Dream vs GS/3 Throwdown

Forums below are for discussions of Dream and Dream T specific issues. However, both machines have as much, if not more, in common than they have differences. This area is for information sharing and questions that you think apply equally to both machines.
Post Reply
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Dream vs GS/3 Throwdown

Post by chas »

Today I expect to complete the cleaning and overhaul of my GS/3 that having the Dream has given me the time to perform without giving up espresso for a week. With a little counter top rearrangement I should be able to put the two machines side-by-side set up with the same boiler temperatures and pressures, using the exact same coffee, both plumbed in and see how they compare. I hope to be able to report the results of those tests this weekend - both for straight shots and milk drinks.

After that, I probably still have at least a week left before the loaner 53mm Scace device arrives. While I wait for that I expect to go through all the menu options on the Dream, photograph and document their operation. I am pondering a few ways I might present that data for optimal clarity though I will probably need to wait until I have all the data before I finally decide. Initially, I'll post this somewhere on the Forum but I'll also provide it to Jason at CCS in case he wants to add it to his official CCS User's Manual.

As I work through that, should I make any interesting discoveries, I'll also be sure to post them here.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
etout00

Re: Dream vs GS/3 Throwdown

Post by etout00 »

This sounds very interesting Chas, looking forward to your results as I have been going back and forth with keeping the Dream or getting the GS/3. I also sent you a PM regarding some specific questions.
Thanks
Dan Bollinger

Re: Dream vs GS/3 Throwdown

Post by Dan Bollinger »

I'm looking forward to this test. It will be interesting to see if double the money buys you better coffee.
PyroKerry
Macchiato
Posts: 38
Joined: Thu Nov 08, 2012 3:37 pm
Location: Illinois

Re: Dream vs GS/3 Throwdown

Post by PyroKerry »

Chas,
Do you use the VST filter baskets with your GS/3?
La Spaziale S1 Dream
Mazzer Mini E
Bunn Trifecta MB
Baratza Virtuoso (Preciso burrs)
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

Yes I do.

Sent from my DROID RAZR using Tapatalk 2
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

Let the throw down begin.....
Let the throw down begin.....
ThrowdownJPG.jpg (90.85 KiB) Viewed 50610 times
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

The first throw down round is over and I am not quite sure what to make of the results. We did four pairs of shots. The difference between each was whether 4s preinfusion was on or off on each machine. I had three very amateur taste testers. Overall, the GS/3 won on three of the four rounds but one taster actually called it for the Dream on two of the four rounds.

None of us are big straight espresso drinkers so anything less than a God shot was a challenge and we had nothing close. I think I probably need to make some adjustments and redo the tests again in the future.

1) To my surprise, the Dream seemed to do better with preinfusion off.
2) Since the tests were done with preinfusion On/On, On/Off, Off/On, Off/Off. That means that there were two shots drawn from each machine with preinfusion On and two shots with preinfusion Off. The matching pairs of shots would be expected to get rated pretty similarly, but they weren't. The ratings were all over the map.
3) Even though both machines were set to 93C the Dream was rated as sour to very sour much of the time compared to the GS/3 which leads me to believe that either the cold nose was coming into play or the machine is not accurately set to really deliver 93C water. I used a 58mm Scace to tweak in the GS/3 before the test, but the 53mm Scace hasn't arrived yet. So I am just trusting that CCS adjusted the temp offset on the Dream as accurately as I did on the GS/3, On the other hand, while the GS/3 shots tasted better some of them were not that great either so I think I need to also adjust the temperature higher on both machines for this coffee.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
etout00

Re: Dream vs GS/3 Throwdown

Post by etout00 »

chas wrote:The first throw down round is over and I am not quite sure what to make of the results. We did four pairs of shots. The difference between each was whether 4s preinfusion was on or off on each machine. I had three very amateur taste testers. Overall, the GS/3 won on three of the four rounds but one taster actually called it for the Dream on two of the four rounds.

None of us are big straight espresso drinkers so anything less than a God shot was a challenge and we had nothing close. I think I probably need to make some adjustments and redo the tests again in the future.

