Voltage Controlled Pump Pressure (Ulka)

Need help with equipment usage or want to share your latest discovery?

#1: Post by mini »

revised 5/5/10 - crimping methods
revised 5/16/10 - reasoning behind rheostat
revised 5/17/10 - "good links" section below; soldering; where to get wire

I've been thinking about this topic a lot. It has been done in a couple of forms by different people, but I think this way is the best. The following post is going to be rather lengthy so that I can get all my ideas out into the open. (read: It's a beast) This post is also a very detailed how-to for others to follow. This information is aimed at people without much/any electrical experience and without a general understanding of vibratory pumps. Bear with me if you don't fall into those categories. I hope this idea catches on, because I really think it's a good one.

Background / Purpose:

Something like 90% of the entry level home espresso machines out there use vibratory pumps - and most of these are Ulka brand. The Ulka EP5/EX5 is a fine work of engineering; efficient, durable, simple and very well suited to making espresso. However, the very nature of its design creates a flow-rate vs. pressure curve...

meaning that the more you limit flow, the higher the pressure gets. By varying the coarseness of your espresso grind, you can certainly alter the amount of pressure the pump generates by changing said flow rate. Still there are obvious limitations. Towards the ristretto end of the shot spectrum, many Ulka pumps will generate much more pressure than the ideal ~9 bar. There is no way around this. High pressure causes an increased chance of channeling, and can also cause undesirable flavors.

For this very reason, a lot of manufactures install adjustable over-pressure valves. Adjustable OPVs allow the user regulate pressure to their liking. While OPVs as a safety feature are extremely common, adjustable ones are much less so. My machine's OPV only kicks in around 13 or 14 bar from what I've read.

I wanted to make a modification to lower the pressure on my machine. I was really tired of 9 out of 10 shots having channeling. Yes, I've tried WDT, NSEW, very rapid doser clacking, etc, but I was still having problems. I've owned my machine long enough to know that it's a bit finicky. I really thought that there was something I could do to improve things. So here's it is.

Why Voltage Control?

Because it's easy to install, broadly applicable, easily reversible, and really, really cheap. Sure, I could probably find a way to install an adjustable OPV on my machine, but it would be none of these things.

I'd never done an electrical project on anything before this, but I had no problems and finished in 20 minutes. The following mod is applicable to every machine with an Ulka EP5 or EX5 pump, and the idea is applicable to almost any vibratory pump. Because of some standardized spade connectors inside, I can reverse this mod any time I want in 30 seconds. And finally, the total cost of parts I used in my prototyped version is $13.80 (plus $5.00 for shipping).


The basic idea is to install resistors in-line with the pump in order to reduce the voltage to the pump and thereby reduce the overall pressure curve. Adding resistance in series lowers the pump strength linearly, but variances in pumps makes this a guess-and-check method. Therefore I decided to use a variable resistor, or rheostat, to find out the optimal resistance for my machine. By means of a knob I can add resistance on the fly from none to some, too.


The following mod is designed to be done completely at your own risk and with your own best judgment. While this mod is easily reversible, I can make no promises that it won't void your machine's warranty or even DESTROY your pump. I have read one posting online specifically stating that similar mods will hurt pump longevity. I am also describing ways to change the path of large voltages and life-threatening currents. I was willing and capable to assume these risks. Make sure that you are as well.

That being said, I sincerely believe that this mod will not damage my pump and that the wiring I have laid out is safe and newbie friendly. I am a second year mechanical engineer who has consulted with several electrical engineering grad students. Take my advice at your discretion.

The how-to describes everything that I've done so far. My prototyped version isn't very pretty or permanent, but I'm going to explain it thoroughly anyways. That way, you can make any customizations with confidence.


I tried to use parts that would exceed the demands placed on them. I used wires, connectors, and a rheostat that are rated much higher than they need to be for safety. You should too.

rheostat - $10.00 (+$5 s/h)
A seller on Ebay had the best selection and prices. It is an Ohmite rheostat, rated 0-50 Ohms, 25 Watts, and .71 Amps. These numbers are important, do not get a lower wattage rheostat. Here is the exact ebay link.
18 to 14 gauge wire (6 ft.) ~$1.80
I got 12 gauge originally, rated at 90C. It was extremely stiff, and way overkill - get a lighter gauge. Home Depot sells wire by the foot and I got two different colors. I recommend two 3ft. sections for mistakes or external mounting. Stranded wire is preferable for bending. I have since bought some marvelous wire from user "ira." See his post below and PM him if interested.
spade connectors (fully insulated) - $2.00
I got these from RadioShack. They came in a bag of 3 - you only need one female and one male. Get connectors to match the gauge of wire you are using.
wine cork
Coke bottle
... to be explained below


The Mod: Step by Step (with tons of pictures!)

Here's the wire I used originally. Blue and black. Yours will probably be a bit thinner

Here's the rheostat. Doesn't it look fun?

Here's what the spade connectors look like. Yours probably won't be yellow; yellow is for 12AWG which I mistakenly used at first.

Super simple.

