Vibe pump profiling - understanding PWM controller

Need help with equipment usage or want to share your latest discovery?
mikeness
Posts: 93
Joined: 9 years ago

#1: Post by mikeness »

Bit of a long story how I stumbled on the info in the Bezzera Strega threads (fueled by upgraditis), but the idea of varying voltage of my vibe pump to reduce brew pressure sent me down a rabbit hole of vibe pump profiling threads. I have a Magister E61 with Ulka pump...

I know there are several threads both here and at CoffeeGeek, but I've found it hard to find clear consolidated information on how to control my vibe pump pressure. My desire would be to achieve a 3 bar preinfusion for example, but I'm not convinced that a) it can be done (mostly for the simple fact that if it were doable, I'd expect it to be more widely adopted/detailed mod), or b) if it is indeed possible, that I have the skills/aptitude to pull it off. So I thought I'd layout two basic conclusions that seem to have been reach based on having read through those threads... feel free to correct me, as its part of exercise of providing an accurate summary that I hope this thread can become:

1. Basic voltage control though dimmer or variac --> Works well and its repeatability is dependent on the quality of the voltage controller, however only really provides variation reliabily above 6-7 bar and will stall the pump below that. Therefore this method is better suited to be used as an OPV type application vs actual full range profiling.

2. PWM controller --> rather than simply reducing the voltage to the pump, which provides variability only in a limited range, users have modified the frequency of the signal through a PWM controller, allowing pressure control from 0-max pressure. Now here is where the threads sort of lost me... how simple is this to do?

Do basic PWM controllers exist? Whereby it would act similar to a light dimmer and it would be up to me to guess and check where the knob on the controller produces X bar pressure? Or must every controller be customized for a particular frequency(ies) combined with a complex algorithm(s)?

Users seem to have integrated PWM controllers into complex systems with an Arduino to create a profiling interfaces that facilitate executing a given pressure profile repeatedly, but I am not interested in that at this point. Ideally, I'd love to find a cheap and dirty way to produce a 3-bar preinfusion pressure, no matter how manual the process is, then evaluate the extent to which I would invest in this mod and continue down the rabbit hole...

Thanks for any contributions :) I'm just trying to wrap my head around a topic that could really change the way I brew if I can just get through the information out there that is often presented based on varying goals (ie. lack of OPV, looking to set pressure to 9bar) or technical knowledge.

691175002
Posts: 89
Joined: 8 years ago

#2: Post by 691175002 »

Since the pump is essentially a "piston" actuated by a solenoid, the flow is tied to the frequency of the pulses being sent to the pump. Methods that reduce the voltage or current provided to the motor will have very coarse control because each actuation moves a set volume of fluid. Any reduction in flow would be a result of stalling the motor or being unable to complete a full pump.

The ideal way to change the flow rate is to slow down the frequency of the pulses provided to the motor, allowing it to complete each actuation at full torque.

This could be done with a commercial single-phase VFD (Variable Frequency Drive). Chinese models start at ~100$ shipped, but they are extreme overkill for this application. You could also hack something together with say a bridge rectifier and a MOSFET.

The definition of PWM is control of duty cycle, not frequency. It should not behave any differently than a dimmer or any other way of reducing the amplitude of the underlying wall power..

pcrussell50
Posts: 4010
Joined: 15 years ago

#3: Post by pcrussell50 »

Breville does it quite nicely with a it's adjustable preinfusion, where you can set anywhere from 55% pump power up to 95%. I've played around with low pressure extraction, by using "preinfusion" as the actual extraction pressure, and setting the pump power to between 75 and 85%. It worked amazingly well. Perfectly, really.

In any event, Breville has some kind of patent on how they do it, IMS. Bill Crossland knows quite a bit about the Breville patent, and about these things in general. But I had the idea that if methods of controlling vibe pump output are getting patented, they must be something more than your typical dimmer switch.

-Peter
LMWDP #553

User avatar
radudanutco
Posts: 147
Joined: 11 years ago

#4: Post by radudanutco »

indeed a vibe pump would require a variable frequency pulse control in input, in order to push upstream a variable quantity of water;
and I suppose, as a vibe pump has such a steep flow-pressure characteristic at nominal voltage/frequency, it should have something highly modified profiles at different frequencies; at 50Hz/230V, my Ulka EX5 from Strega, is pushing over 10 ml/s at 0 (gauge) bar pressure (beginning of preinfusion for about 5s at Strega, or in a flush), while at 10 bar, at the end of a normal pressure ramp, it is pumping about 3 ml/sec!
but what would be the pressure variation at, say 30 Hz and different flow rates? I don't know;

I do not have experience of using PWM's with vibe pumps, but I would suppose the flow rate/pressure charateristics of a vibe should take precedence over the PWM/dimmer choice;
as of PWM controller, even it has a fixed frequency pulse output, it can be modulated pulse width wise, as to have a variable frequency resultant output; but still I wonder if the flow rate - pressure characteristic at a certain frequency does not have the precedence...

