• Grinders

  • DF-64 variable RPM with inverter - experiment

I am planning to make an experiment of using an VFD inverter to vary the RPM of DF-64 grinder. I already bought a single phase output inverter and performed a quick smoke test which went fine. With the motor connected directly to the inverter (bypassing the grinder switch electronics) the rpm goes up and down according to the frequency set in the inverter.

I have not seen any success stories about this, probably there is a reason for that, but I have not seen explicit failures either, so here we go.

Planning to do some testing with coffee in the following days, if you have any suggestion of testing, or if you want to share your opinion about this subject that would help!

Cheers

    Petre I have not seen any success stories about this, probably there is a reason for that, but I have not seen explicit failures either, so here we go.

    It will depend on the controller you use, some will reduce torque a lot, so a little less. You might find, especially at lowish speeds, the grinder may stall with harder beans. e.g. around 800 rpm.

    Contact me

      Most likely - I am also curios to find where that limit is.

      I’ll be using this inverter (which what the hack, it is £82 now? I payed £96 2 days ago..). I read in the manual that it has some “Torque compensation” settings which might help.

      In the meantime I measured some rpm at various frequencies:

      1480rpm mains 50Hz

      1000rpm 35Hz

      837rpm 30Hz

      800rpm 29Hz

      400rpm 17.3Hz

      This topic will likely transform into “Things not to do when moding a grinder”…

      Let me start by listing the setup.

      • Inverter connected directly to the grinder motor. Do not connect the grinder plug directly into inverter output, you will blow up grinder switch electronics, ask me how I know.
      • DF 64 with SSP red speed HU
      • Mythos clump crusher mod
      • Normcore autolevel tamper
      • Silvia Pro espresso machine
      • VST basket (20g used in these tests) and naked portafilter
      • Normcore puck screen
      • Red brick coffee beans roasted on 31 Jan

      Normally, for these beans I dial in 18g in, 34g out, ideally 24-28 sec.

      For today session I tried test the differences between 1480rpm mains 50Hz and 1000rpm 35Hz with inverter. TLDR, I tasted virtually no difference in shots.

      Today session started with a bit of a fail. Previously, I tried some random lower speeds and noticed that grounds did not flow out quite efficiently, clogging up around the clump crusher and I decided to remove it. The result of that was that I got not only clumps, but also the entire chute clogged after 3 grindings of 18g - bad idea.

      I cleaned everything up and put back the clump crusher, dialed in at 50Hz full rpm got 24 sec shot. I repeated the process with exactly the same settings but at 35Hz 1000rmp this time and got a 28 sec shot. Usually these kind of differences in shot times happens often, it is basically hit and miss if I try to make it exactly the same. I gave it another go at full rmp hoping I get the same shot time, and even worst, I got 21 sec. At this point grinder was noticeable warm for which I blame the drift in settings. (Side note, usually with the grinder warm, if I chase the grind settings to bring the shot time back up to 27-28, next time when grinder is cold those settings will be waay to fine, giving 35-50 sec shot, so it needs another dial in and wasted beans).

      So as I said, virtually no difference in taste, maybe the 1000rpm was slightly better extracted but that is probably because of 28 vs 24 sec difference. Definitely no new notes, flavours or anything else. Just like any other good extraction I got from this setup at full rpm.

        Petre So as I said, virtually no difference in taste, maybe the 1000rpm was slightly better extracted but that is probably because of 28 vs 24 sec difference. Definitely no new notes, flavours or anything else. Just like any other good extraction I got from this setup at full rpm.

        I didn’t really want to say anything, but for almost a year I have been conducting tests with various grind speeds….I have not really noticed any change in flavour, just a difference in grind quality. There is quite a bit of stuff being spoken on the web about speed profiling and how it changes the flavour of the shot. This was happening as soon as speed control became more common on the very high end grinders (Kafatek, Levercraft, Lagom etc.).

