Cafelat Robot Versus EspressoForge

A haven dedicated to manual espresso machine aficionados.

#1: Post by LewBK »

I now have both a Robot and an EspressoForge, and I wonder if others have had similar experiences to mine and can provide some advice. I find that the Robot's workflow to be far easier than the EspressoForge, but I find so far the Forge makes better espresso. My Robot's espresso has often had a slight bitterness to it, despite my efforts to tamp well and use the shower screen correctly. Meanwhile my Forge allows me to use an OCD distributor and leveler easily and I think that has improved my results, with numerous sweet shots. Yet the work involved levelling the espresso is a bit of a pain as some espresso invariably sticks to the OCD distributor and I am busy tapping the VST basket I bought and then using a chopstick or fork for WDT before using the OCD. All of that work pays off though with some excellent shots. But I don't do that with the Robot because well an OCD distributor won't reach to the bottom of that basket. I tap and I use chopsticks or fork to WDT with the Robot, then I tamp and make sure the shower screen is inserted correctly. I've adjusted the grind for the Robot to hit 17 kg on a scale for the right pressure, but the puck still cracks sometimes which I imagine means chanelling--it's hard for me to see if it's chanelling while I push the arms down. I wonder if the paper filter might help I keep reading about here? Also, I wonder if people have had similar experiences with the EspressoForge, the work involved and the delicious shots? Also, I wonder if Cafelat or Paul Pratt has announced a date for a self-levelling tamper release or is that not happening anymore? Thanks in advance for responses.


#2: Post by hbk520 »

I have same channeling issue here , I have tried filter paper , WDT , vibration in palm of hand , to eliminated channeling .
It works time to time ( record extraction every time) , that's why I assumed tamping is key process in my case.
As you mentioned Robot 's basket can't use OCD to make bed distribute well . I'm still looking for fitted tool (OCD) . If anyone found it , pls share to us.
BTW , I also experienced medium or dark roast is easier to make great extraction without channeling than light roast on Robot.
Maybe I should try another grinder for light roast bean ?

User avatar

#3: Post by Denis »

WDT is mandatory specially to higher diameter baskets (the smaller diameter basket you have the more forgiving because of the bed height). Forge is not 58 mm, right?

So if you are into light roasts then you must do WDT with a needle device, OCD will do almost nothing, just level the coffee for a straight tamping.

To avoid channeling make sure you are using the right amount of coffee that works (18-20g?) make sure you are tamping, and the most important that has been discussed on robot topic, when you put the screen shower on top of the coffee, push it hard with your fingers. It should sit steady on top, so when you pour the water it doesn't come loose.

When using ultra light coffee, you grind finer, this means you have to do a longer brew time, and the most important is to make softer ramps between low pressure to high pressure, so you soak the bed 100% before going ham.

The channeling can come from the grinder, using a weak grinder that doesn't deliver a uniform grind start to finish can result in channeling.

User avatar

#4: Post by MB »

@Denis: The Espresso Forge uses a standard shower screen and 58mm baskets.

@LewBK: I have both the Robot and the EF. I have similar experiences with the EF providing very tasty shots, but it's more sensitive to grind and prep it seems than the Robot. I used to see cracks occasionally in the Robot puck, but remember that to examine the puck you have to apply negative (or upward) pressure when lifting the arms and/or removing the holder. I imagine this has an impact on cracking, but not so much on shot quality since even on the cracked puck shots, the beading and pour on the bottom of the filter basket has been even, and I get tasty shots.

I feel like I have had some outstanding shots on both devices, with maybe a few percentage more on the EF. However, the Robot just seems to not have any misses for me, once I got in the groove (and that didn't take long at all). Overall it is more forgiving. Also, the workflow is better on the Robot.

If you will wipe off the face of the OCD more often, it should have less fines sticking. Seems an invisible coating of coffee oils attracts the fines.