1) To my surprise, the Dream seemed to do better with preinfusion off.
2) Since the tests were done with preinfusion On/On, On/Off, Off/On, Off/Off. That means that there were two shots drawn from each machine with preinfusion On and two shots with preinfusion Off. The matching pairs of shots would be expected to get rated pretty similarly, but they weren't. The ratings were all over the map.
3) Even though both machines were set to 93C the Dream was rated as sour to very sour much of the time compared to the GS/3 which leads me to believe that either the cold nose was coming into play or the machine is not accurately set to really deliver 93C water. I used a 58mm Scace to tweak in the GS/3 before the test, but the 53mm Scace hasn't arrived yet. So I am just trusting that CCS adjusted the temp offset on the Dream as accurately as I did on the GS/3, On the other hand, while the GS/3 shots tasted better some of them were not that great either so I think I need to also adjust the temperature higher on both machines for this coffee.
Very interesting Chas. I am very curious to see what your 53mm Scace results show as all our machines are set to a 0 offset, which I've been questioning since the beginning.
etout00

Re: Dream vs GS/3 Throwdown

Post by etout00 »

Since you are doing a side by side comparison, is there anyway to compare the sound levels of the machine as well? Maybe a quick video or audio file? Or if you have a smart phone there are free sound meter apps that read out dB.
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

etout00 wrote:Since you are doing a side by side comparison, is there anyway to compare the sound levels of the machine as well? Maybe a quick video or audio file? Or if you have a smart phone there are free sound meter apps that read out dB.
The Dream averages 70dB and the GS/3 73dB. When the GS/3 came out it was roundly criticized for its noise level. For the last couple of years they've been shipping models with a new motor mount design. So I would expect new GS/3s to be about the same as the Dream.

* Readings measured by the SmartTools App on my Android phone.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
timbo

Re: Dream vs GS/3 Throwdown

Post by timbo »

This is a great comparison, though I would expect very very close results. The interesting thing i've noticed about my spaz is the narrower basket really limits the maximum dosage to be in the 16g range. This is much smaller than the LM, even before going to triple baskets. Though I believe the shots on the spaz are as '3rd wave' as any 58mm machine - plus you save a bit of coffee. I'd be interested in seeing different dosages (and the same dosages) on the different machines - though i expect you'd need to grind differently for each machine as the puck height is obviously not the same.

Lovely setup you have there, btw!!
Dan Bollinger

Re: Dream vs GS/3 Throwdown

Post by Dan Bollinger »

Chas, I agree that it is too early to tell anything significant from this initial test. I'm going to chalk this up to improving the testing methodology. As I see it, the influencing factors that need to be dealt with are:
  • Amateur tasters
  • Not using blind test
  • Machines not yet adjusted properly to make a good shot
  • Small population, n=8
No surprise there. I'm sure you came up with the same list, too. I look forward to your findings.
User avatar
slo
Barista
Posts: 419
Joined: Mon Aug 31, 2009 6:51 pm
Location: Montreal, Canada

Re: Dream vs GS/3 Throwdown

Post by slo »

Thank you Chas for this.
I am eagerly reading and waiting for more on this.
The GS/3 has been an option in my home for a while because my wife likes the looks of the machine better! I didn't share any information on pricing! ;-)

I was and still remain skeptical that the differences will ever be such as to allow declaring that the GS/3 is worth a few more $K. The diminishing returns will apply here. But I am very interested to get a comparison of the 2 side by side. Side by side comparison of espresso machines are so rare. We always have to somewhat rely on the comments of new owners who rarely run the old and the new machine.
Vivaldi II, Multiple (a collection really) Lever machines
Currently on deck grinders: Mythos and MXKR
Backup grinders: Robur, Major.
Toper Cafemino Electric and Poppery 1 roaster

I have a serious problem ... Can you guess what?
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

I got a really good deal on my GS/3 but three years ago it was still $2K more that a new Dream. There is no doubt that the GS/3 is a nicer machine in a number of ways so it comes down to whether it is worth the extra money. I currently have both on my counter and am using the Dream extensively for testing and learning purposes, but I still gravitate regularly to using the GS/3. Make no doubt, the Dream really closes the gap over the previous S1 models. With all the programmable features of the new Dream it comes close to the programmability of the GS/3. There are definitely some GUI kinks which I'm sure, at Chris' urging, LaSpaziale will get corrected. The GS/3 has a few GUI issues of its own that will probably never be changed but they're not a big deal.

