Optimizing power delivery

If you want to know my thoughts on the LR2 as well as recommended improvements you can contact me at bleepbleepbloop1234@gmail.com.

The topic of the thread was “performance boost” not “intro advice.” It’s clear there is some hostility here by the bias that is being portrayed & disproportionate reactions given to mere upgrade advice. I apologize if I made you upset with my criticisms of the LR2. Have a nice day.

I thought this point was worth addressing.

Setting aside the clear advantage of being able to run smaller wires and faster movement speeds, the stepper motor torque curve usually reaches a peak at 50-75 rpm or something to that effect (it depends on the motor). Before this point it declines, after this point it declines. Doing slower cuts does mean you are operating in a sub-optimal part of the torque curve. Moving at higher RPMS does produce higher torque at least before the peak, and at those higher RPMs the higher voltages do improve torque. You can see that clearly in examples such as this and this. Now the question is if it is reasonable to cut at the speeds where this higher torque band occurs. I will attempt to estimate this but it really depends on your particular setup.

A GT2 timing belt has 2mm per tooth which combined with a 16 tooth pulley produces 32 mm/rot. 50 RPM is 1600 mm/m which is roughly 27 mm/s. In other words, based on this estimate it is feasible, depending on the type of motor and cutting speeds, that you can reach a more optimal part of the torque curve at a reasonable feedrate for the LR2 and, at that RPM, 24v would probably improve torque. Someone running a 1/8" bit through MDF could, in my opinion, easily reach 27 mm/s. I believe the benefit would be far greater for motors with 400 steps/rev.

If you are running your motors at sub optimal values then, yes, the input voltage doesn’t matter. A racecar and a minivan both perform the same at 25mph. But, as I said before, I would be more comfortable continuing this conversation via email.

I don’t think we are far off in our technical opinions. And I thought we were all being pretty nice about the differences. If you don’t want to deal with a little criticism, then fine. I am disappointed you’re leaving. You seem like you want to help and share, which are both qualities we look for here.

1 Like

I think you got me confused with someone else, because I can’t understand where you got the impression i was upset or that I even had any opinion on the upgrades. I was just trying to show you that this crowd is motivated to help people, not tear them down. Many of the folks in here are also incredibly technical, and many of them professionals, where they’ve gotten used to debating merits and drawbacks with folks who aren’t emotionally attached to their arguments.

And dude, I GET where you’re coming from, but look at it this way (to borrow your auto analogy). Suppose a guy showed up in a car forum and said this :
“Hey guys, i just 350-swapped my miata, and i ran the quarter mile in 18 seconds. I think it should be faster, so if i were to throw money at this, what would be worth doing?”

I really do hope you’d have better ideas than telling him he should go look for an LS or strap a supercharger on his 350 or get better tires.

Yeah, they’re all performance boosts, but something looks REALLY wrong first.

As for other folks, I think everyone… EVERYONE in the thread who had questions about your setup said they wanted to see it. They have understandings of how things work based on other things they know. If those things aren’t right, their understandings may not be right. IIRC, Ryan even said he might have to update his recommendations if 24v really is that much better than 12. (maybe that was another thread?)

The people here change their minds and learn from each other, and they do it by experimenting, discussing, and back to back testing.

Don’t believe me? Search for “concrete filled tubes”. Lots of skepticism, then big claims about epoxy filed tubes, skepticism about that, then some testing, and now we all know that we can expect a little bit of improvement if done a particular way, and how much that improvement is. Or the double decker mpcnc. That one was fun.

Now, I’ve done my absolute best to help you understand what I think is going on here because you seem like a good dude and I think you belong here. Your comments in general (not just in this thread) have been positive and helpful, and I really appreciate your input.

@sparkm, I’m sorry this is your thread now.

3 Likes

Spark asked what could money be best thrown at. You provided an awesome answer. I should have just stayed out of it. 100%.

I can assure you no one here is offended, and I have said that before. I would offer the advice that over the last 7 years no one really gets wound up around her. If you read all the replies like we are very excited kids I think it might read more true. This is not like most other forums in that way. At the same time trust me when I tell you some people around here really really understand some of this stuff to a very very high level, it sounds like you are one of those guys. If we don’t understand something you do, try and help us understand like we are very excited kids.