Bottom of the rheostat. You can see the sweeper mechanism for increasing resistance.

Notice that there are three terminals on the rheostat (on the bottom). You will attach wires to the middle, and only one of the ends. If you use the middle and left terminal, then turning the knob counter-clockwise will increase resistance and decrease pump power. This made the most sense to me - turn up the power with clockwise rotation. You can choose the opposite if you want.

Step 1: strip about a quarter inch off one of your wires. I picked blue. Make sure to cut the end off cleanly before stripping. Home Depot kind of mangled the ends for me.

Step 2: stick the stripped wire into a MALE spade connector. Crimp down firmly on the stalk. I managed to get by with some pliers. This took a lot of force. Crimpers would be the best tool to use. A vice grip will work as well. Don't worry, they are pretty hard to break.

Make sure again that the wire gets crimped well for good contact. It should withstand a good tugging. General note: it's still a good practice to make sure you pull connectors apart by the plastic, not the wire.

Step 3: strip the other wire and crimp on a FEMALE spade connector. This is the black wire.

Now we are going to attach the two wires to the rheostat. Strip the remaining two ends about the 3/8 of an inch each.

Step 4: Thread one wire up through the hole in the right terminal. (Remember our directional discussion.)

Step 5: divide the wire strands roughly into half, and bend them around the outsides of the terminal. I've been calling it a "squid connection." Needle nosed pliers worked best. MAKE SURE it is on there firmly. We don't want any live wires popping off.

This is not a good idea. Repeat, don't make connections like the picture above. It's potentially dangerous, and there is a lot of extra resistance caused by these connections. My pump was behaving strangely even with my knob turned "off" by week 2. So instead, we need to SOLDER the connections.

Don't be scared. This was my first time soldering too. You can hardly use too much solder. You want good contact, a smooth bead surface, and a shiny finish. Here is a good tutorial on soldering, which has even better links in it. All advice on through hole circuit board soldering can be scaled up to wires. Of course, if you really don't want to solder (or don't want to buy an iron), try to find a friend to help out.

Solder the wire in place.

Step 6: use a similar squid connection to attach the other wire to the MIDDLE terminal.

Similarly, solder the other wire to the MIDDLE terminal.

This is what you should have. A white wire replaced my old blue one in a later revision. See! Even a newbie like me can get make half-decent connections. You're done with all the wiring now. Quite exciting, isn't it?

Now, the only problem is that these rheostats are designed to be surface mounted, meaning that only the shaft is external and the rest is hidden inside a machine. Everything that is metallic on the bottom of the rheostat conducts current. We need to find a way to insulate this from any accidental contact. If you plan on mounting the rheostat inside your machine, you will have to find a better solution. For right now, a Coke bottle works.

Step 7: cut a Coke bottle in half and cut out a hole in the very bottom the size of the brass threading. I got by with a pocket knife, but a big drill bit would be advantageous. Cut a notch or hole for the wires as well.

The rheostat conveniently already has a nut and washer. Unscrew these. Bad connection visible!!

Step 8: Attach the rheostat inside our make-shift housing.

If you don't want to buy a knob, you can make one.

Step 9: make a hole in a wine cork with a screw driver.


Open up your machine now.

Ok! You are standing in your kitchen in a puddle of water. I beg of you, unplug your damn machine! Do it RIGHT NOW before you forget. I lose sleep at night thinking about people who touch live lines...

Locate the pump.

Notice that there are two terminals on the pump, each of which has a spade connector attached.

Step 10: it's time to plug your rheostat into your machine. Unplug one of the spade connectors from your pump. It doesn't matter which wire. I picked the top because it was easiest to reach. Wiggle it back and forth if it's stuck on there. This should be a female connector. Plug in the male connector from your rheostat. Now plug the other female spade connector wire from the rheostat back into your pump. Make sure that you bend the wires in a way that they do not contact your boiler. You want them to run in as cool of a path as possible so they don't melt - you have 3 feet, after all.


Here's how I have it running currently. I have taped some of the wires together to make it look neater.



Now, the first thing that everyone is going to ask is, "Does it work?" The answer is yes. It reduces the pump pressure in both of the ways I am currently able to test - audibly and with unrestricted flow rate. The rheostat seems to function in a surprisingly smooth manner. At one end the pump seems to function completely normally, at the other, with much less power. Dropping the voltage too much can cause vibratory pumps to fail and require re-priming, but 50 Ohms does not seem to stop my pump.

The better question is whether or not this mod is useful or worth it. I don't know yet. Obviously, I took the time not only to do this mod, but to describe it in a lot of detail. I think it's going to be great. I think other people are going to use it (frankly, I don't know why they haven't already). Even if all it accomplishes is taking the edge off of my Le'Lit's or some Silvia's ramp up pressure, it will be a success in my book. But before anyone starts dreaming too much of pressure profiling Slayer-style on an entry level machine, I'll need some time to ponder results and pull a fair number or shots.