The main things I like about the GS/3 over the Dream are the external mechanical design items.
1) Much more clearance between the group head and the drip tray allows me to put fairly tall cups right under the group. The Dream as well as previous S1's forces you to use shot glasses and then transfer the espresso to the desired cup. This is the standard old school way of doing things so this might not be an issue with a lot of folks, but I like to be able to cut out the shot glasses and go straight into the final cup.
2) Steam wand comes out of the top of the machine allowing great flexibility as to the size and placement of the container for steaming. You can use a large mug and steam milk directly into the mug for hot chocolate as well. On the other hand while I don't have this level of flexibility with the Dream, I haven't really had any issues with it either. Even if the Dream just had the same S5 steam arm I previously had on my older S1s that would be a big improvement. It was just a little over 1" longer and with a slightly different angle which makes a lot of difference.
3) Hot water. The GS/3 hot water wand is similar to the steam wand in design so it provides lots of flexibility as to the size of container you can dispense the water directly into. At least the hot water nozzle on the Dream pivots, which is an improvement on the previous generations of S1. The GS/3 has a mixing valve so that cold and hot water can be mixed to obtain the desired hot water temperature. I like mine about 185-190F. On the Dream you're only option is pretty much water just off boil.

In general if you don't have a GS/3 or better machine now, you may not have the above items and your don't miss them. Are they worth doubling your investment?
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

I received a UPS tracking number a week ago yesterday for the 53mm Scace loaner that is coming from Portland. However, it doesn't look like it actually got into UPS hands until yesterday. So I should be able to start the temperature profiling next weekend or early the following week.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

The loaner Scace arrived Friday. However, the throw down has hit a snag. The Scace reveals that the water temperature of my Dream is way low, far outside the range of the offset to correct. I suspect I have a bad group boiler temperature sensor. Hopefully, I can get one shipped to me on Monday unless Jason thinks it is something else.

I did discover that the -8C to +8C offset range of the VII has been replaced by a range of 0-4C or 0-8F.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
jfrescki
God Shot
Posts: 100
Joined: Tue Nov 20, 2012 9:30 am
Location: NJ

Re: Dream vs GS/3 Throwdown

Post by jfrescki »

Any update on the brew temp problem?
John

Mini Vivaldi, EPC-8, SJ, Steel Burr Vario
Behmor, Modded Poppery
Coffee Stuff
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

I think the range of the offset needs to be expanded. 0-4C or 0-8F probably isn't enough - at least given the way the machine works at the moment. However, there is a bigger issue that I discovered which Jason at CCS confirmed within the last couple of hours. Offset just doesn't workright. It interferes with properly heating of both boilers if the offset is any value other than zero. Neither 1.07 nor 1.14 are right but the problem is worse in 1.14.

For example, if you use the Scace and note that when you set the coffee boiler temp to 93C you really get 91C group water temp, then you should be able to change the offset to +2C. That way the boiler actually heats to 95C according to the internal temperature sensor but the front panel display shows 93C, the water temp dispensed is 93C and everybody is happy. If you do that now, the group boiler doesn't reheat as quickly as normal and the steam boiler completely shuts down. However, if you set the offset to 0C and the boiler temp to 95C you still get the 93C water you want and both boilers continue to work properly. So it's definitely a software issue. No one has seen this until I started playing with mine because all the units are shipping with the offset set to 0.

