This forum is now read-only

To login to the new support channel and community forums, go to the Support Portal

Cold-Feed Conversion L1 Series 1 Pt 1

I’ve recently completed the conversion of my early 2013 Series 1 L1. There were no dramas but it was an interesting exercise.

This conversion obviously voids any warranties, requires some technical knowledge and tools and won’t be supported.

This conversion can obviously only be made to a plumbed inlet machine, but preferably the outlet should be plumbed as well.

Finished conversion.
From the bottom right we have the water inlet and the non-return valve.
The vertical leg has the expansion chamber and a tee going the OPV Valve and the cold feed inlet.
The OPV outlet tube feeds to the front panel outlet.
Everything else is standard.

The Series 1 machines don’t have a cut-out for the waste tray. This was made by cutting two 60mm holes and carefully removing the central portion with a cold cut blade in an angle grinder.

Attached files

image image


  • A compact waste outlet elbow was machined from a 40mm length of 40mm stainless bar welded in by a local machine shop using the original drip tray. Reiss’s Shallow plumbed in drip tray is a much neater solution but this works well.

    The legs need to be extended to 40mm to give enough underbody clearance. I did this using a 10mm thick plastic spacer and extending the bolt length. Again it would be much easier to just to purchase the 4 x 40mm tall stainless steel legs.

    The cold water inlet and waste outlet were very messy to install which is why I hadn’t plumbed the machine years ago. The Stainless Steel Waste Water Skin Fitting and 19mm waste water hose came from a Marine supplies shop and look very neat. The outlet is plumbed with its own 40mm trap directly to the house waste system.

    The cold water inlet is taken from the mains supply via a pressure regulating valve in the conventional manner.

    Locally I could only find ½” braided steel hose which limited my options as it not nearly as flexible as the 3/8” hose. I also only required a 1.0M length.
    The internal fittings are mainly ¼” BSP with 1/8” and 3/8” used where required. The additional internal pipework is in soft walled ¼” copper tubing using flared fittings as I find these very easy to use and reliable.

    A ¼” BSP cap is required to plug the boiler where the banjo pipe was connected.

    All threaded joints were sealed using Loxeal High Pressure Thread Sealant.

    The Vibe Pump was removed and its electrical connectors securely insulated and taped out of the way. The low water level sensor in the tank was bypassed.

    Attached files

    image image image

  • It’s absolutely essential that a good quality OPV (Over Pressure Valve) is used. The water expansion in the HX loop can generate extreme pressures in the supply line as the Pressure Regulating Valve is a one way valve. A good quality check valve should also really be used to keep all excess pressures within the machine. I initially tried a Fluid-o-Tech OPV valve but these are very poorly made. I’m now using an Expobar OPV valve that’s working well. Its adjusted to about 5 bar. The OPV water out is fed to the drip tray via a hollow bolt in the front panel. (Reiss used to have these. I don’t know how the overpressure water is handled in the L1-L3 machines)



    Out of interest while setting up I’ve installed a pressure gauge in the 18 x 1.5MM hole in the back of the group. My local supplier stocks a tapered 18 x 1.5MM plug that was tapped to take a 1/8” BSP nipple. This shows the pressure profile of the pre-infusion. It also showed that when the lever is released but before it catches a short pressure pulse flows back through the HX loop at up to 9 bar until relieved by the OPV or expansion chamber.
    Initially I was using a rather suspect pressure relief valve and the pressure pulse was taking my pressure gauge off scale. I now have a much better pressure relief valve, but have also added an expansion chamber that absorbs the pressure pulse keeping the maximum pressure to about 4 bar. Its similar to the below device.

    Adding the additional T junction for the expansion chamber made the vertical leg a very tight fit. I shorten the Tee fittings, otherwise 1/8 inch fittings or a custom manifold could have been used.

    Attached files

    image image image

  • image

    I’ve installed a 4 channel data logger that monitors the boiler, group and HX loop In and Out pipe temperatures. This clearly shows that the cold feed has no negative impact on the group temperatures with normal use, but that the group can easily be flushed cold with a prolonged flush.
    The Type K thermocouple bead probes were sandwiched in a piece of Thermal Transfer Tape for electrical isolation and good thermal contact and attached using Aluminium Foil Tape.

    Machine warm up from a cold start. I need about 90 minutes.

    At 120 minutes a single Long Black was made. This was followed by a short flush. The extraction doesn't lower the group temperature at all. The short flush only has a short term effect. Normally I would only flush at the end of a session.

    At 386 minutes a long black was made. This was followed by a short double flush at 390.

    At 401 minutes a long 350ml flush was made. This significantly cools the group, but in reality there is no reason for anything more than a short flush.

    Attached files

    image image
  • Hats off! A very nice job, well documented too.
  • Wow Greg. Nice Job. Thanks for temperature study results too.

    I have a question. Is recovery about 9 minutes with a small flush in your setup? Any idea what the incoming water temp is? My incoming water is pretty close to room temp, but my back to back shot time is often faster than 9 minutes when I have guests. I guess I should probably check and run a quick temperature study, but I haven't noticed any sour shots though. Then again I'm really not pulling too many back to backs and I am taking my time.

    I agree with the sentiment about anyone trying this sort of mod. Use the best parts you can find. I used vibiemme parts (both check and OPV). I adjusted my OPV way down so I get a pretty steady, slow stream during warm up. This then stops once the group and hx is up to temperature. I have no idea what pressure it is set at. I haven't noticed any water dripping during the initial lever release when pulling a shot, but I haven't looked either.
  • My incoming water temperature is about 15 deg C.