        There is definitely a minimum speed a flat burr grinder should run at, which depends on:

        • Burr geometry
        • Burr carrier/chamber design

        There is something I call the optimum burr packing speed, a point at which the speed it fast enough to “pack” the burrs correctly. Going faster only has the advantage of a quicker grind time, but at the expense of slightly more fines

        For conical burrs the situation is a bit different as the burrs self pack quick well, so very low speeds work fine, then the speed becomes the lowest required to get the coffee out of the chamber. e.g. 71mm burrs on the E92S rotating at 240 RPM is slow and the grinds do back up in the grinder, which is why it has massive retention (45g) and has to be converted to single dosing for home use (where it works extremely well). Hand grinders can go at 60 rpm or slower, because the coffee drops straight out the bottom of the burrs. It’s why the Niche Zero runs at the speed it does.

        The stuff about speed profiling for flavour, I have not found to be true….all other things being equal, changing the speed on a flat burr grinder if around the burr packing speed will simply change the modality of the grinds. I suppose when speed control came about on these very expensive grinder, it was a combination of marketing and perhaps a fear of not doing the work to get the optimum speed. Suffice to say, when individuals drop big bucks on a grinder, they will of course notice a huge difference with speed control (expectation bias and placebo effect)

        Contact me

            The stuff I’ve seen shows that faster speeds grind finer with more fines but same could be achieved by adjust the grind setting finer instead.

            Gwilym Davies did some work on a Mythos 2 which showed, adjusting for shot times, that the lower RPM had consistently higher extraction yields (500rpm v 1000rpm). This could translate into a taste difference. Tbh personally I haven’t experimented with different rpms on my Ultra. I use 400 for virtually all brew methods and only ramp it up to 1,500 when bag grinding. Max on the Ultra is 2,500.

            For me there are so many other variables that could have a bigger impact so hence why I kept it at the single rpm.

            DavecUK

            Thanks Dave and Steve for your input, it is very reliefing to hear more people is looking into this matter, even if there are no very clear advantages found yet.

            I would be curious how bigger flat burrs would behave in this kind of testing. I hear about the Mythos 2 (Steve, could you post a link to that?) and tbh the grinders I inspired for this testing were also bigger flats. I suspect that, as David mentioned, there is a different geometry and I believe it also a matter of size:

            • 64mm burr mean approx 200 mm circumference
            • 85mm bur mean approx 267 mm

            This would be the distance traveled by the burr in 1 rotation at the very edge. From here I deduce there will be at least 1.3 rmp ratio between the 2 to achieve the same speed at the burr teeths level. I am sorry if I butcherd some burrs theory, if that exists I would love to learn it.

              Today I tried 800 rpm (29Hz on inverter) vs full rpm connected to mains.

              The overall result was pretty much the same as previous with no additional or better flavours magically popping out at lower rpm, but there were some interesting stuff going on I’d like to mention.

              First, I dialed in the grinder because what was yesterday 24 sec today was 50+ seconds. By the time I manged to get the shot in 24 ish seconds the grinder dial was back in the initial position.. weird… Anyway, I pulled a 24 sec shot with grinder connected to mains and the I’ve done exactly the same thing but with the grinder connected to the inverter set at 29Hz(for which previously I measured 800rpm). The results:

              • grinds were pretty clumpy this time
              • shot was taking 28 seconds (again)

              I use wdt at every shot so clumping should have not affected the shoot too much.

              I can’t help but notice that shots take slighty longer at lower rpm - it happened at both tests. Usually, from shot to shot it is the other way around. Is this a sign of “better” grind quality like you David mentioned? Is it because there are more fines? I have no clue.

              I don’t know if I would continue further down with rpm. The clumping for me suggests that grinder is struggeling to get those grinds out because they don’t carry enough speed to shoot out. Probably a different clump crusher would be necessary and this is a land I don’t wanna go now :)

              What is clear for me is that I could not find anything special or worth the effort to keep this mod, at least for this setup. It might work differently of other grinders, but for this one you probably want do something else with those £100 :)