Simple Profiling on a La Marzocco GS/3 AV - Page 9

Need help with equipment usage or want to share your latest discovery?
rccoleman
Supporter ♡

#81: Post by rccoleman »

I'm still trying to get my proverbial sea legs with the wheel and I'm finding it a bit challenging. I haven't confirmed it yet, but it seems like the meaningful working range is pretty small on the needle valve, while the number of turns from stop to stop is much more than needed. That makes it harder for me to know, at any given point, where I am in the range and whether I'll need a small or large turn to adjust the flow up or down. I'm not sure how it would work on this valve, but I wonder if some sort of lock washer-style mechanism could restrict the range so that it's easier to identify when the valve is sufficiently open for full flow/pressure. As it is, I often find myself wanting to restrict the flow during the last bit of the shot and failing to hit it because I ended up at an unknown level of openness when starting the shot.

This is basically what I was afraid of without having the position of a paddle to directly indicate the flow rate. I know that I can play around with flushing and adjusting the wheel pre-shot to get the initial setting, but I find it harder to use during a shot where timing counts. I think part of the challenge is that there's no way to measure the flow rate at a given point in time, and the reading on the machine pressure gauge is a combination of factors like the flow rate, amount of water in the group, and resistance of the puck. I can see the value of a group-mounted pressure gauge to help reduce some of those variables. Is there a trick that I'm missing for accurate and repeatable wheel usage?

pcrussell50

#82: Post by pcrussell50 »

I have that valve. And while there are several turns available, you never really need much more than 180-210 past "closed" in order to hold a pressure between 7-9 bar for as long as the puck can stand it. So I start every shot with the valve either lightly closed, or barely open a crack. Over time, you will learn to hear when the headspace fills and you can predict when you will begin to see a pressure rise on your "surrogate" pressure gauge. My machine has group pressure so I get it right away, where you might have a slight lag. BUT over time, pressure itself becomes a surrogate for flow... a curiosity, if you will. On the odd occasion where I don't want to pre infuse, and I just want to whack the puck with nine bar as fast as possible, I do have to kind of guess at a starting valve position, and open the valve more than a crack. When I turn on the pump, pressure rises rapidly, and I close the needle quickly as the pressure needle springs to life. You will see the response, and have control over it. It all sounds overwhelming, I know. But it's actually easier done, than said. Keep at it.

-Peter
LMWDP #553

User avatar
Jake_G
Team HB

#83: Post by Jake_G »

I agree with everything Peter said.

Here's what I do:

I start my shots with the panel pressure gauge registering about a half bar of pressure, which is with the needle raised off the pin, but still inside the zero mark.

I choose to do this with a brief flush before pulling the shot, but you could quickly adjust the pressure at the beginning of your shot, too. Doesn't really matter.

I leave the knob alone for 10-12s. This is pre-brew, and the pressure gauge will do something during pre-brew that tells you a little bit about the puck. If you have a tightly ground puck with little headspace, the pressure will come up to 1 or 2 bar in those 1st 10s. If it is a coarser puck or has more headspace, the pressure could sit down low and not budge.

Knowing the headspace before hand let's you use this initial response to tell you how long to stay in pre-brew.
Tighter pucks will need more time in pre-brew than looser pucks, so a lot of headspace with a raise in pressure in the first 10s means a very tight puck that needs more PI.
A puck with less headspace that reacts slowly means a looser puck that could use a quick ramp to pressure.
And of course you have all the cases in between.

So, 10s into the shot in pre-brew and it's time to go to brew pressure. With my valve, I simply roll my finger from the back to the front. That's all it takes to initiate the rise to brew pressure.

As the pressure rises, I get ready to roll my finger back to "catch" the pressure before it spikes. I slow the pressure ramp and give the shot a gentle landing at whatever brew pressure I like. I generally grind fine enough that there isn't a single drop in the cup at this point. I then continue to monitor the brew pressure and slowly open the valve to maintain a steady pressure while the first 5-10g hit the cup. Sometimes, if the pressure wasn't responding at all during PI, I might "slam" the puck will full pump pressure instead of "catching" it, though the "catch" is the single most important part of controlling your shot. If you let the pressure spike, the pump bypass valve is in charge of the shot pressure instead of you, and getting back in control of the shot is tricky.