    I would never flush between back to back shots and in that case the recovery time is virtually instantaneous.

    I always flush at the end of each session but then the recovery time within reason is irrelevant.

    It would be very handy to have a source of good quality reasonably priced parts. My supplier was very helpful but they normally only sell parts for specific machines.

    My OPV is set to about 5 bar and I used to get a steady drip during warmup. With the expansion chamber the pressure doesn’t rise above about 4 bar and the OPV doesn’t drip at all. The inclusion of an OPV in the design though is essential for safety reasons. Most factory set OPV’s will be set to about 9 – 10 bar and hence need to be adjustable.
  • Great job! Thank you for the precise documentation. Interesting this expansion chamber. I would have thought that the OPV would handle the pressure easily but obviously not good enough.

  • I must admit I tend to over engineer things and strive for the best.

    The OPV easily handles any overpressure but it constantly dribbles whenever cold water is introduced into the HX loop and almost squirts when the lever is released.

    The expansion chamber contains all within the machine and I think is a much neater solution.

    With the original hot feed HX loop the steam above the water in the boiler acts as an expansion chamber.
  • Been looking at this for a while. Please correct me if my understanding is wrong.

    Pressure to the entire machine is regulated via valve to desired preinfusion pressure - 4 bar for example
    Something like this at the house main feeding the machine

    Boiler is fed normally and managed with the existing solenoid valve.
    The group head comes off the tee to apply 4 bar pressure to head vs picked up off boiler. Still uses heat exchanger to warm water as needed.
    A brass plug is used to cap off the boiler fitting not used anymore from rerouting group head feed.

    Exobar over pressure valve bleeds off excess pressure in group head but this valve is not managing the preinfusion pressure. It’s just keeping things from blowing up from over pressure. Over pressure from exobar valve goes to drip pan.

    Is that right?

    How much dribble comes from the OVP valve? Is it just during warmup? I can’t plumb the drain. Is it a few ounces a day? We empty the pan once a week. Making about 10 coffees every day.
  • Hi David
    My machine has worked perfectly since this modification and I have no desire to make any further changes. I would do the same thing again.

    The machine pressure is regulated via an identical looking valve in the cold feed line to the machine.

    Your machine plumbing description appears to be correct. All modifications are easily seen from my first photo.

    An over pressure valve is absolutely essential, but the amount of dribble is only a few drops.

    I think everything else is covered in the previous discussion.

    Its in another thread, but any plumbed machine really should have some form of water leak detection.
    It works as I've tested it several times

  • Thanks!
    Just st wanted to ensure I understood what I was reading. I’m quite handy with these sorts of things and might give it a try. Will report back here when I do.

    The leak detection is something I don’t have but often worry about and will be sure to add with this conversion.

  • Hi all,

    I'm in the process of doing a similar conversion on my old L1 - Mostly collecting needed parts for now. Instead of starting a new thread, I thought I'd simply add on / contribute to this one.

    As a start point, I put a gauge on the group head via the 18x1.5mm plug on top of the group and didn't get the values I expected.
    The boiler pressure is set to 1.5 bar but the group gauge only shows 1 bar.
    Also, when making a coffee the group gauge remains at 1 bar.
    From diagrams, I thought there was a check ball in the group and that gauge would read high(7 bar-ish?) during extraction but it remained flat at 1 bar. I even tested the gauge to make sure it was working properly.

    Am I misunderstanding the pressure at the 18x1.5mm fitting?
  • hi david

    1. check that your gauge reads zero when the machine is cold. if the anti vac valve on your machine has stuck at some point since 2016 a vacuum is created as the machine cools and this has enough force to cause the needle on the pressure gauge to turn on the centre pin as the vacuum tries to pull the needle below the stop. this then results in a gauge that reads greater than zero when the machine is off and cold, and overstates the actual pressure when the boiler is pressurised by the value of the overstatement when the system is cold. this will be true even if the anti-vac valve was subsequently cleaned or changed

    2. you only put a check valve in the dipper configurations of this group. if you put a check valve in with a thermosiphon configuration it is likely to result in the thermosiphon running cold as the machine ages/scale forms in the non return valve

    3. to get an accurate pressure reading you will have to bleed out every last cc of air between the water and your pressure gauge or the air compresses under load and acts as a shock absorber and soaks up the increase in water pressure, so you won't see it on the gauge unless you get all the air out

    kind regards

  • Hi Reiss,

    I checked the gauge and sure enough, It reads 0.1 bar high when the machine is cold.
    I can hear the anti-vac pop up when the machine is warming up but I'll probably pull it out and clean it to be sure.

    I also re-checked the group head gauge I was using and it's not accurate below 4 bar. Cheap gauge :(
    Grabbed another gauge, bled all the water out of the fittings and everything is agreeing.


    The lack of a check valve explains the pressure spike when releasing the lever mentioned earlier as the pressure will feed back into the boiler until the piston seal passes over the cylinder water holes. A quick release of the lever can bump the needle.

    Question, without a check valve and the back feed of some water from the piston to the boiler (or at least into the copper tubes) : Is there any possibility of coffee getting back to the boiler?
  • hi david

    no one has managed to do it yet, so i think we're safe

    kind regards

Sign In or Register to comment.