Brads Gas & Gear

[BradC]

Well you pegged the correct weather station. Belmont at the airport is the most reliable one close to me (<10 mins in the car).

I've been in contact with an EEV manufacturer in China. I can get them at less than $20 each, but I have to buy 1000 at a time. I've ordered a fair qty in 4 different sizes on a prototype basis and I'll see what sort of price I can negotiate for smaller Qty's.

Buying Sporlan or Carel at my local fridge store run about $200 each. I see in the US you can sometimes grab them for ~$50 surplus on e-bay.

The only thing I've not been able to find a great deal on is the ratiometric pressure sensors. There a dude selling a range of them that look like they'd be suitable on e-bay, but they are still $30-$35 each (I pay $90 for them trade here). Also, the e-bay units are 1/8" NPT rather than male or female flare. No biggie as adaptors are easy to get, but just makes the sensors less attractive from a price POV.

Now I've seen you can manage with temperature only, it places less emphasis on needing a pressure sensor and the guff to drive/condition it, but I still want to know why my high and low side pressures are so I can monitor the overall compression ration and therefore system efficiency.

[BradC]

So I've had this heat exchanger sitting around for a year or so while I figured out the best way to use it.

I went and picked up some copper reducers yesterday, and swaged out the two big ones with a blowtorch and a 32mm Sidchrome socket in a vice, to make them a snug fit. I then borrowed my neighbours Oxy to braze them in (actually probably could have done it with the double burner torch in hindsight). Brazed them in with 45% silver (always a good move to stainless). The rest of the fittings are done with 5%.

The fittings are a little over 1 3/4" ID and the water fittings are 2" NPT.

The HX had been used for a hot water supply in an apartment block, and it fed hot water to 20 units. There was a lot of gunge in the boiler side of the unit from the coolant they used, so against all advice I flushed it out repeatedly with a dilute hydrochloric acid mixture and it gently cleaned it all out. I did this about 4 times, then left it connected to a circulating pump overnight feeding clean water through it. This afternoon I gravity drained all the water out of it. and filled it with Acetone (it took over 8L). Sloshed that around for a while before emptying it.

Brazed on the service fittings, gave it a couple of high pressure nitrogen flushes and hooked it to the Vac pump.

I've had to change the Vac pump oil 4 times thus far as it keeps pulling the moisture out (my vac pump does not have a ballast valve).

The stainless rule against the side of the HX is just over 12" for scale.




The HX surface is a little over 8.2 Square meters (from memory thats somewhere around 80 square feet). The unit is an genuine Alfa Laval, rated to 30 bar and 200 degrees C. It appears to be in pretty good nick as it was removed as a going concern. Time will tell I suppose.

There are still all sorts of tiny fragments of crap in the refrigerant side, but nothing a couple of big liquid line driers won't eventually solve.

Using this as my condenser has removed the requirement for me to have a liquid receiver. Each of my Evaporators (I have 2) holds a little under 2L of refrigerant, + another 1/2 L for the liquid lines and a bit for good luck. This HX holds 8.5L on that side, so I can fully pump down the system into the HX and isolate it with a solenoid and ball valve to be able to work on the remainder of the system.

My water table here is at 10ft, so yesterday I sunk a 40mm PVC casing down into the water table so I can feed in a bore spear to re-inject the water I pull out to run through the HX. My supply bore has a spear down at 30ft and is 15M away from the injection bore, so I should be able to push a _lot_ of heat into the ground before I even notice a temperature change on the supply water (if I ever do).

Still have to build up the remainder of the system and modify the supply bore to allow a second pump to draw water (don't want to use the reticulation pump as it moves _way_ too much water for this job).

[AC_Hacker]

BradC,

That's the biggest brazed plate HX I have ever seen!

Do you have any idea what the capacity of the monster is?

If you give me the dimensions, I'll try to find out.

BTW, from my research into the HX thing, brazed plate HXs are not recommended for open loop applications because the small passageways which give them their phenomenal heat-transfer ability, also tend to clog easily.

But if you got it for pretty cheap, why not give it a go.

* * *

At one point in the 'Manifesto' I was trying to come up with a general formula that would allow you to get a reasonable estimate of the heat transfer capacity of a brazed plate HX, knowing the width, length and number of plates.

I wasn't terrifically satisfied with my effort, but I want to give it another try. It would be very useful for when you might come across a HX on ebay or in the scrap yard, to be able to get a rough estimate of capacity.

I also have come across water-to-water brazed plate HXs that already stipulate the capacity, and it would be very useful to have a multiplier that could be applied against that figure, in order to estimate refrigerant-to-water capacity.

I looked at some charts for various refrigerants and they were all different, but not so terribly different as water-to-water capacity would be.

Any ideas or information on this quest?

-AC_Hacker

[BradC]

Originally Posted by AC_Hacker

BradC,

That's the biggest brazed plate HX I have ever seen!

Do you have any idea what the capacity of the monster is?

If you give me the dimensions, I'll try to find out.

Well, it came off a 200KW closed loop boiler. I've actually found no direct data on the unit, but its 60 plate brother (this one is 70) is listed in a cross reference table for another HX manufacturer as a 20 Ton condenser (which is about 70KW). My application is only going to require 21KW maximum, so I'm expecting a nice low condensation temperature.

Originally Posted by AC_Hacker

BTW, from my research into the HX thing, brazed plate HXs are not recommended for open loop applications because the small passageways which give them their phenomenal heat-transfer ability, also tend to clog easily.

But if you got it for pretty cheap, why not give it a go.