All the offset really needs to do is show a delta temperature on the display versus what the temperature probe is displaying so that you can adjust for the LaSpaz being a "cold nose" device and for the fact that there will be some internal temperature drop between the heater element and the top of the puck. I am not quite sure how you screw that up such that it impacts the overall operation of both boilers, but obviously it's possible. :bom:
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
richardcoffee
God Shot
Posts: 172
Joined: Tue Dec 07, 2010 4:59 pm
Location: Michigan City, Indiana

Re: Dream vs GS/3 Throwdown

Post by richardcoffee »

Does this affect both boilers when operating in the 15 amp mode? Or does it matter?
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

I can check but so far I am running in 20A mode so I don't know.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
richardcoffee
God Shot
Posts: 172
Joined: Tue Dec 07, 2010 4:59 pm
Location: Michigan City, Indiana

Re: Dream vs GS/3 Throwdown

Post by richardcoffee »

I had a theory that the problem would have been confusion only when running in the 15 amp mode. Apparently that's not the case.
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

I'm starting to see a new issue that may be related to the temperature problem you've observed. Up until a few days ago, the displayed brew temp when idle corresponded with the setpoint. In other words, my brew temp is set for 92 C, and the display shows actual temp at 92 deg C when idle. But a few days ago, the displayed temp has risen to 96 C when idle. I haven't changed a thing in the settings, but just to make sure, I rechecked by setpoint at 92 C, and I check that the offset was still set to 0 in the factory settings. Not sure whats going on here. I don't have a scace, so I can't tell what the actual temp is at the brew head.
Bob
La Spaziale Dream and Baratza Sette 270W
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

Just as a followup to my last post, this may very well be a software issue. Yesterday, the brew head temp as reported on the display drifted high. In other words, over the course of a few hours, with a setpoint of 94 C, the displayed temp went as high as 106 C. I tried rebooting the machine by powering off and on. I also went into the factory settings and chose the "reset" option to reset the eeprom. Neither attempt worked and it continued to drift and would not hold the setpoint. I also tried flushes etc. My machine is set to turn off at 11 pm and come back on at 5 am. This morning everything is back to normal and it is holding to a steady 94 C. So, perhaps this is a software glitch that caused bad control action until the machine went through a cool down cycle. As per my previous post, I have not touched the PID settings at all. One other thing that I am observing is that there is a wavy line under the brew temp that illuminates to show that the brew boiler heater is on. This shows correctly when you pull a shot or first turn the machine on. But if the machine is just sitting idle, then you don't see the wavy line even though the heater should be turning on occasionally. In other words, at idle, with a 94 setpoint, when the temp drops to 93, you should see the wavy line momentarily as the heater comes on to return it to 94. You see the temp return to 94, but you never see the wavy line. I will cross post a reference to this in the 1.14 bug report thread.
Bob
La Spaziale Dream and Baratza Sette 270W
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

When I used the Offset feature and noticed that the wavy line stayed on all the time, I did not believe that the group heater was coming on and staying on. So I got out my voltmeter and confirmed that the group element was really cycling on and off at a fairly quick rate - perhaps on the order of 4 sec off and 1 sec on. What I did not do, but will put on my to-do list - is to check this with the offset back at 0 so that the wavy line is turning on and off and see if there is any correlation whatsoever between when it is on and when the group heater is actually on.

My GS/3 has two side-by-side rectangles on its front panel display that pulse to black, then clear indicating when the group or steam elements are on. They pulse on and off rapidly like I saw when I checked the Dream using a voltmeter. Frankly the duty cycle on the wavy line is very slow, much less than I'd expect on a PID'd machine so it will be interesting to check this in more detail. I may actually need to use my oscilloscope as the response time of the voltmeter is a little slow to see exactly what is going on.

I am hoping that this is not a serious issue with the real-time programming necessary to keep up with these rapidly occurring events, especially if they also implemented the PID in software rather than using a dedicated, off-shelf HW PID.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

I have not experienced the "temperature wander" incident again. Temp readout is stable. So it really looks like a glitch in the PID algorithm or somehow something got corrupted then cleared itself. Very strange.
Bob
La Spaziale Dream and Baratza Sette 270W
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

I am almost ready to commence the throw down. With the latest Dream SW ver 1.17, the capability is now there to adjust the shot temperature and offset properly in order to achieve a good result. I spent time yesterday re-learning how to use the download and plot feature of my Fluke digital thermometer. So I should be ready to gather all the data today. Although it may take me into tomorrow to create and label the various plots.