For those of you who think that this mod is a bad idea and won't change things... well we're going to find out, now aren't we. 8) but be gentle, please

I plan to follow this post up with a slightly more technical discussion of pump function aimed at defending my stance of pump longevity in relation to resistors. I will post posted some sources (below) to consolidate a fair amount of information that is floating around the web on this topic.

And of course, I will keep you posted on how I like this mod, any revisions, and any improvements in my espresso. Hopefully I can make some comparison videos as well.

Thanks for your time.


#2: Post by JimG »

Very nice approach, Matt. I've done something similar using a couple of triac-based dimmers, but I like your rheostat better. The dimmers chop up the sine wave so much at high phase angles that the pump seems to become "confused."

A few comments (picky, picky):
  • 12AWG wire is overkill to the extreme, and might cause problems due to the stiffness. 18AWG is more than enough, and you could probably get by with 20AWG.
  • Use a real crimper. Pliers do not make good crimped joints.
  • You should definitely solder those connections to the rheostat. There's going to be a lot of resistance if you just bend/wrap the wires around the tabs. Worst case, the connections could overheat and mess up the tabs and/or wires.
  • UL1015 hookup wire, available at big box stores (Lowes, etc) and automotive stores, is a good choice for the wires. It has a little higher temp rating than your's, and it is has a UL rating (AWM) for use in appliances. This wire is nice and flexible, and the strands are tin coated to avoid oxidation.

mini (original poster)

#3: Post by mini (original poster) »

I appreciate the feedback! Like I said, this is my first electrical project so I don't know all of these practical things.

I realized that 12 gauge was overkill when I brought it home and compared it to my internal wiring. :lol: It shouldn't need to carry more than 0.4A, and 12 gauge is rated at... 10A, but I didn't feel like buying more. I certainly will look into the wire you mentioned and revise this as needed. I was able to bend mine ok, but it's certainly stiff. I think mine might be UL rated, but I didn't pay too much attention.

I didn't notice crimpers at the store, but the pliers were not the easiest method. Good tip.

Yeah... I'm going to solder it pretty soon. I was just excited and thought that it would make do for a little while at least. FWIW, my multimeter didn't pick up too much resistance from the connections, but oxidation is always a threat.


#4: Post by randytsuch »

Instead of crimpers, I cheat and use a pair of vise grips.
Not as good as a real crimper, but it seems to do a reasonable job.
After you finish a crimp, make sure you grab the wire with one hand, and the connector with the other, and give it a good pull. You should be able to pull as hard as you can, and the wire won't come out.

In my Classic, I squeezed in a pressure gauge on the side of the machine. It's very nice to be able to monitor the pressure as you pull a shot, especially since you can now control the pressure. The gauge is liquid filled from Amazon, and wasn't expensive. Fittings were from Mcmaster.

I am assuming you can't find anywhere inside your machine to mount the rheostat, so that made you mount it outside.
Maybe you could mount it in a project box, and attach the project box to the back of your machine?

My dimmer is mounted on the rear side of my Classic, close to the side corner. So, to adjust the pressure, I just have to reach around the back, and turn the knob.


Team HB

#5: Post by ira »

I make this offer again. I have some proper 14ga stranded wire with silicon insulation with a cloth braid for strength rated to 150C in black and white. I'll be happy to supply short lengths for cost and postage for those who want to do work on their espresso machine internals.


mini (original poster)

#6: Post by mini (original poster) »

I went back and checked my crimp connections. They are snug, but not unbreakable. I'll revise the instructions to reflect your advice.
randytsuch wrote:I am assuming you can't find anywhere inside your machine to mount the rheostat, so that made you mount it outside.
Maybe you could mount it in a project box, and attach the project box to the back of your machine?
Actually, I do have room inside I think. I just wanted to do a proof of concept and see how well it works before I start drilling holes in my machine casing. That and it's exam week for me.

ira, I might have to take you up on that offer. Is the wire pretty flexible?

Team HB

#7: Post by ira »

I'd guess it's a lot more flexible than what you're using.



#8: Post by frankmoss »

If using an older Ulka with an external diode, should the rheostat be placed before or after the diode?

mini (original poster)

#9: Post by mini (original poster) »

I didn't know that the older models were like that...

However, it should not matter either way because of electrical properties.

Oh, and I'm at home away from my equipment this week. I should start testing out my new mod in full next week. Like I said, I will post all my opinions about how it works.


#10: Post by frankmoss »

I'm eager to see how your tests turn out. I'm planing on doing the same mod, provided that yours works well.

However, I must respectfully disagree with this:
Adding resistors in series changes the pump strength in an impossible to predict non-linear fashion
Since the force created by a solenoid is F=(mu)NI/h, V=IR, and P=F/A, you should be able to approximate the pressure created by the pump. You would have to calculate/measure the resistance in circuit, the number of turns (N) in the solenoid, and the area (A) of the piston. Then, P=(mu)NV/(hRA). Therefore, it is theoretically possible to calculate to pressure created by a vibe pump, and the pressure is linear with respect to R. However, most resistors have high tolerances (~10%), so this makes the actual pressure deviate from the theoretical pressure.

In any case, we'll see how well it works in practice.