What are the variations in performance specifications between "Pro" and "Hobby" servos?
I've heard a lot of claims, but it's easy enough just to go look at the published specs. Here, I'll take a Yaskawa 750W Sigma-5 AC Servo (typical axis drive for a Haas VMC) against the HomeshopCNC 850 oz-in DC servos I have installed on my IH mill. Here's what I get from the respective (linked) spec pages:
- Power: Yaskawa = 750W / HomeshopCNC = 612W (a little smaller servo)
- Peak Torque (oz-in): Yaskawa = 1184 / HomeshopCNC = 850
- Max Speed: Yaskawa = 6000 rpm / HomeshopCNC = 4200 rpm
- Voltage: Yaskawa = 200V AC / HomeshopCNC = 72V DC
One interesting figure is to take the power consumption of the two and equalize what the torque should be from the HomeshopCNC were it to consume 750W. That value is 1042 oz-in, which is still less than the 1184 oz-in from the Yaskawa. AC servos are known to be more efficient, and that sort of quantifies it a little bit at 12-15% more efficient in producing torque for a given amount of power.
This is all just out of idle curiousity and interest, but there are intangibles to consider too.
I correspond with a lot of machinists who bring me ideas for this blog or otherwise just want to talk over something they see here. My post on Yaskawa vs “Hobby” servos shook loose an interesting one. This is a machinist I’ve had a lot of great discussions with who does a lot of retrofits. One of his special niches (every shop needs a special niche the competitors can’t go!) is that his machines are simply bigger than the others nearby. If you have a rush job needing 60 inches of travel on a mill, he’s the man you have to talk to.
He’s gotten a lot of these machines together by retrofiting older machines. In response to my Yaskawa musings, he wanted to point out that there are a lot of intangibles that the raw specifications don’t speak to. He went through an evolution of sorts on the retrofitting. First he went with big steppers. Like any successful machine shop, he’s cranking out tons of parts. That’s how you make machine tools print money. As he puts it:
I would run a ton of parts. Come back and do a home check. It would be off .005 or .010″. Crap, you wonder how many parts are scrap. Was it loss of lube? Junk in the ways? I could really never tell. Maybe the ballscrews had rubbish in them?
The CNCZone crowd are fond of saying that a properly designed stepper system never loses steps, so there are no advantages to servos. I’ve had one of the leading luminaries in stepper drive design say that you can clearly hear when a stepper motor has dropped steps. He’s a great stepper drive designer (he really is, his drives are awesome), but he is not a machinist. Heck, I’m lucky to hear myself think when a big machine gets to moving some chips big time, let alone hear whether some stepper briefly squealed a little bit differently. Imagine just losing 1 or 2 steps on every 2nd or 3rd part all day long. That’s how you get to be off 5 or 10 thou at the end of the day.
I’ve also had my servos fault often enough, and known why they faulted, to be completely confident that peeps are losing steps out there from time to time, whether or not they want to admit it. Get a gib acting too tight due to not enough lube or junk in the ways like my machinist friend mentions and that’s all it takes. The machine will overcome it in many cases rather than stalling, but you’ve already lost the step. So you’re off by the end of the day. Crap.
From steppers, he went on to the “Hobby Grade” DC servos with brushes and all. I’ll let him tell the story again:
I went to the DC servo stuff. Way better than stepper stuff. I can actually run at 100 IPM!! Well, this stuff zeros the same all the time. Closed loop is nice. BUT, the motors got hot after extended use. I had several motors smoke and fail. I took them apart and looked at the charred brushes, they looked OK. Maybe the brush hung up in the holder? I sent in the analog servo card for factory evaluation. No error codes etc. just a red led telling me of error. Servo guys say it is fine…. I still have to this day marginal electrical connectors on those cards. Molex pin type connectors. Not designed for high current use. They get brown from the heat. I’d junk the stuff in a heartbeat if it was not for the downtime and expense. Think of all the money spent on Tagamet.
I can imagine the Tagamet. Rushing to get a job done, behind the 8 ball, and the machine is not working right after it always had. It’s like having a car that breaks down right when you need it most. Drives you crazy.
The comments about type of connector and their reliability are spot on. When you think about it, these servos and steppers live in a pretty nasty environment. Metal chips are flying everywhere, and will short out a contact in an instant. The airwaves are rife with electrical noise from the VFD on the main spindle to the guy that fires up the welder at the other end of the shop. Coolant is pooling around these bad boys hour after hour, are we running a machine tool or a submarine? And there is no rest. Got to grind out the parts constantly. No time for preventative maintenance to speak of. Run it until it breaks and then fix it ASAP.
As my friend puts it, a lot of hobby systems are a cascade of “it will work” type thinking. Not the result of ” let’s build the best” type thinking. It really does make a difference.
Yaskawas have the mojo, they are the best, and that’s why so many machines use them.
Be the first to know about updates at CNC Cookbook
Join our newsletter to get updates on what's next at CNC Cookbook.