One thing I noticed is that the Offset for the Dream can be set as high as +6C or +10F depending on which mode you are operating in. However, +6C is really equivalent to +10.8F. So I may duplicate the tests in both C and F modes to see whether the +10F offset is really +10F or whether it is actually +10.8F (assuming the results are stable enough to measure this). What I am not sure of is whether LaSpaziale actually measures directly in the chosen scale or whether they always measure in one and convert to the other. Testing both C & F modes separately might clarify that.

As noted previously, I will duplicate the tests on my GS/3 with it set as closely as possible to the same 94C output temp to see which comes closest to 94C and which machine appears to do so with the flattest, most stable profile.

Starting with the Scace on both devices in place for a few hours and the machine idle during that entire time. I will do 5 back-to-back tests. This will show for both machines how many warm-up shots it takes to get the most stable result and, once there, how stable the output is when continuing to pull back to back shots. As CCS does when initially adjusting the machines, I will remove the Scace between each run for the equivalent time it takes for the boilers of each machine to turn back off and/or for the time it would take to grind and tamp coffee for another shot whichever is the longer of the two times.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

Below is a PDF file that contains 10 graphs for just the Dream. I am about to gather the data for the GS/3 as soon as I submit this post. Starting with the Dream being idle but with the Scace device installed for a little over an hour. I took 5 back-to-back measurements with about 20 seconds between each measurement during which I removed the Scace device. This emulates removing the PF to grind and tamp coffee and give the boiler element a chance to cycle and come back up to temp.

For each of the five measurements, the Dream was set to 94C and a +6C offset. I then let the meter software convert to Fahrenheit for comparison since many of you use F rather than C. This is why there are 10 plots in the file. For each run there is a plot in Celsius followed by a converted version in Fahrenheit.

In the next day or two I will repeat the measurements with the machine set to 201F and +10F offset for comparision in case there are significant differences between C and F modes of the machine. (None are expected but you never know until you check!)

One thing I have noticed is that the Dream keeps two sets of books relative to C and F modes. for example 94C = 201F. If I set my GS/3 to 94C and switch to Fahrenheit mode it will display 201F. This is not necessarily true of the Dream. Rather it will remember what value you used the last time you were in that mode and return to that value. So if I am running in F mode at 199F and then switch to C mode and change the setting to 94C, then later go back to F mode, the Dream will go to 199F not 201F. The same is true for the offset and also the steam boiler setting.

There are two conclusions I make from these plots. 1) As with the original V1 and V2 models, you need two warming flushes if the machine has been idle more than a few minutes. Starting with the 3rd pull and all those that follow, the temp vs time plot is reasonably flat. From start to finish of each shot from #3-5, the temp rapidly levels off and then increases by less than 1C over the duration of the shot. 2) at least with my machine +6C offset is not quite enough. I really need +7C. For my machine if I want the water temp over the duration of a shot to not only be pretty flat but also average 94C, I need to set the machine to 95C with a +6C offset since I have no +7C offset.

Jason at CCS says that the machines he is bench testing are coming out spot on with the offset at 0C. So we are trying to compare notes to see why we are getting such different values. We have both checked the calibration of our Scace devices and believe them to be accurate.
DreamTempProfile3-3-13.pdf
(70.81 KiB) Downloaded 524 times
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

The file below contains the 5 graphs for temperature profile of the GS/3. The basic test procedures are the same as for the Dream which was covered in the previous post. There are only 5 because I did not duplicate the Celsius measurements in Fahrenheit. For comparison, the GS/3 uses a -2C offset.

Like the Dream, the profiles are much flatter after two warm-up shots. The difference, however, is that the average temp for the first two shots still averages 94C even though there is more variability in beginning and end temps that in shots 3-5. So you really could pull shots without doing warm-up shots and you may not taste a difference in the cup.