Again, I am sorry for even poking my head in this one. I just got excited about what you have achieved with your build, I did not mean to sound so skeptical, I was just excited and wording things poorly.

4 Likes

PS I am very much in the weeds of the new LR build and I want it to be a significant jump in performance with no price increase. Some of your tweaks could do that…so that just proves why I am extremely interested in what you have to say. Please let me see this thing in action or help beta test the new build when it is time to do so!

2 Likes

FWIW, I have made some progress on the teething troubles I was having earlier. Thanks to the forum members who have been incredibly helpful and have provided insightful ideas. I just posted the first piece of furniture I built using LR2 (Rocking Chair).

The last cut I made was at 10mm DOC @ 10 mm/sec in plywood and the machine appeared more than up to the task. This is without any mods and following Ryan’s original specs. My plan is to incrementally push it more to a point where things start to fail. Then figure out the bottleneck and may be try out some of the enhancements proposed in this thread.

3 Likes

So, more than twice as deep and almost 50% faster? That sounds like a huge win to me, nice!

That rocking chair looks great, too!

1 Like

It’s all good Ryan, I don’t see any daggers being thrown in these posts. If anything a bit of tempering of the support lines… you are the guy who would be answering questions about why 24V didn’t work etc so no problem highlighting that line in the sand for yourself here. If anything, this thread displays the exceptional humility and understanding to be found on these forums.

No bridges burned I don’t think… maybe just some time away to think about things before returning to the office. I also hope he returns to partake in mpcnc tech discussions. If this discussion could be categorized as abrasive or whatever… this is nothing compared to some of the spirited talks I’ve had at work… only to make the best of friends with my adversaries over beers afterwards. The important ingredient, is that the truth is the pillar that we all find ourselves wrapped around in the end… west, east, north, south side of that pillar does not matter.

3 Likes

A 24v power supply costs basically the same as a 12v one, has the benefit of allowing for smaller/cheaper wiring, drastically faster movement speeds, and moderately faster cutting speeds. Why would you go 12v? Given the option between the two, it makes no sense to go with 12v. I don’t really see where this hostility is coming from towards such an obvious design choice. The only reasons I could imagine is because you already have board that you want to reuse and it doesn’t support 24v. That’s incredibly niche scenario. SKR boards are cheap, work super well, and support 24v. If you are going to buy a board and a power supply, just go for 24v. Similar logic can apply for everything else. The cost differential between 70 oz/inch motors and 92 oz/inch motors is negligible. It cost me like 20 bucks to add in second x motor. We’re not talking about breaking the bank here. These are cheap, common sense upgrades.

1 Like

I made these choices, not anyone else.

The real reason we use 12V is I have a supplier for reliable 12v and the same suppliers 24v are not as good. The initial reason I even sought out 12V suppliers was large format 3D printing was by far the most common use for MPCNC’s for the first few years…7 years ago.

As for changing to 24v, we had a user to a very extensive study and I did not see the same reasult you did in his numbers. You keep saying drastically better, his numbers say otherwise. He shows that wired in series there is no power left at 200mm/s. Considering we now run individually and I limit the firmware to 50mm/s that still not all that important.

Our wires are under 5’ in length so I am not really seeing a difference there either, Wire Size & Voltage Drop Calculator | GRE Alpha we use just under 1A per stepper typically.

I am still consider switching to 24V though. It is no doubt a better option, the only debate is to what degree, I feel it is mild and should have no day to day effect for the average user.

1 Like

Yeah, i don’t know if you’re confusing me with someone else, but i haven’t made an opinion on your suggestions. Or maybe you’re still trying to convince me that the pushback is hostility that I’m just not seeing?

Either way, I’m not going to make more posts in this thread unless the OP has more questions.

It might be useful to have a new thread dedicated to debating any or all of your machine choices where you can open with your argument and/or some video of how well it works.

OP seems to have solved his original issue and is on a good path to find the limits of his current setup.

1 Like

Here is the other one, MPCNC: Stepper Drive vs. Back EMF at the Edge of Madness – The Smell of Molten Projects in the Morning. Still, the issues are with series wired steppers above 133mm/s…I stopped selling those and limited the speed… This is also with a4988 drivers.

I now tell people to work deeper and slower. This is best for the endmill and the much improved machines 5 years later handle the load much better.

