I haven't gotten an answer to this one on Gs3cafe (which seems to be a very sleepy forum at times), so I thought a somewhat more detailed post here might get a response.

When I first got my GS3 about 3 1/2 weeks ago, the pressure would ramp up from 9 BAR to 9.5-10 BAR against the blind filter in about 3-4 seconds (for example, when doing a plain-water backflush.) At that time, I checked the GS3 pressure gauge with a home-made portafilter pressure gauge and the two gauges agreed on the pressure -- i.e., the built-in gauge appeared to accurately reflect pressure at the group head.

Starting about three days ago, the pressure has been ramping more slowly. Now it takes 6-7 seconds for the needle in the built-in gauge to start moving up from 9 BAR when the blind filter is used. Also, the first drops of espresso are taking longer to appear when I pull a shot. As I recall, it took 5 seconds or less previously, and now it takes closer to 10 seconds (varies depending on dose and grind, of course, but it's definitely taking longer.)

Checking the pressure again with my home-made PF gauge, the two gauges give me about the same max reading against the blind filter, around 9.5 BAR, but my home-made gauge is reading around 8.0 BAR with a 2ml/sec flow while the built-in gauge reads 9 BAR. The caveat is that I use a needle valve to adjust the flow rate on my homemade gauge, and it's hard to dial in the flow accurately. Nonetheless, it does appear to me that there's less pressure at the PF than there was before, despite what the built-in gauge is reading. I think the pressure may even be less than 8 BAR.

I tried adjusting the pump pressure up and down a half BAR or so from 9 BAR, but this doesn't seem to change the ramp up timing or pressure reading on my home-made gauge.

Has anyone seen a change like this? Can anyone come up with a possible explanation for the change? The only thing I can think of is that there's an obstruction near the group head that's restricting the flow.

NB, I do a manual water backflush every day and a programmed detergent backflush once a week. The daily backflush is done with the shower screen in place. The first couple of times I did the detergent backflush, I left the screen in place, but the instructions in the GS3 manual say it should be removed, so I did that the last couple of times. I was very careful to remove coffee debris from around the dispersion block before doing the backflush, but I suppose something could have gotten in there.

As for the impact on performance, the shots are just as good as before, but they do seem to be taking longer. The CHRONOS feature starts when the pump is activated, so I guess the extra few seconds of ramp up time are being counted.

Peppersass wrote:Has anyone seen a change like this? Can anyone come up with a possible explanation for the change? The only thing I can think of is that there's an obstruction near the group head that's restricting the flow.

I don't have a GS/3 to compare against, but what you describe regularly occurs to the La Marzoccos in Counter Culture's espresso lab. Their tech says they installed small diameter gicleur valves for slower pressure build-up and these valves need regular cleaning/descaling to avoid clogging.

Dunno where the gicleur is located on the GS3, but it is rather small at 0.6mm as i recall. A call to your local LM manufacturer should get you straight on that. If it has experienced scaling, then you have some pretty hard water and you should do something about it so that you don't need the educational experience of full-machine descaling.

-Greg

Did you the to bleed the group? Maybe a little air got in there, just a thought

Thanks for the replies so far. I've been using treated water in the range of 30 ppm hardness and 30-50 ppm alkalinity (TDS = 30 ppm), so I'd be surprised if there was scaling after less than four weeks. I did increase to about 70 ppm hardness a few days ago to experiment with taste, but again it seems odd that scale would build up so quickly with that level of hardness.

Be that as it may, is the gicleur accessible in such a way that I can get to it to clean it out without doing a full-blown descaling?

BTW, I did bleed the group -- first thing I thought of.

Follow-up:

I called LM USA today and spoke with John Blackwell. I described the change in ramp up time, and John's reaction was that there's nothing wrong. He said he understood what I was describing, but he didn't think it meant anything in terms of the machine's performance (I can confirm that -- the shots are fine. If anything, the slower ramp up may be beneficial.) He also said he knows of nothing that could cause the change. He said it was very unlikely that the gicleur is obstructed. If that happens, it's an all-or-nothing situation and he couldn't see how there could be a partial obstruction. He also thought scale buildup was very unlikely given the position of the gicleur in the flow path, the short length of time the machine has been in service and the low level of hardness/alkalinity in my treated water.

John though it might be possible that the change has something to do with the pump breaking in. I didn't completely follow the explanation, but he said excess pressure circulates through the pump and that the relief mechanism might change slightly as the pump breaks in, resulting in a slightly increased ramp up time. He said he would check a brand-new GS/3 and one that has been in use for a while to see if there's any difference in ramp up time. He promised to call me back with an answer.

More follow-up:

Check out this post on Gs3Cafe: Slow waterflow out of the brewgroup. Also see my reply.