If you compare shots 3-5 from both machines just based on the shape and flatness of the plots, they seem to hold within 1 degree once the temp has ramped up and flattened out. the main difference is that I can't really set the Dream to the desired temperature and have the offset range needed to tweak it in.
GS3TempProfile3-3-13.pdf
(46.33 KiB) Downloaded 483 times
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

FYI, for those thinking about buying the S1VII rather than the Dream, the file below contains the temperature measurements I made several years ago when I had the VII. This was before I owned the logging digital voltmeter so this is just a PDF version of an Excel spreadsheet I created which allowed me to optimize the offset to -2C for that machine.
VIITempMeasurements1.pdf
(13.19 KiB) Downloaded 634 times
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

Chas: I know you are in contact with Jason at CCS about this, but I am wondering how a machine with an improved location and design of the temperature sensor needs a bigger offset than the previous version of the machine. In the back of my mind is my problem with the brew temp sensor. I wonder whether there is more to this?
Bob
La Spaziale Dream and Baratza Sette 270W
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

It's all software which has largely been rewritten for the Dream. There are two components to the offset. In the V1/V2 there is an internal offset that is hardcoded which is supposed to compensate as closely as possible across all machines for the cool down effect between the boiler and the group head. Then the user programmable offset can take care of any machine to machine or even lot to lot variances. On many of the V2s the internal offset actually overcompensated on most machines requiring a negative offset.

Jason as CCS is under the impression that the Dream is supposed to be the same way. However, if I had my guess based on what I am seeing LaSpaziale either forgot to add the code for the internal offset, screwed up the code, or forgot about it.

On the other hand, Jason still claims using his Scace he is getting 0 as the proper offset for the machines he is bench testing. Both of us think that our Scaces are properly calibrated so we are still trying to figure out what it really going on.

This morning I switched my machine from C to F and set the offset to +12F. From what Jason had told me about SW version 1.17 I thought in both C and F modes the offset range was going to be +-10 degrees. I haven't checked to see if the setting will go negative yet but the Max positive ranges are actual +6C and +12F. +12F is actually a higher absolute temp than +6C so I should have a better change of getting a good 201F out of the Dream even though I could not get a good solid equivalent 94C when using the +6C offset.

It is also possible (and probable) that Jason sets the offset in F mode only. So perhaps I will find that there is a code problem in C mode that does not exist in F mode. That would be an interesting discovery! I'll let you know later what I find out.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

Great info! Thanks, Chas. As for my Dream, the replacement temperature sensors are due today from CCS, so with any luck I will be back in business. I will report back on any other issues discovered.
Bob
La Spaziale Dream and Baratza Sette 270W
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

I gathered the temperature profile data for the Dream this morning in native Fahrenheit mode. I found that the offset goes to +12F so I used that. I'll be able to tell for certain tomorrow when I plot the data, but from watching the meter while pulling the test shots, I am thinking that +11F offset will do the trick.

When I initially suspected that the large offset might be due to a bad temperature sensor I got a new one and found that wasn't the issue. For test purposes I put one of the sensors in the normal spot on top of the boiler where they are in all the Dreams and I put the other one in the port on that back boiler cover where it has been for all the S1 Vx models.

In the next few days, I'll retest in Fahrenheit mode using the rear temp sensor to see what, if any, difference there is in the temp profile.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

The Scace I am using initially passed 2oz of water in about 45 seconds. Since this is not really adjustable, I went with that for the first set of graphs below. Later I discovered a tiny particle of lime which was restricting the flow out of the Scace. After a cleaning, I am now getting 2oz in 25-30 seconds as expected for the Scace design. This is really how it is supposed to work when used for testing. So I redid the graphs again.

Even though at this point the initial 45 seconds graphs should be ignored, it is very interesting to see the significant difference in the set of curves. The longer curves take a little longer to ramp up to the expected temperature then are pretty flat. However, the 25 sec shots exhibit a strong overshoot, then immediately drop back down below the desired temp and ramp back up.