I just want you to see I do take these things into consideration but not just from some posts, I need to see proof or contradictory findings. You say your works amazing, and I just really want to see that in action and see the accuracy it holds at those speeds. As you can imagine, I have people telling me things on all platforms about how my machines should and shouldn’t work. I only trust tests and proof.

1 Like

You keep saying this, but it doesn’t make any sense.

W use current controlled drivers. Wiring gauge needed is a function of the current carried and the length of the wire. Since we have the same current regulated, how can you reduce the size of the wires?

Edit for clarity: I’m asking for an explanation, because nothing that I’ve read anywhere else suggests that thinner gauge wiring is or should be possible. If you have 1.5A of current over 2m runs, you need at least a certain thickness of copper, or else you will lose significant power to parasitic losses.

1 Like

The only thing I can think of is that the voltage drop in the wiring is greater if the resistance is higher (which happens with smaller wires). If you had very small wire or very long runs, you would have enough voltage drop in the wires to have problems with 12V. But with 24V, the voltage drop from the smaller wires would be tolerable because you have more voltage across the coils (only when on, the average voltage would be the same).

But you can’t have your cake and eat it too. If you use smaller wires, you are effectively using a lower voltage power supply. At some point, the voltage drop would be 12V and you would have the same qualities of a 12V PSU. Maybe you would prefer somewhere in between, like 6V of drop in the wires, so you can use tiny wires, and then you get some of the high end speed boost, but not as much as using 24V and larger wires.

My first impression was that it was being confused with advice for 3D printer beds. With a constant voltage power supply in a 3D printer, using a 24V heater with the same wattage as a 12V heater means less current through the wire. So you can have the same heat and power loss in smaller wires for a 24V bed. But that does not continue to stepper drivers.

I am giving the benefit of the doubt that this isn’t the obvious reason to use smaller wires, and instead, it is based on the voltage drop.

Right. When I did stuff for automotive installations, we always calculated wire gauge based off of the current load and the length of the wire run. It’s severely oversimplified, because we were always dealing with nominal 12V circuits, but several handbooks noted that 24V automotive systems should be run with the same guidelines, except that with 24V, you would be dealing with twice the power for the same current draw. (Which is why a 24V heater cartridge is 4X the DC resistance of a 12V heater cartridge.) For an EE course that I took (34 years ago, egad!) we got into a bunch of calculus, but what I remember is that we were looking at power dissipation based on current and length. This was all AC, but the difference between 9VAC an 50kV transmission lines were basically that the 50kV transmission lines were useful because they needed to carry less current for the same amount of power.

Not related to wire thickness:

Taking some sample numbers from the NEMA17 motors supplied by V1, these are nominal 3.36V coils. taking this and figuring out their likely back EMF numbers, I come up with about 0.65V/radian/second. For this, they generate enough back EMF to no longer be able to hold peak torque (2A) at about 13.25radians/second, which is a bit over 125RPM. This more or less agrees with the chart that @BleepBleepBloop posted to earlier, so I think my math is reasonable. This is a little better than 4000mm/min. My fastest cutting profiles are 2400mm/min. which means that I’m in no danger of slowing my machines down at 12V.

One function that I’ve come across reading up on this though is acceleration. There is apparently a difference with some drivers in how fast the driver can saturate the inductive load based on V(mot) which leads to peak torque sooner. This pretty much only applies to a motor at rest, but the formula for inductive saturation does take into account voltage. This may mean that higher voltage can be used to increase the acceleration factor, but some care needs to be applied depending on top speed, as the numbers change rapidly as soon as the motor is moving. (I got to use calculus!) In terms of our usage it probably really only applies as “instantaneous speed change” and is only going to have millisecond impacts on machine time. Insofar as I could tell, the drivers where this is even mentioned are the large stgand-alone drivers more commonly found for 6-wire NEMA23 or larger motors.

1 Like

From what i read 24V seems to be better for versatility, doesn’t it ?
If you want to use a laser, it seems even milliseconds on instantaneous speed change would be good to take.

I don’t have anything to debate or compare with here, from the start i use 24V on my SKR Pro as i reused it from an old disassembled 3D printer. It’s working nicely at no cost, i don’t ask for more :smiley:

3 Likes

If you have it use it!

