Something to mention before I go into detail. I have not used a pressure scace on my machine so I don't know the true output pressure.

With that being said, I have used my expansion valve to adjust brew pressure. I had it set at 9.5 with a blind which would give me between 9.0-9.5 while brewing a double. I ended up watching a video where Luca Bezzera talked about 9 bar only being the correct setting for rotary vane pumps because their flow rate is much higher. He stated that they flow at about 100 L/hr where vibe pumps usually only flow around 34 L/hr, so 11 bar is needed for vibe pumps to ensure proper pressure at the puck. Any disturbance in flow with a vibe pump will adversely affect flow so a higher pressure is needed. I calculated mine and I got 39 L/hr and that makes sense because I have the larger Ulka than what he was using in his demonstration.

I would also like to say that I have confirmed his theory about flow. When I had my machine set to 9.5 bar, it didn't use nearly as much water as it does now that I set it to 10.5 bar on the gauge. I believe this to be because the expansion valve recirculates more water to the tank when the pressure is set lower. The tank runs out of water markedly faster than before. I turned the expansion screw all the way clockwise to see what max pressure is and it won't register above 11. I backed it off to 10.5 on the backflush disk and I believe my shots are actually coming better. Not only is there better taste but a nicer crema.

I'm willing to bet that the puck probably is seeing 9-9.5 bar of pressure now due to two things:

1. The brew pressure is always lower with a puck than at zero flow against a disk.
2. The manometers on these machines are probably not accurate.

I won't drop the money on a scace to really find out because I am enjoying what the machine is doing now.

I noticed that when I had the machine set at 9.5 it took a bit longer to ramp up to brew pressure and that the preinfusion seemed to flow right through. Since now I can pull excellent shots, I feel what Luca Bezzera says is accurate. I have read many threads on here discussing 9 bar being proper on any machine and I have to disagree respectfully. I think AT LEAST 9 bar or more is required.

I should also add that I would love to hear discussion on this topic as everyone just seems to go along with the whole 9 bar thing.

I'm going through this with my machine, an andreja premium. I was not happy guessing at the temp, so I bought a scace2 gauge. It's expensive but worth it. I was not really thinking of the brew pressure at the puck until I got the scace2. And vibe pump pressure has always been in my head, but it didn't click until your brought it up. That vibe pumps need to be closer to 11 bar. I'm at 9 with the scace2 and close to 12 on the machine. I need to turn it up.

Gj91. I would think as long as you're at 9 with the scace2 at the group head you should be ok. With my expansion valve fully closed, I could only get 11 bar on the manometer anyways which would lead me to believe yours will probably not go much higher than it is now. After all, the pressure at the brew head is what counts and I believe when adjusting brew pressure they were using the gauge on the machine. Maybe I'm wrong, but more importantly, how are your shots?

The say a rotary pump is internally regulated so pressure against the puck isn't as important. With the vibe pump, setting it to 11ish on the manometer should give 9 at the puck (or so I'm lead to believe).

On a side note, I closed my OPV fully and there is still a good amount water which flows back to the tank. This means that these OPV valves are set to either go to 11 bar or not much higher. I thought for a while that my gauge wasn't working properly but now I believe it's working just as well as it should. So in short, there will always be flow from the OPV back to the tank, but how much depends on how high you set your brew pressure. Like I said, my water consumption went up when I turned my pressure up. That can only mean that flow increased.

I want to update this thread to discuss some results I had when lowering pressure on the manometer when pulling a ristretto.

It pulled a much nicer tasting shot with the OPV valve set to 9.5 against a blind and 9 on brewing. Not only was crema nicer but flow was reduced so it allowed more contact time with the coffee. I'm probably going to leave it here for now but its an easy turn of the screw to go back up to full pressure if I need to. It was nice to be able to keep the pressure high for regular shots.

all is about solubility (Ben Kaminsky); and they say, many decades ago, that at 9 bar, there is a peak in extraction; so it could be a good thing to have extraction pressure to peak at around this value;
from above the puck all the way to the OPV output, there is a pressure drop of 0,5-1 bar; and yes, the pressure gauge could be wrong

about the difference in working pressure between vibe and rotary pumps, this is of no effect during a good extraction: the pressure is suposedly the same 9 bar, the hydraulic resistance of the puck is the same, so the flow is the same; the extraction volume should not depend on the type of pump;
having the same flow through the puck, the excess flow is returned back to tank by the OPV in the case of the vibe pump, or recirculated via bypass inside the rotary pump;