Once 5 or 10g is in the cup I let the pressure decay or even help it a bit as the shot progresses by rolling my finger back on the knob. I usually peak at 8 bar and decline to somewhere between 3 and 5 bar over however long the shot takes. When I'm ready to cut the shot, I put my finger on the front of the wheel and roll it back to close the valve entirely and then stop the pump.

After my shot, I knock the puck and rinse the group. I always end my shots with the needle valve closed, and I always open the needle valve till the pressure gauge shows 3 bar when I rinse the screen, and I always set the PI pressure to 0.5 bar before starting a shot. With all these constants in my daily routine, I am pretty much never hunting to find a specific position in the range of the needle valve, as it's all muscle memory at this point. There's really 3 positions that matter, and I modulate the flow around them.

Cheers!

- Jake

rccoleman
Supporter ♡

#84: Post by rccoleman »

Thanks, guys. One thing that I've been trying to wrap my brain around is that as I'm reducing pressure toward the end of a shot, I have to close the valve more than expected before I see any reflection in the pressure on the gauge. Often, I'll stop the shot from the front panel after having rolled the wheel pretty far toward the back (but not closed) with the gauge still showing 9+ bars. I'll knock the puck out, and just running a quick flush will show the gauge at 0-3 bar. So I clearly was affecting the flow, but it wasn't obvious by examining the machine or the shot. I guess that's just an example of the pressure being governed by the bypass valve and my not "catching" it before it spiked. Do you recommend setting the pump bypass higher to make it easier to modulate with the wheel?

User avatar
Peppersass
Supporter ❤

#85: Post by Peppersass »

I finally have all the parts to install my Jake Valve!

That includes 5 sets of the wheel and bracket, professionally printed (should have a set or two available when I'm done.)

The discussion of wheel position interests me a lot because my GS/3 has a gear pump with no bypass valve to protect it, and the expansion valve comes after the needle valve. What I was hoping to do is find the smallest opening of the needle valve that won't exceed 12 BAR at the gear pump (I have a pressure transducer there), then insert a small screw into one of the holes of the wheel that will prevent the wheel from rotating any further (same principle as the stop screw in the wheel of the tea water mixing valve.)

But this discussion sounds like there may be more than 180-degrees of rotation between a needle valve opening that's safe for my gear pump and a reasonable amount of free flow. I'm wondering if the screw stop I described will get anywhere near the stock 450-500 ml/min free flow with the gicleur in place. If not, I guess an option could be to remove the stock gicleur altogether, which on my GS/3 is really easy -- the newest TL-30 tube has the gicleur at the interface with the 3-way valve.

Speaking of the gicleur, I believe Jake has removed his gicleur. That could alter the flow dynamics enough to explain the difficulty other Jake Valve users may be having.

I'm also thinking my flow rate dynamics will be different because I'm using a variable speed gear pump with no bypass valve.

Fortunately, my Arduino interface to the GS/3 monitors the flow meter and reports the flow rate in ml/min. I've been using that to set the flow rate with my jury-rigged pressure regulator solution (soon to be made obsolete by the Jake Valve) and to achieve a constant flow rate during the back end of the shot. Hopefully, that will help me to avoid the puck slamming problem.

User avatar
Jake_G
Team HB

#86: Post by Jake_G »

rccoleman wrote:Thanks, guys. One thing that I've been trying to wrap my brain around is that as I'm reducing pressure toward the end of a shot, I have to close the valve more than expected before I see any reflection in the pressure on the gauge. Often, I'll stop the shot from the front panel after having rolled the wheel pretty far toward the back (but not closed) with the gauge still showing 9+ bars. I'll knock the puck out, and just running a quick flush will show the gauge at 0-3 bar. So I clearly was affecting the flow, but it wasn't obvious by examining the machine or the shot. I guess that's just an example of the pressure being governed by the bypass valve and my not "catching" it before it spiked. Do you recommend setting the pump bypass higher to make it easier to modulate with the wheel?
I think there is a tendency to think that full pressure requires a large valve opening, but that isn't the case at all. You only need to nudge the valve open enough for the pressure to start climbing, and it will climb all the way up to 9 bar or whatever you have set if you grind fine enough. I don't know that my valve is ever opened more than it is when I flush at 3 bar. In fact, quite sure that it isn't.