with a dimmer, the pump voltage supply has a fixed frquency, but a variable RMS value; the pump - I imagine - is functioning with some incomplete cycles, resulting in a somewhat inconsistent control; but, if the setup has an added pressure gauge, I cannot see what's wrong;

in fact, I am using such a mod on my Strega for over a year now;
with a (fan/light) dimmer and a pressure gauge on output of the pump, I can control a pressure(flow) profile in preinfusion as intended; that is, there is a position of the dimmer knob, for which I have a very slow preinfusion: Strega group chamber filling in about 17 sec (cca. 65 ml), then a pressure ramp up to 3 bar; but usually I do not go so low on pressure, my lowest setting is for a 12s, 0 bar PI, then a pump pressure ramp up to 4 bar (well, then up on springs etc.);
what can I add: I had 5 different dimmers to start with, and I had chosen the one which seemed to make the most responsive pump at lower pressures, by using a DIY portafilter with a pressure gauge;

User avatar
redbone
Posts: 3564
Joined: 12 years ago

#5: Post by redbone »

Would a potentiometers (gain controller) be sufficient ? Similar to those used for amplifier controlling.
Between order and chaos there is espresso.
Semper discens.


Rob
LMWDP #549

mikeness (original poster)
Posts: 93
Joined: 9 years ago

#6: Post by mikeness (original poster) »

Thanks all for your replies - this is just the type of discussion I was hoping to have... hoping to keep this going until we get to a solution, then I'd be happy to post a DIY for all to benefit as a reference.

It seems that I have misunderstood the PWM's application, in that it doesn't actually modulate the frequency, but rather modulates the on/off duty cycle of the vibe pump. This would therefore modify the pulsing behaviour of the pump piston in a more erratic fashion vs. actually modifying the frequency of the pulses through say a Variable Frequency Drive (looked into these, seem really awesome but as 691175002 stated, seems overkill... I can't find much under $200 CAD), but could lead to a similar repeatable result of maintaining a reduce pressure at said spot on the dial.
radudanutco wrote:with a (fan/light) dimmer and a pressure gauge on output of the pump, I can control a pressure(flow) profile in preinfusion as intended; that is, there is a position of the dimmer knob, for which I have a very slow preinfusion: Strega group chamber filling in about 17 sec (cca. 65 ml), then a pressure ramp up to 3 bar; but usually I do not go so low on pressure, my lowest setting is for a 12s, 0 bar PI, then a pump pressure ramp up to 4 bar (well, then up on springs etc.);
what can I add: I had 5 different dimmers to start with, and I had chosen the one which seemed to make the most responsive pump at lower pressures, by using a DIY portafilter with a pressure gauge;
Interesting that you've found a way to get consistently such low pressure using a light dimmer, as most have seen their pump stall. Are you able to share which dimmer seemed to give you the best results? Also, just to confirm, does the pump maintain the 3 or 4 bar pressure? Or are you simply releasing your lever when your reach the desired temperature cutting off the power to the pump altogether? The ramp up time on the E61 would obviously be much quicker vs the Strega so I'm not necessarily concerned about that.

mikeness (original poster)
Posts: 93
Joined: 9 years ago

#7: Post by mikeness (original poster) »

Thought I'd provide an update... I went ahead and installed a dimmer on the hot wire to the pump. The cost of doing so is less than $10 really so figured I had nothing to lose.

Result:
I can modify the pumps output to result in virtually any pressure.

I've worded that sentence carefully as I don't believe the dimmer relibly limits the pressure output of the pump, but limits the flow of water, thus resulting in reduced pressure (may be a distinction without a differnce). If I'm shooting for 3 bar with the dimmer dial at the dimmed position, for example, it takes easily twice as long as when the pump is at full power to simply build any pressure at all. Then the pressure slowly builds to the desired 3 bar, but could continue to rise if left long enough.

I'll post some videos with a pf gauge to better illustrate the behaviour. In the meantime, here are a couple of photos of the setup. Notice how close 20, 40 and 80 psi (measured with a pf gauge) reference marks are from one another. Without a pressure gauge, it's almost like guess work when pulling the shot. That being said, every "profiled shot" pulled has been better than before the mod, so my initial inclination is to recommend the mod.




Does anyone know of a good DIY guide to instal a brew pressure gauge? Might not be exact, but I could get a sense for the accuracy by comparing with a PF gauge and would give me a bit of a reference point when pulling a shot.

jonr
Posts: 610
Joined: 11 years ago

#8: Post by jonr »

Here is a summary of vibe pump control:

Variac - works well
High speed PWM (Khz)- works well
synthesized, lower frequency, 1/2 AC cycle sine wave - don't know, why bother
Dimmer - most people report OK
Low speed, unsyncronized PMW (10s of Hz) - works poorly