Luca Bezzera could be right that in my machine (a Bezzera Magica), 12 bar would give a better cup than 9 bar; but only in theory; in practice, at 12 bar, the extraction is more prone to channeling and less forgiving; I have set mine to 9.5 bar on the gauge; I do not have a machine to do pressure profiling...

there is another rationale for 12 bar and an ULKA EX5 pump: at this pressure, the pump flow is at about the nominal value for a normal extraction: 2-3 ml/sec; at 9 bar, the flow is double in value (cca. 260 ml/min) so, at least for normale and ristretto, the OPV will open, to limit the pressure at 9 bar;
because for a vibe pump with OPV, a higher flow resistance would reduce the flow and increase the pressure to the limited value (with return flow), and viceversa, a lower resistance (channelig) would increase the flow and lower the pressure under the limited value (no return flow); as opposed to a rotary pump which has normaly enough flow to mantain its 9 bar set at bypass;

I do not see how you have different water consumptions in extraction with 9 and 12 bar settings, as the extracted volume of espresso is the same, and the excess is returned into the tank? and, by its characteristic flow/pressure, a vibe pump has more flow at lower pressure and viceversa; so, yes, at 9 bar there is more water returned into the tank than at 12 bar;

Radu, while brewing, the water consumption is almost the same, but when doing a cooling flush the water consumption went up markedly. What happens with that opv valve is once set on a lower bar setting, it lets more water return to the tank, and that's how it adjusts for pressure. If you crank it all the way up, more water is sent to the group head and less back to the tank. It doesn't really adjust pressure, it adjusts flow and pressure is incidentally increased when bypass flow is decreased.

Those opv valves are basically just an in line flow regulator so the tighter you make them, the less flow they divert. Even when just doing a cooling flush, I can see there is less water from the group head at a lower setting.

well, the OPV is limiting the pressure, that is, at a certain value as set, the valve opens abruptly and all the excess flow is diverted, so the pressure remains constant;
the OPV is not controlling the flow of the pump, it is limiting the pressure of the pump; the flow depends of the pressure and the hydraulic resistance of the circuit;
the pressure of the vibe pump is related to the flow, as I said: higer pressure - lower flow and viceversa;
OPV is not directly related to the flow: when hydraulic resistance is high, the flow decrease, the pressure increase, usually to the limiting value as set at OPV; with an ULKA EX5 this happens during extraction of a normale, or ristretto;

when OPV is open, the flow of the pump is higher than the flow towards the group;
as said, at 9 bar setting of the OPV, the flow of the pump is, say, 4.5 ml/sec, from which 3 ml/sec are returning back into the tank;

as for the flush, yes the flow is of course higher as the pressure is way below the usual setting for the OPV;
OPV remains closed, and the pressure decrease at a value corresponding to the pressure drop along the hydraulic ciruit open during flush; so, normally, during the flush, there is no influence of the OPV to the flow;
considering two different setting for the OPV, during extractions there is the same flow to the outside in both cases; and during the flushes, the OPV remains closed, and the flow/water consumption is the same;
shouldn't be any difference;
maybe there is an OPV problem in your case;

Radu,

I don't think there is a problem with my OPV. I even tested to see how fast the free flow from the group into a cup was. With the OPV closed, flow was faster into the cup doing a cooling flush. Putting the OPV back to 9.5 bar caused the machine to take longer to fill the cup. Even the manometer went from a 3 bar to 2 bar under free flow.

I am willing to bet that the OPV valve is always open to some degree when the pump is pushing water through the group (puck or no puck), and the lower you set it, the less flow you will have through the group head. I'll confirm this when I look at the return hose while the pump is running next time I use the machine.

The only reason flow is reduced under high pressure is because the outlet is being obstructed by some obstacle. That is exactly how the OPV valve is working; as a controlled partial obstruction.

So to analyze what was happening when I increased the max pressure of the OPV to 11 bar, at free flow the OPV valve was opening less or not at all, while at 9.5 bar max pressure on the OPV, it was opening slightly to allow some water to return to the tank. This would correspond with the 3 bar cooling flush with the higher setting vs. the 2 bar cooling flush with the lower setting.