I haven't tested WDT with a chopstick, but with other items it seems like the thicker the implement or more wires of the stirring device, the more likely I would compact some grinds (and unevenly). I have been wanting to reduce the wire thickness of my WDT implement, since when I was using just a paperclip, the results seemed slightly better. So, on the Robot, I grind into the basket, WDT with a four wire pronged implement, shake side to side lightly to level, give it a light tap on the counter, then nutate the tamper around a few times at progressively less and less of an angle till essentially flat, and finally a light tamp.
LMWDP #472

LewBK (original poster)

#5: Post by LewBK (original poster) »

Thanks for the advice everyone. I use 18 grams, occasionally 17 grams, of medium or dark roast coffee, never light roast, with the Robot. I may try to grind a little coarser as some of my best shots on the Robot have been when the bars of pressure were actually a little bit less than I've been getting lately with the 17 Kg on the scale. I would like some sort of distribution or leveling tool for the Robot. I know the OCD only distributes the surface of the Forge, but for whatever reason it seems to be working when combined with WDT from fork or chopstick.

Team HB

#6: Post by samuellaw178 »

I can't help but thinking you may be looking at the wrong place for a solution (I could certainly be wrong). In my own experience, the OCD-style polisher doesn't do all that much. If the grind needs to be stirred/mixed, WDT does everything the OCD-style groomer does and better.

Are you seeing channeling or lop-sided extraction on the Robot? OCD can help with lop-sided extraction (due to uneven tamp), but not so much on channeling. I wouldn't worry about the cracks at all for reasons others have mentioned.

I would also focus more on the extraction variables, namely the brew time (how many sec preinfusion & extraction time) and the brew ratio (17g in how many gram out)? Do you preheat and if yes, what's your protocol? If no, you can consider doing that and see how that impacts the taste.

Could you be confusing bitter with sour (it can be common to mistake one for the other)? Can you describe the bitterness that you're tasting? If any, I did find that it can be easy to 'overextract' on the Robot by pulling too much water through (because there's a lot of water available). When that happens, the flavour doesn't suffer but the aftertaste/finish of the shot can be unpleasant (but don't think I would describe that as bitter).

Sorry that it's a response with more questions than answers. :P But if you have all of the above addressed, I think it'll be quite difficult for you to get bad/bitter shot..

LewBK (original poster)

#7: Post by LewBK (original poster) »

OK, I just made a delicious shot on the Robot by significantly increasing the grind size from a 1.7 on a Kinu m47 grinder to 2.0. Same 18 grams, same coffee, same pre-infusion of 5 to 6 seconds, WDT with a fork, same firm insertion of shower screen. What happened though is instead of hitting 17 kg on the scale for pressure, I only hit 8 to 9 kg and the shot came much more easily. I assume--please correct me if wrong--8 to 9 kg is only about 4 or 5 bars of pressure. The result was a much smoother sweeter shot than before. Interestingly, I can hit much higher bars of pressure with the EspressoForge and still get a smooth sweet shot.


#8: Post by jpender »

LewBK wrote:I assume--please correct me if wrong--8 to 9 kg is only about 4 or 5 bars of pressure.
In the Robot manual 17 kg is equated to 6-7 bar. So 8-9 kg would be only 3-4 bar.

LewBK (original poster)

#9: Post by LewBK (original poster) »

Interesting, maybe I'll try grinding one notch finer to get the pressure up slightly.


#10: Post by jevenator »

Denis wrote:
When using ultra light coffee, you grind finer, this means you have to do a longer brew time, and the most important is to make softer ramps between low pressure to high pressure, so you soak the bed 100% before going ham.

The channeling can come from the grinder, using a weak grinder that doesn't deliver a uniform grind start to finish can result in channeling.
I wish I knew about this earlier. I owned the Robot for a few months now and for some reason based upon some videos I saw and my own understand I should like mimic a regular machine with like a PI and then full ramp to pressure quickly.

I had an amazing shot (dry process, light roast, (fruit bomb)) and all I changed was the technique. I do 17.5g in, WDT, light tamp, push shower screen down, all that jazz. But when I was pulling down on the levers I took it very gently and smooth. I slowly pressed to imitate PI and then just kept applying pressure slowly (over a period of 6-8 seconds) to 6-7 bar. I keep it there until I reach approx 2/3 of my desired shot weight and taper off to 2 bar before lifting up on the levers quickly to cut off the shot.

BEFORE I would do a long 20 seconds PI to make sure it was soaked and quickly put the power down to get 6-7 bar.
LMWDP #643