[rtbarton]

Some pics, more on My Flickr Page

[rtbarton]

This diagram shows the operation of the pump

Note the exhaust valve, it is the horseshoe shaped thing in the previous post, the opening to the piston is 4mm dia, the opening to the outside world is 1mm, this means the pressure in the chamber in which the valve sits must get quite high and a lot of heat will be generated, which could explain the gunge.

I have measured the temperature on the cylinder head at 100 degC.

I think I might drill this hole out a bit, the pump may run cooler and possibly a bit faster.

[rick]

No, they need to sit in an oil bath for cooling. AFAIK The bearings are pressure lubricated, with the lid off and when warm they'll fling oil a good couple of feet radius all over the place. The oil runs down the inside of the housing for cooling and some of the mist gets sucked into the pump.

naa, those little suckers are only splash feed

[rtbarton]

naa, those little suckers are only splash feed

Something sucks the oil up from the bottom of the pot to the top, the drive shaft is hollow. It's expelled at the top and the piston is splash lubricated. If I have to take a motor to bits I'll find out then.

How do you release the mounting springs, and more importantly, how do you get them back!!

[rtbarton]

Got the exhaust valve cleaned up, rebuilt and re-fitted.

No binding anymore, runs sweet as nut. Reached a pressure of 150 psi and didn't leak when the motor was switched off.

I managed to get some decent gasket material off eBay. The gasket that goes against the block is important because the piston is slightly proud of the block at TDC.

I drilled the 1mm hole out to 3mm, this is the size of the matching hole in the inlet valve plate. I didn't want to increase the size of the hole in the plate for fear of damaging it.

I'll measure the time it took to reach full pressure and the temperature rise tomorrow.

[rtbarton]

I'll measure the time it took to reach full pressure and the temperature rise tomorrow.

Not as big a difference as I expected, took about 10% less time than the fridge motor (This one is a freezer motor) but because of the way the pipes run the volume was slightly higher. Temperature rise was about the same. The freezer motor only took two attempts to start at a temperature of 7 degC which is much better than the fridge motor, but again the freezer motor is more powerful.

[rtbarton]

Refitting the lid and trying to make it oil tight!

I set an O-ring in Hylomar in the groove between the welded-in Jubilee clip, and soldered a small stub of copper pipe into the hole in the lid.

Even so I got a small leak I guess I'll have to devise some method of screwing the lid down tight to compress the O-ring, the cable-tie wasn't upto the job. Also some tiny splashes of oil found their way out of the copper stub due to slight "crankcase pressure", this will need sealing and a breather providing.

Note the spring inside the inlet pipe, this is provided to prevent the plastic pipe collapsing, it gets very soft due to the heat. Nylon rather than PVC would have been better. The spring is a piece of an old MIG torch liner, stretched so that the end can be inserted into the small inlet pipe on the pump casting.

I got the O-ring from these people Polymax It was Pt No BS259N70 - 158.34mm x 3.53mm (ID x Width). This was a slight stretch fit on the freezer lid, maybe the next size down 151.99 x 3.53 would have been a better fit on the fridge motor.

[rtbarton]

Solved the splashing problem with an outer inlet pipe running outside the extended inlet pipe, and the leak was solved by clamping the lid down hard with a Jubilee clip.

The unit now breathes through the outer pipe and the fumes are sucked into the pump. This going to contaminate the air slightly but not as much as not extending the inlet. This does mean that fumes don't get into the room, although I could extend the inlet to outside the building!

I'm going to try an in-line fuel filter as an air filter, like the Clarke Ssssssshhhhhhh compressors do, should quieten it down a bit too!

I've just found this place Silicon Hoses they do small bore silicon vacuum pipe whch may be more appropiate for the inlet pipe than the screenwasher tubing I'm using at the moment.

[rtbarton]

Got some 3mm ID silicon vacuum hose from the link above. Perfect fit on the inlet pipe on the pump casting and a close fit in the 10mm stub soldered onto the lid. No need for the outside plastic pipe or putting a spring inside the inlet pipe, the silicon stuff bends quite sharpley without kinking and is good for 120 degC.

Whilst testing this particular pump it stalled under pressure after about 30 mins. Rather than dismantle it I flushed it out with meths and a load of brown gunge washed out and lo and behold it started working again.

[rtbarton]

I've just taken another compressor to bits, this time it was a Sterne from a large fridge.

Rather a novel crank arrangement!

The lid is modified to take the inlet pipe (the vertical copper tube in the background) and a longer pipe to take a dipstick or use with an oil circulating pump. The large hole is there to enable the crank to be un-seized if necessary without taking the lid off. It will be filled with a blind grommet or rubber bung.

This pump had seized up due to using cooking oil as a lubricant, it freed off OK, but flushing with meths didn't fix it, fortunately flushing with white spirit did.