My bypass is set at 9.5 bar with a blind basket, and I generally "catch" my shots around 8 bar or so and hold them there. Since I am limiting the pressure to something less than my bypass pressure, the needle valve has instantaneous control over the shot pressure, but the only vent of pressure is through the puck and into the cup. I mention this because the pressure doesn't change on a dime with this. Its gradual, but you can control how gradual it is as long as you are limiting the peak pressure.
Peppersass wrote:I finally have all the parts to install my Jake Valve!
That's great!!
Peppersass wrote:But this discussion sounds like there may be more than 180-degrees of rotation between a needle valve opening that's safe for my gear pump and a reasonable amount of free flow. I'm wondering if the screw stop I described will get anywhere near the stock 450-500 ml/min free flow with the gicleur in place. If not, I guess an option could be to remove the stock gicleur altogether, which on my GS/3 is really easy -- the newest TL-30 tube has the gicleur at the interface with the 3-way valve.
There is another alternative for your implementation. You could simply run your gear pump at its minimum speed and adjust the needle valve to get you a very low water debit, then speed the pump up and see what the max pump pressure is. If the peak pressure acceptable, you would be using the Jake valve as a tuned gicleur to optimize the output of your gear pump. Not as sexy as full needle valve profiling, but then again, you have a gear pump, which is pretty sexy all by itself.

However, I don't think I ever use more than 180 degrees of rotation. It ends up working like the volume wheel on a clock radio and I give it one full front to back sweep and then a little more to return from closed after a shot to 3 bar on the panel gauge for flushing, which is plenty of flow for me. I have yet to feel the need to use more flow for any reason...
Peppersass wrote:Speaking of the gicleur, I believe Jake has removed his gicleur. That could alter the flow dynamics enough to explain the difficulty other Jake Valve users may be having.
One of the reasons I haven't removed it is so I will be living in the same realm as others. I don't find the control hard at all, but I only open the valve enough to get the pressure on an upward trajectory. There is no reason to open it further, but I think there is a desire at first to reach "wide open throttle" in order to reach full brew pressure. Once you overcome this desire and use your grind to ensure the puck is tight enough, that urge dissipates quite quickly.
Peppersass wrote:Fortunately, my Arduino interface to the GS/3 monitors the flow meter and reports the flow rate in ml/min. I've been using that to set the flow rate with my jury-rigged pressure regulator solution (soon to be made obsolete by the Jake Valve) and to achieve a constant flow rate during the back end of the shot. Hopefully, that will help me to avoid the puck slamming problem.
I'm excited to see how it works!

One thing we know for sure is that Slayer uses a needle valve for prebrew at the same peak pump output as full brew, so there is very little reason to assume there would be a problem so long as you don't dead-head the pump. Could you implement a pwm clamp in your code based on pump pressure? That way if you close the valve too tight, the pump will simply back off to avoid going over 12 bar or so... seems simple enough.

Cheers!

- Jake

rccoleman
Supporter ♡

#87: Post by rccoleman »

I think I'm seeing some positive results from my profiling experiments. I mostly use darker roasts (Malabar AU, Vivace Dolce/Vita) and I find that they don't benefit from long preinfusion (tends to make the shots more bitter), but they do respond well to a declining flow toward the end of the shot. Using an 18g dose and with the flow barely open, I start the pump and quickly roll the wheel open until the machine gauge reads 3-5bar. Pressure continues to increase, I wait until about 15g have accumulated, and then slowly close the valve, ending around 25g out. It's pretty easy to do and I think it's an improvement over the push the button and go strategy that I was using before.