1 Like

Anytime the resistance in the wiring becomes an issue, higher voltage performs better. This is the case when running small and/or long wires. If your build has fans, relays, endstops, etc, these are things that can use higher gauged wiring with a 24v power supply. It depends on your exact build and can vary wildly. As someone else pointed out you don’t want to use them on your motors if your priority is having high voltage as measured at the motor. The wires create a voltage drop. If that’s not a concern then you can just run smaller wires for the motors too.

I am not criticizing your kit. I couldn’t possibly know all the many reasons and considerations that went into planning that. The topic is upgrades and a 24v power supply is an option with nice benefits. If someone is going to upgrade their board or power supply, or are just looking for an upgrade for an upgrade’s sake, it makes sense to go for a 24v power supply. That’s all I am saying.

1 Like

If all else was equal, sure. The plethora of 12V accessories and other things makes 12V useful for more things, not to mention economy of scale when it comes to things like fans. 24V does have some possibilities that would be nice to haves.

It would be, yes. You have to tune for it though. The biggest barrier to me there is that practically all diode lasers that I’ve seen want 12V power, so they either require an additional buck-boost converter or a separate power supply for the laser itself. It seemed much easier to me to just power my laser control board from the same 12V supply that I was using for the laser. (Or, more accurately, power the laser from the control board’s Vin power.) So, perhaps ironically, my laser is one place where I deliberately avoided 24V power.

My LowRider runs on 12V, and because I’m not overly concerned with acceleration, and won’t ever be running faster than 4000mm/min, I’m not worried about switching it to 24V. I scavenged the power supply from somewhere else. If it dies, I’ll replace it with a 24V supply, because as Ryan noted, they cost about the same, and if I need to buy a new power supply, I might as well go for the higher voltage option.


Given the nature of our stepper drivers, voltage remains the same. Net voltage on the line, accounting for inductive loading, including the parasitic inductance of the wire itself will never exceed the voltage required to drive the set point of the motor driver. That is the very definition of a current controlled driver. If the net voltage to the stepper exceeds the amount required to drive the rated current to the stepper, the stepper will exceed its ability to dissipate heat, and the motor coils will fail. While a higher “instantaneous” voltage may occur, (If such a thing actually existed in the real world) the inductive nature of the load itself will resist it, resulting in an effective voltage drop. It is the driver’s job and nature to prevent this from occurring for long enough to make an actual difference. Any driver that fails to do this will, over time, cause the motor to fail, unless you set a larger “safety margin” whereby you effectively reduce the motor’s capability. (Most of us actually do this anyway, because we want to reduce the motor temperature so as not to soften the PLA plastic parts.)

If you thin the wires so that parasitic resistance becomes an issue worth noticing at what are still relatively low voltages, you’re introducing all kinds of trouble with approaching power limits for stepper motors.

So I see where you’re coming from for the thinner wires, however, I believe that you are mistaken in this case.

The real limit with 12V is the max torque speed limit of about 4000mm/min, (Assuming 16T GT2 pulleys) which the vast majority of users here will never run into, as few of us even run travel moves at that speed. (I actually have travel moves limited in firmware to 7200mm/min on my Primo, but I think they’re programmed in CAM at 3600mm/min.)

The max acceleration is based on the max torque (before skipping steps), which is roughly proportional to current, not source voltage.

You don’t get more acceleration with 24V vs. 12V. At least not at low speeds.

There are a lot of rumors around the benefits of 24V. And I think that is what makes me speak up about it.

At higher speeds, you can lose torque with 12V vs. 24V. That (IMO) is the big benefit.

24V doesn’t hurt though. Yes, you have trouble with some accessories. But it isn’t worse at driving stepper motors. I wouldn’t buy a 24V PSU if you already have a 12V PSU. It will just spend more time idle than a 12V.

The coil resistance on the motors is very small. 3Ohms. All you have to get is 1Ohms in the wiring to lose a quarter of your voltage to the wiring. I’m just talking regular resistance, not inductance.

28awg at 5m is 1 Ohm. 2.5m each way. So 3V loss.
24awg at 5m is 0.4 Ohm or so, so only 1.4V loss.
22awg at 5m is 0.16 Ohm. 0.6V

It’s not insignificant.

1 Like