Latest update: My flow rate problem has been solved, but not by fixing anything. After we tried just about everything short of disassembling the group head and replacing the feed tube, Chris Coffee replaced my GS/3 with a new one. The new machine consistently runs at about 450 ml/min.
Chris Coffee is, in my opinion, the best espresso equipment vendor in the universe.
For the technically inclined, some specifics follow.
erics wrote:What this shows is the contribution of the pump's bypass valve (amoungst other factors) during the course of any flow measurements. Hence my initial suggestion to remove/inspect the bypass valve for any obvious descrepancies.
After the low flow rate problem returned, I disassembled the bypass valve and inspected it. Nothing obvious wrong, but it would be difficult to know for sure if there was a defective spring or ill fitting part. At any rate, when I reassembled the valve, the flow rate was still low.
Chris Coffee then sent me a replacement pump, which included its own bypass valve. Replacing the pump had no effect on the flow rate. It remained at about 360 ml/min. I inspected the feed tube from the reservoir and the pump outlet tube, both of which were clear. This pretty much eliminated the water path from the tank to the tea water mixing valve.
I happened to plumb in the machine during this time, and this did not change the flow rate. But I didn't think it would.
We also tried a neat experiment devised by Chris Kramer to test John Blackwell's theory that a bad check valve on the tea water mixing valve was bypassing brew water into the steam boiler. We had discounted that theory because it seemed that there would be evidence such as changes in steam boiler pressure, reduced steam, etc. But I realized that the amount of water bypassed during one of my tests would be pretty small, on the order of 50-75 ml, not likely to noticeably reduce available steam. Also, my wife uses a lot of tea water, so the level in the steam boiler constantly gets reset throughout the day. Anyway, what I did was drain the steam boiler and disconnect the level probe and heater so that the boiler would be empty and not refill. Then I disconnected the tube that connects the suspect check valve to the steam boiler and hit the brew button. Not a single drop of water came out of the check valve.
It was after this that we decided the only remaining thing to do was disassemble the group head and inspect/replace the feed tube. Understandably, I didn't want to do that particular bit of surgery. It was expensive for both Chris and me to send the machine in for service, and the intermittent nature of the problem made it more than likely that a "fix" might be found, only to have the flow rate drop again later. That's why the decision was made to replace the machine. It helped a lot that I was willing to drive 3 hours each way to Albany to switch machines. Hopefully, Chris's people will find a definitive cause of the problem.
erics wrote:In addition to this, there is no information provided by anyone as to how this flushing flowrate is being measured, i.e. are we using stopwatches, laboratory beakers, multiplying a 10 second water debit by 6, etc., etc. Everybody can be using different methods and providing intentionally honest numbers.
I guess there could be significant variation between the measurements provided by various people. However, everyone who reported back to me had a flow rate in the 425-475 ml/min range. Further, and probably more significant, Chris Kramer at Chris Coffee has setup and measured the flow rate on at least 15 GS/3s since I reported my problem to him, and they were all in the 450 ml/min range. Presumably, Chris was using the same measuring equipment and methodology each time, which I'm pretty sure was the same as the method I told him I was using (remove PF, run for max brew limit of 50 seconds, divide resulting milliliters by 50, multiply by 60.)
Perhaps more significant are the variations in flow rate I saw. Since they covered a 120 ml/min range, there's little doubt that something was wrong with the machine. The odd delay in pressure ramp up time (it doubled), which was also intermittent and not strictly correlated with the flow rate variations, indicated a problem as well.
Problem aside, what's the right flow rate? I don't know how to reconcile all the reports of flow rates in the 450 ml/min range with the 325-350 ml/min design spec reported to me by Bill Crossland. Perhaps this is due to a different measurement regimen. Bill told me he was drawing a full liter and keeping track of the time, which is a little different than the way I was doing it. Perhaps LM Florence didn't get Bill's memo or changed the flow rate somewhere along the line? All I know is that there are a lot of GS/3s out there with flow rates in the 450 ml/min range, including my new replacement GS/3, serial number 613 manufactured 09/08. So far, the only GS/3s with flow rates under 400 ml/min that I know of are Bill Crossland's personal GS/3 and the one I took back to Chris Coffee. Oh, and then there's the guy in Australia with a flow rate of 180 ml/min, but that's another story, perhaps due to some very seriously hard water down under.
by Peppersass on Sun Oct 04, 2009 1:22 am 
. Most definitely, I believe that knowing your machine's flowrate during flushing is important and, once a reasonable, repeatable number is attained, it should be penciled in your book as an aid to future troubleshooting - hoping you never have to use it 
