Wondering what the compelling reasons would be to upgrade a machine ( or build a new machine) to use Servos vs Steppers.
Servos are more powerful, smoother and are closed loop. The latter meaning if you lose position the controller immediately knows about it, which means you know about it, which means you can save your work in most cases should something go wrong. Steppers are dumb and don’t know when they lose position.
On the other hand, Nema 34 steppers are cheaper, still pretty powerful, and if used right can crank out the work. Before working here I had a Nema 34 system (different machine, similar power) and I ran the heck out of it, while consciously knowing it’s limits and I did fine.
Back then servos were WAY more money, now they’re more, but not much more. If I was buying a machine today I’d go servos, but that’s me personally.
Thanks Eric …. I think 3K on a 10k machine is significant so I’m just looking for the compelling reason(s) to consider them. Given that we arent a production shop, but a makerspace, perhaps they would be overkill.
I’d say in a makerspace you SHOULD have them… With potentially a lot of users who aren’t good at using the machine servos provide a really nice safety system should someone run a job too fast, or if they crash the machine.
Here’s the problem .. its just too expensive.
I have two machines now, with 4 more sometime in the next 12 months. At 3k per that’s 18k … nearly the cost of two machines … (our 3636 configs come in at about 11k)
When a customer considers this, there is a value assessment made … your website does little to help us understand that and given our years of experience with First machine, a PRO, with literally no crashes Its hard to see.
I would like to understand more, where can I find out more?
Jim
Eric, are there any statistics on how much less noise and electricity servo systems make/use? In my experience, servos run cooler (i.e. less electricity so less $$) and quieter, which might be a key factor in a maker space.
It would be amusing if the ROI on servos just for electricity use was short.
We could certainly do better, but we do have some information here:
and here: https://youtu.be/xWnxrotEBkY?t=300
I think the opening line of the video is very relevant to you.
If your users aren’t crashing the machine due to lost position than I think you’ve answered your question, steppers will be great for you!
I highly doubt there’s any real difference between the systems. They both use the same power supplies.
Thanks Eric, we did see that.
As a non-production shop, where speed isn’t a factor, it’s hard to see from those descriptions why we would choose anything but NEMA23 … AVID represents it as “extremely reliable” and compromising speed only. This is consistent with our experience with our Pro machine. Mechanically it has been solid, Mach4 on the other hand has been a trainwreck.
The price differential as a percentage of total acquisition cost is huge … for us it’s the difference between 5 and 6 machines …. In my opinion AVID should quantify the benefits. With what you have provided I could not defend the increased cost.
While power wont be compelling for us (we don’t pay for it) if there is a significant savings that’s the kind of thing you should highlight, along with quantified speed
We have a CRP800 to EX upgrade in process to address the mach issues.
Jim
If you are going to push your system hard, like you are trying to cut sheets of plywood for a cabinet shop all day long, the higher power and not losing steps is nice. This would be for a pro system.
If you are running a desktop with ballscrews, you aren’t likely to be pushing those kinds of speeds, so that probably doesn’t matter, and I doubt if you would have problems losing steps with ballscrews due to the higher turns ratio. Also, a desktop has a lot less moving mass in the gantry than a large pro system.
If you are doing precision work, you may consider that the steppers offer a lot more resolution in the motor movement than these particular servos. It can be many times better with the stepper in a system with the R&P drive (like 0.001” vs 2-8x less than that). With ballscrews, the resolution is already much better than what matters on a cnc router so it doesn’t matter.
The servos are definitly smoother in operation. So if you have problems running details carvings that have a lot of rapid direction changes, or similar with laser work, you can definitly reduce some vibration in the whole system. Kinda depends on how you are using your machine as to if it will really be beneficial or not.
As far as electricity goes, a 4hp spindle plus your 15A, 120V control box power input running full out is about 5400W. So if your electricity is 10c/KwH, the machine is costing you about $0.54/hr to run. Again thats full out max power, which you are never gonna do. The spindle is 66% of that overall. The motors are using the same power supplies, and have to move the same mass, so no, the different motors aren’t going to change the cost. A 3hp dust collector is likely going to average more power that the whole cnc.
The servos do have a coolness factor to them though 
1 Like
I’m a little surprised to hear this. IIRC, Mark Rober’s puzzle video talks about using servos to improve resolution, and the specs have better resolution listed for servos. Is there something you noticed when you switched to servos?
The steppers rely on proximity switches for homing. If I want an accurate rotary centerline, I sometimes have to adjust my work offset a little after homing (I have NEMA 23 steppers and legacy Mach 4 controls). Would switching to servos (with their hard-stop homing) address this issue?
Likely yes. The hard stop homing is a VERY accurate way of homing.
This is a very cool video showing how it works: https://youtu.be/KLUKn59T1GQ
Great, detailed response Jim N … for our application (a makerspace) NEMA23 steppers have worked out fine. The only time we had any mechanical issues is when we neglected to replace the gears as prescribed in the maintenance procedures for our PRO machine. Once that was addressed no mechanical issues. Any issues we had with stalls was traced to MACH and the i5 PC we were using .. once we replaced the i5 with an i7 stalling was vanquished. We chose nema 23 for cost, and, we aren’t a production shop … the machine doesn’t cut cabinet panels all day, rather its pumping out smaller projects. Some are demanding and might benefit from more power but its not essential. As we add more machines the cost savings of steppers over servos multiply for us As for power costs, we are fortunate to be part of a community woodshop & makerspace … power is supplied by the community association, so its cost is not a primary concern. Thanks again for a very detailed and helpful response.
There are two types of clearpath servos - the regular ones have 800-line encoders, and the “enhanced option” one has 6400-line encoders. So the regular ones, like Avid uses, are comparable to a 200-step stepper with 4x microstepping, and the enhanced are like 32x microstepping. Regular steppers are somewhere between those.
Teknic may have some software magic to improve the resolution beyond what the encoders use (like using all four line edges to simulate 3200/25600 “lines”, but you need some of those lines to avoid hunting), of course, but if you wanted “servos are more accurate” you’d want to spend the extra $70/motor to get the enhanced option.
FYI on a PRO system, 1000 steps per rotation works out to about 0.001” per step, well within the 0.005” tolerance Avid claims for the machine.
These servos have 12800 step resolution on the encoder, but only 800 commandable steps. The motor needs more resolution in the encoder so its internal controller can observe and manage to hold itself to the commanded position. If you just push on an axis slightly you can here it and see it doing that, especially if you put a dial indicator on it. Clearpath does have a higher resolution motor that has the same encoder but much higher commandable resolution.
The default stepper setup AVID has microstepping at 2000, and you can set it a couple multiples of 2 higher if you want.
correct, its not a big deal for most applications.
I did a bunch of repeatability testing on the Mach4 based system with steppers and the standard proximity sensors and was able to pretty reliably repeat a position on the gantry (measured with a dial indicator) to +/- 0.001” or so. So maybe you have another issue, like worn belts/gears?
Hi SunCityJim,
Maybe consider setting up only one of your machines with servo’s, then your use cases over time will help solidify the decision for your other macbhines.
Just a thought…
Sounds like I need to start by installing the replacement gears I ordered a while back.