The 2nd set of included graphs - the 25s ones - have 10 plots. The first five are using a +11F offset and the latter five use +12F. The next to last graph is actually a perfect example of what I would expect each shot to look like after the first two warm up shots. That's a God Shot profile right there! I am not sure why it happened once and only once. The other graphs looks like the PID settings need some tweaking. A properly tweaked PID should have fast recovery but without the overshoot. Actually in revisiting the 45s plots there is a hint of overshoot there as well, but nowhere near as pronounced as in the 25s shots.

I also still need to work with CCS and LS via CCS on why I am seeing such a large offset. I would think that the Dream should have a built-in offset that attempts to get to a 0 offset and then perhaps require no more than a degree or two of offset to account for machine to machine or lot to lot variances.
DreamNativeFModePlots.pdf
~45s Shot - Temp vs Time Plots
(253.8 KiB) Downloaded 537 times
DreamNativeFTopPort25sTests.pdf
~25s Shot - Temp vs Time Plots
(1.99 MiB) Downloaded 456 times
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

I'm running with the factory 0 deg offset in F mode with a setpoint of 203 F. I had it at 201 and shots were sour. At 203 they are perfect. Hard to believe mine is off by 11 or 12 degrees. My guess is that if I use an offset that big on my machine, I will get flashing to steam. I might try just to see.
Bob
La Spaziale Dream and Baratza Sette 270W
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

Look out, I am now playing with the PID settings and flying without a net!

Changing the PID settings is a real PITA. In looking at typical PID settings, some of them are either fractional or low single digits whereas P=300 and D=10,000 on the Dream. Since the processor in the DREAM can't work in fractions/decimal numbers I wonder if they scaled the values up so that P=300 on the Dream is really equivalent to 3.00 or even 0.3 on a hardware PID. So if you want to change the values you have to change them a lot and the "hold down the arrow to repeat" function is NOT implemented in Factory settings mode.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
Endo

Re: Dream vs GS/3 Throwdown

Post by Endo »

This new "Dream" seems like about as much fun as programming a first generation VCR. :roll:

After using a lever (only electronics is a heater and switch), I think moving over to the Dream would be more like a nightmare.
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

I read up a bit on PIDs yesterday. I am especially trying to find layman's explanations on what changes I can expect of the temp vs time plot if I adjust any one of the three parameters up or down. The article I read yesterday indicated that the D value in general controls the overshoot. So I did a series of plots with D changed from 10000 to 10200.

I saw no difference in undershoot at all. The shape of the graphs was pretty much the same as with D=10,000 except as if I had dropped the offset by about 5F! So rather than overshooting to 203 and then coming back down to 201, the overshoot went to 198 and then settled to about 196!

I have now changed the D setting to 9800 and will see what that does. In theory at this setting I will need to drop the offset. We'll see!
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

Typical PID units have a function that auto tunes the values based on a self check. Its a shame that La Spaz did not emulate that function in software.
Bob
La Spaziale Dream and Baratza Sette 270W
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

I came up with a potentially better value for "P" this weekend and had just started to work on homing in on a value for "I" when I got sidetracked by another project.

I realized before I started that this maybe an exercise in futility. The PID algorithm must be tuned by looking at time versus temp profiles where "temp" is the temperature of the internal sensor, not the Scace measured temperature. After all, it is the temp from the internal sensor that is driving the PID algorithm not the Scace temp. The difference between the internal boiler temp and the group measured temp will always be different and not necessarily in a linear way. Trying to accurately tune the PID using the Scace is not really a valid way to do it.

On the other hand, if I can come up with values for P,I, and D that give me a superior temp vs time profile at the group, should I care whether the PID is truly tuned to perfection?

Hardware PIDs will actually auto-tune themselves. However, the Dream has what I would call a 1st generation software PID algorithm. It may or may not be coded very well, I can't be sure. If I start with the water temp at 201F and begin to pull a shot, the wavy line indicating that the group boiler is ON doesn't come on until it hits 199F. The PID needs to be tuned so that the boiler comes on long before there is a 2F drop. The boiler should probably be coming on much more often but for shorter periods of time. If the PID cannot be set to achieve this type of performance, then I think the algorithm needs to be tweaked.