During our call, John Blackwell advised me to test the flow rate of my GS/3. He said it should produce 500 ml per minute, +/- 10% (presumably with an empty basket or no PF in place.) I did the test and found that the machine is producing about 150 ml in 50 seconds with no PF in place, and about 175 ml in 50 seconds with an empty basket. Of course, the firmware cut off the pump at 50 seconds, so I couldn't run the test for a full minute. But the calculated rate is 200 ml per minute, less than half what John said it should be. I've emailed him the results of the test, and queried whether my habit of doing a plain-water backflush with the (dirty) dispersion screen in place may have clogged the water path with coffee debris. I guess It's either that, some other obstruction, or a bad pump.

I'd appreciate it if one or more of you GS/3 owners out there could test your flow rate and let me know what it is.

Peppersass wrote:During our call, John Blackwell advised me to test the flow rate of my GS/3. He said it should produce 500 ml per minute, +/- 10% (presumably with an empty basket or no PF in place.) I did the test and found that the machine is producing about 150 ml in 50 seconds with no PF in place, and about 175 ml in 50 seconds with an empty basket. Of course, the firmware cut off the pump at 50 seconds, so I couldn't run the test for a full minute. But the calculated rate is 200 ml per minute, less than half what John said it should be. I've emailed him the results of the test, and queried whether my habit of doing a plain-water backflush with the (dirty) dispersion screen in place may have clogged the water path with coffee debris. I guess It's either that, some other obstruction, or a bad pump.

I'd appreciate it if one or more of you GS/3 owners out there could test your flow rate and let me know what it is.

My machine produces 370ml in 50 seconds, before the automatic shutoff. This equates to a flow rate of around 445ml per minute.

One suggestion regarding chemical backflushing: I was originally instructed, with my Cimbalis, to begin with several short water backflushes, then to rinse the blind filter with water and put the detergent into the pf. I always run the machine a few seconds before reintroducing the PF with the detergent in it, back into the group. I was told to then backflush 2-4 times with the detergent in the blind filter, until a lot of detergent shows up in the discharge water.

THEN, you should wait a few minutes, at least 2 or 3 minutes, to allow the detergent to do its job inside the machine. Only after that would I use the automated backflush cycle (or finish it manually).

I'm new to this machine also. It could be that (as Dan suggested) the gicleur is small enough that this machine is very sensitive to grunge accumulating in the water path. If so, it is also possible that it will require more frequent chemical backflushing than some other machines, perhaps needing a chemical backflush a couple of times a week (in low volume home use) as opposed to other machines needing this less frequently.

My impression is that something is wrong with the functioning of your machine based upon your description. I'd be amazed if it was scale. It could well be that there is just a little bit of grunge in the water path, or possibly something more serious such as the early failure of the rotary pump. I would start with a chemical backflush, and in this case I'd do 2 to 4 backflushes with detergent, then let the machine sit for half an hour, then finish the backflushing. If there is some coffee residue in there gumming up the water path, this should clear it out. If there is no improvement then I would get on the phone with your seller, Chris Coffee, and have them sort it out.

ken

Got a reply from John at LM USA. He confirms the approximate flow rate should be 500 ml per minute, +/- 10%. He did a test on a new GS/3 and it ran 400 ml in 50 seconds before the cutoff. Ken's flow rate is within spec, mine isn't.

John is guessing that water destined for the brew group is partially being bypassed backwards through the check valve, B.2.008, between the hot water mix block B.2.007 and tube T.2.105 (page 6 of the part manual, for those of a technical persuasion.) Below is a picture of the part in question, labeled "J".



John thinks something may be blocking the check valve open or the part has prematurely failed. I've asked John how to proceed. I'm guessing drain both boilers, remove the check valve, clean/clear it, and see if that fixes the problem. If not, replace the check valve. I've also asked for tips on getting the check valve out. Judging from the picture, it doesn't look too hard, but you never know.

On backflushing, I've also thought that with the automatic cycle you don't get the benefit of the detergent sitting in the machine for a few minutes. I like Ken's suggestion to do a manual detergent backflush for that purpose, followed by the automatic cycle (plus another cycle with clean blind PF to clear out the detergent.)

BTW, based on a post on Gs3cafe, I've been conservative with the amount of detergent: only 1/2 tsp instead of the full tablespoon recommended in the manual, which seemed like an awful lot to me. Any suggestions on that, Ken?

In any case, whether or not there's gunk in the check valve or anywhere else in the water path, I'm not going to do any more backflushes, plain or chemical, with a dirty shower screen in place! This will probably lead to cleaning both sides of the shower screen more than once a week. Seems like a good idea to backflush with a clean shower screen in place, rather than without the shower screen as instructed in the manual, to avoid any possibility of solid particles getting into the water path.

On pressure being different at the group head, that doesn't surprise me. As I said, my method of adjusting the flow rate on my home-made PF gauge probably isn't accurate enough. The gauge is more useful for measuring blind pressure, and there it agrees with the built-in gauge.

I guess the real issue all along has been flow rate, not pressure. This probably explains why the shots haven't been adversely affected, just the timing.