Well, it was about $400 so not exactly cheap. I've had a couple of ideas. One is just using very fine screen filters and a settling tank before the HX. The other is a much cheaper low pressure brazed plate HX in a water to coolant role to cool a secondary loop feeding this HX. I've had it priced for replacement and its upwards of $2k, so I'd like to look after it if I can.

Originally Posted by AC_Hacker

At one point in the 'Manifesto' I was trying to come up with a general formula that would allow you to get a reasonable estimate of the heat transfer capacity of a brazed plate HX, knowing the width, length and number of plates.

I wasn't terrifically satisfied with my effort, but I want to give it another try. It would be very useful for when you might come across a HX on ebay or in the scrap yard, to be able to get a rough estimate of capacity.

Once you have the total HX surface area, I have some rule of thumb calcs that seem to hold up.

I did the calcs a couple of years ago. I then gave my application requirements to SWEP and they did their own calcs, which matched mine pretty closely as a sanity check.

The key is the U value (Watts/Kelvin/Square Meter).

With refrigerants to water, you can roughly use 800 as a rule of thumb.
In this application I'm aiming for 21,000 Watts.

So, 21,000 / 800 / 8.5 = 3.08

Because its simple division, you can just transpose to find the missing value.

I can roughly expect an LMTD of 3.08k across the HX, which correlates with my expected input temp of 20 degrees and discharge ~25, condensing somewhere between the two.

This is of course configured in a counterflow arrangement with the vapor going in the top and liquid out the bottom.

That equation seems to hold vaguely close for most configurations of water to refrigerant HX I've come across (tube in tube, shell and tube and BPHX).

You can use the same rule for evaporation, however the reality is you can get more heat across the same surface area in the evaporator as you have more liquid and less vapor. It's a good starting point though.

If you take my 21Kw and use a more reasonable condensing temperature, something like 35 degrees you get :
21,000 / 800 / 15 (35-20) = 1.75
Which is a much more reasonably sized HX.

[BradC]

So flushing with Acetone seemed like a good idea at the time. Don't do that. I'm 7 vac pump oil changes and 24 hours in, and it's still not even looking like it's made a dent. It may even have damaged my Vac pump seals.

[BradC]

Ok, so it was not the Acetone that damaged the vac pump. A very quick "pop the top off" look revealed a broken valve on the second stage. Obviously I must have copped a liquid slug, or condensed a _lot_ of heavy vapor in the second stage, and it tore the reed valve in half.




The broken reed was 0.15mm thick, so I sacrificed a cheap set of feeler gauges and grabbed a 0.20mm leaf. A bit of quick work with the pliers (snap to length), the drill press and some fine emery paper and we have a Q&D replacement reed.

[BradC]

Well, the HX is a bust. It has a leak internally. Now on the search for a replacement.

Also, I had to change that reed from 0.20mm to 0.10mm to get it to pull down nicely.

[AC_Hacker]

Originally Posted by BradC

The broken reed was 0.15mm thick, so I sacrificed a cheap set of feeler gauges and grabbed a 0.20mm leaf. A bit of quick work with the pliers (snap to length), the drill press and some fine emery paper and we have a Q&D replacement reed.

Clever devil!

-AC_Hacker

[BradC]

Originally Posted by AC_Hacker

Clever devil!

Compliment coming from you! Ta.

While the pump works a _lot_ better with the reed replaced, I fear I may have damaged a vane also as it makes a god awful noise that it never used to. I've ordered a replacement (it was a cheap e-bay model in any case). When it arrives I'll pull down the original and see if it can be repaired effectively. Worst case, I'll take off the second stage and make it a single stage pump for pulling a rough vacuum (< 200 microns) on contaminated systems.

One thing the pull down did demonstrate to me is that if you vac out a grungy system or the oil gets dirty, you _must_ dump the oil immediately you switch off. I failed to do that a number of times, thinking I would just change the oil before I next used it, and the contaminants separated out and refused to re-dissolve leaving lots of grunge in the pump housing.

The other thing I learned is do _not_ kick the pump over with the valves off. You will paint the ceiling and entire contents of your garage with vac pump oil. I could not believe how much oil this thing geysered out and with how much force.

Ordered a brand new HX today. Should arrive in about week and a half. I had the guy verify all my measurements with his simulation software, and they checked out, so...

Moving 18KW of heat into ground water at 40L/minute should give me a condensation temp of ~26 degrees with 2 deg subcooling. That gives a COP of just over 7.5, where a really good 18KW air sourced unit is under 4.

As the house load decreases, I can reduce the compressor speed with the VSD, and increase the evaporation temp to see COP's exceeding 9.5.

[AC_Hacker]

Originally Posted by BradC

As the house load decreases, I can reduce the compressor speed with the VSD, and increase the evaporation temp to see COP's exceeding 9.5.

So, when you day VSD are you talking about something like this VARIABLE FREQUENCY DRIVE INVERTER?

Also, could you please go through your HX selection process step by step? That's a mystery for some folks. I tried to clarify it on a design, but there still seems to be remaining questions...

* * *

...also, there is a conversation going on that you might want to weigh in on...

Morgan (AKA: Ko_deZ), who resides in Norway, wants to make an Arduino-based heat pump controller. He has experience with embedded controllers. This could be a project that all of us could benefit from.

I know that not having a good controller for my hacked machines is a serious problem.

Hopefully it could be sufficiently general purpose to have wide application. Xringer is collecting parts to make a hacked air conditioner water heater, he'll need a controller also.

-AC_Hacker