If I do come up with PID settings that give me a better result at the group, I'll let you know. Hopefully, after much more trial and error testing, I'll know a lot more by the end of the week.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

I totally agree with your observations and conclusions on this pid algorithm that's in the Dream. What is frustrating is that there are easily obtained well implemented software based pid routines that la spaz could easily have obtained rather than apparently doing this themselves at a rudimentary level.
Bob
La Spaziale Dream and Baratza Sette 270W
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

I heard from Jason at CCS yesterday. He did some major calibration on his Scace/Digital Thermometer and now believes that his Dream works best with a +10F offset. Using that and a base temp of 203F, he sees the temp jump up to about 204.5F, then settles back to 203F out to about 40s when the temp then actually starts to rise.

I feel vindicated now that he is seeing the need for an offset in the range that I am seeing and also is seeing the initial overshoot that I am seeing. The testing, comparisons, and tweaking between Jason and I will be ramping up over the next week.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

Awesome. It sounds like at the end of this we will get some rock solid recommendations for offset and pid values. Can't wait!
Bob
La Spaziale Dream and Baratza Sette 270W
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

Of course the ideal would be to hard code this into a new firmware rev.
Bob
La Spaziale Dream and Baratza Sette 270W
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

If further testing corroborates what we have seen so far, I would recommend that LaSpaz hard code a +10F/+5C offset so that user offsets are close to 0 on average.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

Chas: Is there any update on settings, etc. based on discussions and testing from You and Jason? I did note that the new version of the Dream manual on the CCS website recommends an offset of 10 deg F. Should I just put this in or wait for new PID settings as well?
Bob
La Spaziale Dream and Baratza Sette 270W
User avatar
chas
Vivaldi Dreamer
Posts: 3044
Joined: Wed Jun 02, 2004 11:52 pm
Location: Central Maryland
Contact:

Re: Dream vs GS/3 Throwdown

Post by chas »

We haven't done anymore testing recently because we were waiting for LaSpaziale to come out with the next software update on which they were asked to address a number of things including PID and offsets.

HOWEVER, I got a Beta of 1.8 this morning from Jason at CCS. From his e-mail it sounds like LaSpaz only addresses the DST issue and a lights issue in this release. I wasn't even aware that there was a lights issue but supposedly the lights would turn off when you switched profiles even if you had both profiles set for lights always on in both. It sounds like LS definitely went just for the low hanging fruit with this update.

Since we now have no idea how long before we'll see the update re: the offset and PID settings, I suppose it wouldn't hurt to go with a +10F offset. Definitely taste test it with a straight espresso before and after the offset change to see what you think.
Chas
LM GS/3 & LaSpaziale Dream v 1.25 (US 120V)
Mazzer Kony E, Customized Rocky
Hottop P/B
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

Based on a taste test with the setpoint at 200 F, I would say that an offset of 7 or 8 F works best for me. At 10 F, the first shot out, no flush, is good, and subsequent shots start to get dull and bitter. At 7, the first shot is too bright or sour and subsequent shots are good.
Bob
La Spaziale Dream and Baratza Sette 270W
GZim

Re: Dream vs GS/3 Throwdown

Post by GZim »

I have enjoyed reading this thread very much, but I notice no activity for a long time. As I am in the process of going from a Mini to a Dream, I am wondering if any of these problems still exist, or if it has all been addressed with the most recent model and software.

Thanks.
goodboyr
Barista
Posts: 255
Joined: Thu Jan 10, 2013 10:05 pm
Location: Canada

Re: Dream vs GS/3 Throwdown

Post by goodboyr »

All good, except for frequent temp sensor failures for some of us.
Bob
La Spaziale Dream and Baratza Sette 270W
GZim

Re: Dream vs GS/3 Throwdown

Post by GZim »

thanks....
is there any other machines I should consider in the same price range?
The automatic timer feature is really key for me, to wake up and have the machine hot.
Post Reply

Return to “Dream and Dream T Common Feature Discussions”